diff options
Diffstat (limited to 'samples/source')
28 files changed, 1757 insertions, 5155 deletions
diff --git a/samples/source/CustomSchema.cpp b/samples/source/CustomSchema.cpp index 7e159b4..d29c10f 100644 --- a/samples/source/CustomSchema.cpp +++ b/samples/source/CustomSchema.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms diff --git a/samples/source/DumpMainXMP.cpp b/samples/source/DumpMainXMP.cpp index 06d367a..389e10a 100644 --- a/samples/source/DumpMainXMP.cpp +++ b/samples/source/DumpMainXMP.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2002 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -19,7 +19,7 @@ #include <stdexcept> #include <errno.h> -#if WIN_ENV +#if XMP_WinBuild #pragma warning ( disable : 4127 ) // conditional expression is constant #pragma warning ( disable : 4996 ) // '...' was declared deprecated #endif @@ -108,19 +108,19 @@ main ( int argc, const char * argv [] ) printf ( "## SXMPMeta::Initialize failed!\n" ); return -1; } - - if ( ! SXMPFiles::Initialize() ) { + + XMP_OptionBits options = 0; + #if UNIX_ENV + options |= kXMPFiles_ServerMode; + #endif + + if ( ! SXMPFiles::Initialize ( options ) ) { + printf ( "## SXMPFiles::Initialize failed!\n" ); return -1; } - if ( argc != 2 ) // 2 := command and 1 parameter - { - printf( "usage: DumpMainXMP (filename)\n"); - return 0; - } - - ProcessFile ( argv[1] ); + for ( int i = 1; i < argc; ++i ) ProcessFile ( argv[i] ); SXMPFiles::Terminate(); SXMPMeta::Terminate(); diff --git a/samples/source/DumpScannedXMP.cpp b/samples/source/DumpScannedXMP.cpp index 37929c6..5ef137b 100644 --- a/samples/source/DumpScannedXMP.cpp +++ b/samples/source/DumpScannedXMP.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2002 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -19,7 +19,7 @@ #include <stdexcept> #include <errno.h> -#if WIN_ENV +#if XMP_WinBuild #pragma warning ( disable : 4127 ) // conditional expression is constant #pragma warning ( disable : 4996 ) // '...' was declared deprecated #endif @@ -154,7 +154,7 @@ main ( int argc, const char * argv [] ) return 0; } - ProcessFile ( argv[1] ); + for ( int i = 1; i < argc; ++i ) ProcessFile ( argv[i] ); SXMPMeta::Terminate(); return 0; diff --git a/samples/source/ModifyingXMP.cpp b/samples/source/ModifyingXMP.cpp index 7b73bc0..5ff0003 100644 --- a/samples/source/ModifyingXMP.cpp +++ b/samples/source/ModifyingXMP.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -165,9 +165,14 @@ int main ( int argc, const char * argv[] ) cout << "Could not initialize toolkit!"; return -1; } - + + XMP_OptionBits options = 0; + #if UNIX_ENV + options |= kXMPFiles_ServerMode; + #endif + // Must initialize SXMPFiles before we use it - if(SXMPFiles::Initialize()) + if(SXMPFiles::Initialize(options)) { try { @@ -240,7 +245,7 @@ int main ( int argc, const char * argv[] ) // This will: // a) Add ANY new TOP LEVEL properties in the source (rdfMeta) to the destination (meta) // b) Replace any top level properties in the source with the matching properties from the destination - SXMPUtils::AppendProperties(rdfMeta, &meta, (kXMPUtil_DoAllProperties | kXMPUtil_ReplaceOldValues )); + SXMPUtils::ApplyTemplate(&meta, rdfMeta, kXMPTemplate_AddNewProperties | kXMPTemplate_ReplaceExistingProperties | kXMPTemplate_IncludeInternalProperties); // Display the properties again to show changes cout << "After Appending Properties:" << endl; diff --git a/samples/source/ReadingXMP.cpp b/samples/source/ReadingXMP.cpp index b1ec3c4..f3f1333 100644 --- a/samples/source/ReadingXMP.cpp +++ b/samples/source/ReadingXMP.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -76,9 +76,12 @@ int main ( int argc, const char * argv[] ) cout << "Could not initialize toolkit!"; return -1; } - + XMP_OptionBits options = 0; + #if UNIX_ENV + options |= kXMPFiles_ServerMode; + #endif // Must initialize SXMPFiles before we use it - if ( ! SXMPFiles::Initialize() ) + if ( ! SXMPFiles::Initialize ( options ) ) { cout << "Could not initialize SXMPFiles."; return -1; diff --git a/samples/source/XMPCoreCoverage.cpp b/samples/source/XMPCoreCoverage.cpp index c0c2cef..0fb766c 100644 --- a/samples/source/XMPCoreCoverage.cpp +++ b/samples/source/XMPCoreCoverage.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2002 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -287,6 +287,30 @@ static const char * kDateTimeRDF = // ------------------------------------------------------------------------------------------------- +static void FillDateTime ( XMP_DateTime * dateTime, XMP_Int32 year, XMP_Int32 month, XMP_Int32 day, + XMP_Int32 hour, XMP_Int32 minute, XMP_Int32 second, + XMP_Bool hasDate, XMP_Bool hasTime, XMP_Bool hasTimeZone, + XMP_Int8 tzSign, XMP_Int32 tzHour, XMP_Int32 tzMinute, XMP_Int32 nanoSecond ) +{ + + dateTime->year = year; + dateTime->month = month; + dateTime->day = day; + dateTime->hour = hour; + dateTime->minute = minute; + dateTime->second = second; + dateTime->hasDate = hasDate; + dateTime->hasTime = hasTime; + dateTime->hasTimeZone = hasTimeZone; + dateTime->tzSign = tzSign; + dateTime->tzHour = tzHour; + dateTime->tzMinute = tzMinute; + dateTime->nanoSecond = nanoSecond; + +} // FillDateTime + +// ------------------------------------------------------------------------------------------------- + static void WriteMajorLabel ( FILE * log, const char * title ) { @@ -397,11 +421,9 @@ static void DoXMPCoreCoverage ( FILE * log ) // -------------------------------------------------------------------------------------------- - WriteMajorLabel ( log, "Dump predefined namespaces and aliases" ); + WriteMajorLabel ( log, "Dump predefined namespaces" ); SXMPMeta::DumpNamespaces ( DumpToFile, log ); - fprintf ( log, "\n" ); - SXMPMeta::DumpAliases ( DumpToFile, log ); // -------------------------------------------------------------------------------------------- @@ -480,118 +502,6 @@ static void DoXMPCoreCoverage ( FILE * log ) #endif // -------------------------------------------------------------------------------------------- - // Static alias functions - // ---------------------- - - WriteMajorLabel ( log, "Initial default aliases" ); - fprintf ( log, "\n" ); - - SXMPMeta::DumpAliases ( DumpToFile, log ); - - WriteMajorLabel ( log, "Add ns2: to ns1: aliases" ); - fprintf ( log, "\n" ); - - SXMPMeta::RegisterAlias ( kNS2, "SimpleAlias", kNS1, "SimpleActual" ); - - SXMPMeta::RegisterAlias ( kNS2, "BagAlias", kNS1, "BagActual" ); - SXMPMeta::RegisterAlias ( kNS2, "ns2:SeqAlias", kNS1, "SeqActual" ); - SXMPMeta::RegisterAlias ( kNS2, "AltAlias", kNS1, "AltActual" ); - SXMPMeta::RegisterAlias ( kNS2, "AltTextAlias", kNS1, "AltTextActual" ); - - SXMPMeta::RegisterAlias ( kNS2, "BagItemAlias", kNS1, "BagActual", kXMP_PropValueIsArray ); - SXMPMeta::RegisterAlias ( kNS2, "SeqItemAlias", kNS1, "ns1:SeqActual", kXMP_PropArrayIsOrdered ); - SXMPMeta::RegisterAlias ( kNS2, "AltItemAlias", kNS1, "ns1:AltActual", kXMP_PropArrayIsAlternate ); - SXMPMeta::RegisterAlias ( kNS2, "AltTextItemAlias", kNS1, "ns1:AltTextActual", kXMP_PropArrayIsAltText ); - - SXMPMeta::DumpAliases ( DumpToFile, log ); - - WriteMajorLabel ( log, "Resolve ns2: to ns1: aliases" ); - fprintf ( log, "\n" ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS1, "SimpleActual", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns1:SimpleActual : %s\n", FoundOrNot ( ok ) ); - - fprintf ( log, "\n" ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "SimpleAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:SimpleAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - fprintf ( log, "\n" ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "ns2:BagAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:BagAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "SeqAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:SeqAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "AltAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:AltAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "AltTextAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:AltTextAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - fprintf ( log, "\n" ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "BagItemAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:BagItemAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "SeqItemAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:SeqItemAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "AltItemAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:AltItemAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - tmpStr1.erase(); tmpStr2.erase(); - ok = SXMPMeta::ResolveAlias ( kNS2, "AltTextItemAlias", &tmpStr1, &tmpStr2, &options ); - fprintf ( log, "ResolveAlias ns2:AltTextItemAlias : %s, %s %s, 0x%X\n", FoundOrNot ( ok ), tmpStr1.c_str(), tmpStr2.c_str(), options ); - - { - SXMPMeta meta; - - WriteMajorLabel ( log, "Test SetProperty through ns2: simple aliases" ); - - meta.SetProperty ( kNS2, "SimpleAlias", "Simple value" ); - meta.SetProperty ( kNS2, "ns2:BagItemAlias", "BagItem value" ); - meta.SetProperty ( kNS2, "SeqItemAlias", "SeqItem value" ); - meta.SetProperty ( kNS2, "AltItemAlias", "AltItem value" ); - meta.SetProperty ( kNS2, "AltTextItemAlias", "AltTextItem value" ); - - DumpXMPObj ( log, meta, "Check for aliases and bases" ); - - } - - #if 0 - - WriteMajorLabel ( log, "Delete some ns2: to ns1: aliases" ); - fprintf ( log, "\n" ); - - SXMPMeta::DeleteAlias ( kNS2, "ns2:SimpleAlias" ); - SXMPMeta::DeleteAlias ( kNS2, "SeqAlias" ); - SXMPMeta::DeleteAlias ( kNS2, "AltAlias" ); - SXMPMeta::DeleteAlias ( kNS2, "SeqItemAlias" ); - SXMPMeta::DeleteAlias ( kNS2, "AltItemAlias" ); - - SXMPMeta::DumpAliases ( DumpToFile, log ); - - #endif - - WriteMajorLabel ( log, "Register standard EXIF aliases" ); - fprintf ( log, "\n" ); - - SXMPMeta::RegisterStandardAliases ( kXMP_NS_EXIF ); - - SXMPMeta::DumpAliases ( DumpToFile, log ); - - // -------------------------------------------------------------------------------------------- // Basic set/get methods // --------------------- @@ -943,13 +853,15 @@ static void DoXMPCoreCoverage ( FILE * log ) { SXMPMeta meta ( kDateTimeRDF, strlen(kDateTimeRDF) ); - XMP_DateTime dateValue = { 2000, 1, 2, 3, 4, 5, 0, 0, 0, 0 }; - bool boolValue; - long intValue; - double floatValue; - char dateName [8]; + XMP_DateTime dateValue; + bool boolValue; + XMP_Int32 intValue; + double floatValue; + char dateName [8]; WriteMajorLabel ( log, "Test SetProperty... and GetProperty... methods (set/get with binary values)" ); + + FillDateTime ( &dateValue, 2000, 1, 2, 3, 4, 5, true, true, false, 0, 0, 0, 0 ); meta.SetProperty_Bool ( kNS1, "Bool0", false ); meta.SetProperty_Bool ( kNS1, "Bool1", true ); @@ -1129,7 +1041,7 @@ static void DoXMPCoreCoverage ( FILE * log ) meta2.SetProperty ( kXMP_NS_PDF, "Author", "PDF Author" ); tmpStr1.erase(); - meta2.SerializeToBuffer ( &tmpStr1, kXMP_ReadOnlyPacket | kXMP_WriteAliasComments ); + meta2.SerializeToBuffer ( &tmpStr1, kXMP_ReadOnlyPacket ); WriteMinorLabel ( log, "Read-only serialize with alias comments" ); fprintf ( log, "%s\n", tmpStr1.c_str() ); @@ -1137,7 +1049,7 @@ static void DoXMPCoreCoverage ( FILE * log ) meta2.SetProperty ( kXMP_NS_XMP, "Actual", "XMP Actual" ); tmpStr1.erase(); - meta2.SerializeToBuffer ( &tmpStr1, kXMP_ReadOnlyPacket | kXMP_WriteAliasComments ); + meta2.SerializeToBuffer ( &tmpStr1, kXMP_ReadOnlyPacket ); WriteMinorLabel ( log, "Read-only serialize with alias comments (more actuals)" ); fprintf ( log, "%s\n", tmpStr1.c_str() ); } @@ -1475,22 +1387,6 @@ static void DoXMPCoreCoverage ( FILE * log ) } } } - - SXMPIterator iter2 ( meta, kXMP_IterIncludeAliases ); - WriteMinorLabel ( log, "Iterate showing aliases" ); - while ( true ) { - tmpStr1.erase(); tmpStr2.erase(); tmpStr3.erase(); - if ( ! iter2.Next ( &tmpStr1, &tmpStr2, &tmpStr3, &options ) ) break; - fprintf ( log, " %s %s = \"%s\", 0x%X\n", tmpStr1.c_str(), tmpStr2.c_str(), tmpStr3.c_str(), options ); - if ( ! (options & kXMP_SchemaNode) ) { - tmpStr4.erase(); - options &= ~(kXMP_PropIsAlias | kXMP_PropHasAliases); // So the comparison below works. - ok = meta.GetProperty ( tmpStr1.c_str(), tmpStr2.c_str(), &tmpStr4, &opt2 ); - if ( (! ok) || (tmpStr4 != tmpStr3) || (opt2 != options) ) { - fprintf ( log, " ** GetProperty failed: %s, \"%s\", 0x%X\n", FoundOrNot(ok), tmpStr4.c_str(), opt2 ); - } - } - } } } @@ -1688,8 +1584,9 @@ static void DoXMPCoreCoverage ( FILE * log ) fprintf ( log, "\n" ); - XMP_DateTime date1 = { 2000, 1, 31, 12, 34, 56, -1, 8, 0, 0 }; - XMP_DateTime date2 = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; + XMP_DateTime date1, date2; + FillDateTime ( &date1, 2000, 1, 31, 12, 34, 56, true, true, true, -1, 8, 0, 0 ); + tmpStr1.erase(); SXMPUtils::ConvertFromDate ( date1, &tmpStr1 ); @@ -1708,8 +1605,7 @@ static void DoXMPCoreCoverage ( FILE * log ) WriteMajorLabel ( log, "Test date/time utilities and special values" ); fprintf ( log, "\n" ); - XMP_DateTime utcNow = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; - XMP_DateTime localNow = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }; + XMP_DateTime utcNow, localNow; SXMPUtils::SetTimeZone ( &utcNow ); fprintf ( log, "SetTimeZone : %d-%02d-%02d %02d:%02d:%02d %d*%02d:%02d %d\n", @@ -1825,25 +1721,25 @@ static void DoXMPCoreCoverage ( FILE * log ) meta2.SetProperty ( kXMP_NS_XMP, "Format", "new Format" ); DumpXMPObj ( log, meta2, "Create 2nd XMP object with new values" ); - SXMPUtils::AppendProperties ( meta2, &meta1 ); + SXMPUtils::ApplyTemplate ( &meta1, meta2, kXMPTemplate_AddNewProperties ); DumpXMPObj ( log, meta1, "Append 2nd to 1st, keeping old values, external only" ); meta2.SetProperty ( kXMP_NS_XMP, "CreatorTool", "newer CreatorTool" ); meta2.SetProperty ( kXMP_NS_XMP, "Nickname", "newer Nickname" ); meta2.SetProperty ( kXMP_NS_XMP, "Format", "newer Format" ); - SXMPUtils::AppendProperties ( meta2, &meta1, kXMPUtil_DoAllProperties ); + SXMPUtils::ApplyTemplate ( &meta1, meta2, kXMPTemplate_AddNewProperties | kXMPTemplate_IncludeInternalProperties ); DumpXMPObj ( log, meta1, "Append 2nd to 1st, keeping old values, internal also" ); meta2.SetProperty ( kXMP_NS_XMP, "CreatorTool", "newest CreatorTool" ); meta2.SetProperty ( kXMP_NS_XMP, "Nickname", "newest Nickname" ); meta2.SetProperty ( kXMP_NS_XMP, "Format", "newest Format" ); - SXMPUtils::AppendProperties ( meta2, &meta1, kXMPUtil_ReplaceOldValues ); + SXMPUtils::ApplyTemplate ( &meta1, meta2, kXMPTemplate_AddNewProperties | kXMPTemplate_ReplaceExistingProperties ); DumpXMPObj ( log, meta1, "Append 2nd to 1st, replacing old values, external only" ); meta2.SetProperty ( kXMP_NS_XMP, "CreatorTool", "final CreatorTool" ); meta2.SetProperty ( kXMP_NS_XMP, "Nickname", "final Nickname" ); meta2.SetProperty ( kXMP_NS_XMP, "Format", "final Format" ); - SXMPUtils::AppendProperties ( meta2, &meta1, (kXMPUtil_ReplaceOldValues | kXMPUtil_DoAllProperties) ); + SXMPUtils::ApplyTemplate ( &meta1, meta2, kXMPTemplate_AddNewProperties | kXMPTemplate_ReplaceExistingProperties | kXMPTemplate_IncludeInternalProperties ); DumpXMPObj ( log, meta1, "Append 2nd to 1st, replacing old values, internal also" ); } diff --git a/samples/source/XMPFilesCoverage.cpp b/samples/source/XMPFilesCoverage.cpp index 84e5e27..759baef 100644 --- a/samples/source/XMPFilesCoverage.cpp +++ b/samples/source/XMPFilesCoverage.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2002 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -141,7 +141,7 @@ static void OpenTestFile ( const char * fileName, XMP_OptionBits rwMode, SXMPMet static const char * charForms[] = { "UTF-8", "unknown char form", "UTF-16BE", "UTF-16LE", "UTF-32BE", "UTF-32LE" }; - XMP_OptionBits smartFlags = rwMode | kXMPFiles_OpenUseSmartHandler | kXMPFiles_OpenCacheTNail; + XMP_OptionBits smartFlags = rwMode | kXMPFiles_OpenUseSmartHandler; XMP_OptionBits scanFlags = rwMode | kXMPFiles_OpenUsePacketScanning; ok = xmpFile->OpenFile ( fileName, kXMP_UnknownFile, smartFlags ); @@ -156,19 +156,6 @@ static void OpenTestFile ( const char * fileName, XMP_OptionBits rwMode, SXMPMet fprintf ( sLogFile, "File info : format = \"%.4s\", handler flags = 0x%X, open flags = 0x%X (%s)\n", &format, handlerFlags, openFlags, (isUpdate ? "update" : "read-only") ); - - XMP_ThumbnailInfo tnail; - ok = xmpFile->GetThumbnail ( &tnail ); - if ( ! ok ) { - fprintf ( sLogFile, "No thumbnail\n" ); - } else { - fprintf ( sLogFile, "Thumbnail info : file format = \"%.4s\", tnail format = %d, tnail size = %d\n", - &tnail.fileFormat, tnail.tnailFormat, tnail.tnailSize ); - fprintf ( sLogFile, " Image width x height, orientation : %d x %d, %d\n", - tnail.fullWidth, tnail.fullHeight, tnail.fullOrientation ); - fprintf ( sLogFile, " TNail width x height, orientation : %d x %d, %d\n", - tnail.tnailWidth, tnail.tnailHeight, tnail.tnailOrientation ); - } ok = xmpFile->GetXMP ( xmpMeta, 0, &xmpPacket ); if ( ! ok ) { @@ -299,7 +286,11 @@ extern "C" int main ( int argc, const char * argv[] ) fprintf ( sLogFile, "## XMPMeta::Initialize failed!\n" ); return -1; } - if ( ! SXMPFiles::Initialize() ) { + XMP_OptionBits options = 0; + #if UNIX_ENV + options |= kXMPFiles_ServerMode; + #endif + if ( ! SXMPFiles::Initialize ( options ) ) { fprintf ( sLogFile, "## SXMPFiles::Initialize failed!\n" ); return -1; } diff --git a/samples/source/XMPIterations.cpp b/samples/source/XMPIterations.cpp new file mode 100644 index 0000000..7a4273a --- /dev/null +++ b/samples/source/XMPIterations.cpp @@ -0,0 +1,321 @@ +// ================================================================================================= +// Copyright 2008 Adobe Systems Incorporated +// All Rights Reserved. +// +// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms +// of the Adobe license agreement accompanying it. +// ================================================================================================= + +/** +* Demonstrates how to use the iteration utility in the XMPCore component to walk through property trees. +*/ + +#include <string> + +// Must be defined to instantiate template classes +#define TXMP_STRING_TYPE std::string + +// Must be defined to give access to XMPFiles +#define XMP_INCLUDE_XMPFILES 1 + +// Ensure XMP templates are instantiated +#include "XMP.incl_cpp" + +// Provide access to the API +#include "XMP.hpp" + +#include <iostream> + +using namespace std; + +// Provide some custom XMP +static const char * rdf = + "<rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'>" + " <rdf:Description rdf:about='' xmlns:xmpTest='http://ns.adobe.com/xmpTest/'>" + "" + " <xmpTest:MySimpleProp rdf:parseType='Resource'>" + " <rdf:value>A Value</rdf:value>" + " <xmpTest:MyQual>Qual Value</xmpTest:MyQual>" + " </xmpTest:MySimpleProp>" + "" + " <xmpTest:MyTopStruct rdf:parseType='Resource'>" + " <xmpTest:MySecondStruct rdf:parseType='Resource'>" + " <xmpTest:MyThirdStruct rdf:parseType='Resource'>" + " <xmpTest:MyThirdStructField>Field Value 3</xmpTest:MyThirdStructField>" + " </xmpTest:MyThirdStruct>" + " <xmpTest:MySecondStructField>Field Value 2</xmpTest:MySecondStructField>" + " </xmpTest:MySecondStruct>" + " <xmpTest:MyTopStructField>Field Value 1</xmpTest:MyTopStructField>" + " </xmpTest:MyTopStruct>" + + " <xmpTest:MyArrayWithNestedArray>" + " <rdf:Bag>" + " <rdf:li>" + " <rdf:Seq>" + " <rdf:li>Item 1</rdf:li>" + " <rdf:li>Item 2</rdf:li>" + " </rdf:Seq>" + " </rdf:li>" + " </rdf:Bag>" + " </xmpTest:MyArrayWithNestedArray>" + + " <xmpTest:MyArrayWithStructures>" + " <rdf:Seq>" + " <rdf:li rdf:parseType='Resource'>" + " <rdf:value>Field Value 1</rdf:value>" + " <xmpTest:FirstQual>Qual Value 1</xmpTest:FirstQual>" + " <xmpTest:SecondQual>Qual Value 2</xmpTest:SecondQual>" + " </rdf:li>" + " <rdf:li rdf:parseType='Resource'>" + " <rdf:value>Field Value 2</rdf:value>" + " <xmpTest:FirstQual>Qual Value 3</xmpTest:FirstQual>" + " <xmpTest:SecondQual>Qual Value 4</xmpTest:SecondQual>" + " </rdf:li>" + " </rdf:Seq>" + " </xmpTest:MyArrayWithStructures>" + "" + " <xmpTest:MyStructureWithArray rdf:parseType='Resource'>" + " <xmpTest:NestedArray>" + " <rdf:Bag>" + " <rdf:li>Item 3</rdf:li>" + " <rdf:li>Item 4</rdf:li>" + " <rdf:li>Item 5</rdf:li>" + " <rdf:li>Item 6</rdf:li>" + " </rdf:Bag>" + " </xmpTest:NestedArray>" + " <xmpTest:NestedArray2>" + " <rdf:Bag>" + " <rdf:li>Item 66</rdf:li>" + " <rdf:li>Item 46</rdf:li>" + " <rdf:li>Item 56</rdf:li>" + " <rdf:li>Item 66</rdf:li>" + " </rdf:Bag>" + " </xmpTest:NestedArray2>" + " </xmpTest:MyStructureWithArray>" + "" + " </rdf:Description>" + "</rdf:RDF>"; + +// The namespace to be used. This will be automatically registered +// when the RDF is parsed. +const XMP_StringPtr kXMP_NS_SDK = "http://ns.adobe.com/xmpTest/"; + +/** +* Reads some metadata from a file and appends some custom XMP to it. Then does several +* iterations, using various iterators. Each iteration is displayed in the console window. +*/ +int main() +{ + if(SXMPMeta::Initialize()) + { + XMP_OptionBits options = 0; + #if UNIX_ENV + options |= kXMPFiles_ServerMode; + #endif + if ( SXMPFiles::Initialize ( options ) ) { + bool ok; + SXMPFiles myFile; + + XMP_OptionBits opts = kXMPFiles_OpenForRead | kXMPFiles_OpenUseSmartHandler; + ok = myFile.OpenFile("../../../testfiles/Image1.jpg", kXMP_UnknownFile, opts); + if(ok) + { + SXMPMeta xmp; + myFile.GetXMP(&xmp); + + // Add some custom metadata to the XMP object + SXMPMeta custXMP(rdf, (XMP_StringLen) strlen(rdf)); + SXMPUtils::ApplyTemplate(&xmp, custXMP, kXMPTemplate_AddNewProperties); + + // Store any details from the iter.Next() call + string schemaNS, propPath, propVal; + + // Only visit the immediate children that are leaf properties of the Dublin Core schema + SXMPIterator dcLeafIter(xmp, kXMP_NS_DC, (kXMP_IterJustChildren | kXMP_IterJustLeafNodes)); + while(dcLeafIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit one property from the XMP Basic schema + SXMPIterator xmpKeywordsIter(xmp, kXMP_NS_XMP, "Keywords", kXMP_IterJustLeafNodes); + while(xmpKeywordsIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit the Dublin Core schema, omit any quailifiers and only + // show the leaf properties + SXMPIterator dcIter(xmp, kXMP_NS_DC, (kXMP_IterOmitQualifiers | kXMP_IterJustLeafNodes)); + while(dcIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit the Dublin Core schema, omit any quailifiers, + // show the leaf properties but only return the leaf name and not the full path + SXMPIterator dcIter2(xmp, kXMP_NS_DC, (kXMP_IterOmitQualifiers | kXMP_IterJustLeafNodes | kXMP_IterJustLeafName)); + while(dcIter2.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Iterate over a single namespace. Show all properties within + // the Photoshop schema + SXMPIterator exifIter(xmp, kXMP_NS_Photoshop); + while(exifIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Just visit the leaf nodes of EXIF properties. That is just + // properties that may have values. + SXMPIterator exifLeafIter(xmp, kXMP_NS_EXIF, kXMP_IterJustLeafNodes); + while(exifLeafIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Iterate over all properties but skip the EXIF schema and skip the custom schema + // and continue visiting nodes + SXMPIterator skipExifIter (xmp); + while(skipExifIter.Next(&schemaNS, &propPath, &propVal)) + { + if(schemaNS == kXMP_NS_EXIF || schemaNS == kXMP_NS_SDK) + { + skipExifIter.Skip(kXMP_IterSkipSubtree); + } + else + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + } + + cout << "----------------------------------" << endl; + + // Iterate over all properties but skip the EXIF schema + // and any remaining siblings of the current node. + SXMPIterator stopAfterExifIter ( xmp ); + while(stopAfterExifIter.Next(&schemaNS, &propPath, &propVal)) + { + if(schemaNS == kXMP_NS_EXIF || schemaNS == kXMP_NS_SDK) + { + stopAfterExifIter.Skip(kXMP_IterSkipSiblings); + } + else + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + } + + cout << "----------------------------------" << endl; + + ////////////////////////////////////////////////////////////////////////////////////// + + // Iterate over the custom XMP + + // Visit the immediate children of this node. + // No qualifiers are visisted as they are below the property being visisted. + SXMPIterator justChildrenIter(xmp, kXMP_NS_SDK, kXMP_IterJustChildren); + while(justChildrenIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit the immediate children of this node but only those that may have values. + // No qualifiers are visisted as they are below the property being visisted. + SXMPIterator justChildrenAndLeafIter(xmp, kXMP_NS_SDK, (kXMP_IterJustChildren | kXMP_IterJustLeafNodes)); + while(justChildrenAndLeafIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit the leaf nodes of TopStructProperty + SXMPIterator myTopStructIter(xmp, kXMP_NS_SDK, "MyTopStruct", kXMP_IterJustLeafNodes); + while(myTopStructIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit the leaf nodes of the TopStructProperty but only return the names for + // the leaf components and not the full path + SXMPIterator xmyTopStructIterShortNames(xmp, kXMP_NS_SDK, "MyTopStruct", (kXMP_IterJustLeafNodes | kXMP_IterJustLeafName)); + while(xmyTopStructIterShortNames.Next(&schemaNS, &propPath, &propVal)) + { + cout << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit a property and all of the qualifiers + SXMPIterator iterArrayProp (xmp, kXMP_NS_SDK, "ArrayWithStructures", kXMP_IterJustLeafNodes ); + while(iterArrayProp.Next(&schemaNS, &propPath, &propVal)) + { + cout << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Visit a property and omit all of the qualifiers + SXMPIterator iterArrayPropNoQual (xmp, kXMP_NS_SDK, "ArrayWithStructures", (kXMP_IterJustLeafNodes | kXMP_IterOmitQualifiers)); + while(iterArrayPropNoQual.Next(&schemaNS, &propPath, &propVal)) + { + cout << propPath << " = " << propVal << endl; + } + + cout << "----------------------------------" << endl; + + // Skip a subtree and continue onwards. Once 'Item 4' is found then the we can skip all of the + // siblings of the current node. If the the current node were a top level node the iteration + // would be complete as all siblings would be skipped. However, when 'Item 4' is found the current + // node is not at the top level so there are other nodes further up the tree that still need to be + // visited. + SXMPIterator skipIter (xmp, kXMP_NS_SDK, (kXMP_IterJustLeafNodes | kXMP_IterOmitQualifiers | kXMP_IterJustLeafName)); + while(skipIter.Next(&schemaNS, &propPath, &propVal)) + { + if(propVal == "Item 4") + { + skipIter.Skip(kXMP_IterSkipSiblings); + } + else + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + } + + /* + // Visit all properties and qualifiers + SXMPIterator allPropsIter(xmp); + while(allPropsIter.Next(&schemaNS, &propPath, &propVal)) + { + cout << schemaNS << " " << propPath << " = " << propVal << endl; + } + */ + } + } + } + + SXMPFiles::Terminate(); + SXMPMeta::Terminate(); + + return 0; +} + diff --git a/samples/source/XMPScanner.cpp b/samples/source/XMPScanner.cpp deleted file mode 100644 index 63f03d0..0000000 --- a/samples/source/XMPScanner.cpp +++ /dev/null @@ -1,1480 +0,0 @@ -// ================================================================================================= -// Copyright 2002-2007 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// -// Adobe patent application tracking #P435, entitled 'Unique markers to simplify embedding data of -// one format in a file with a different format', inventors: Sean Parent, Greg Gilley. -// ================================================================================================= - -#if WIN32 - #pragma warning ( disable : 4127 ) // conditional expression is constant - #pragma warning ( disable : 4510 ) // default constructor could not be generated - #pragma warning ( disable : 4610 ) // user defined constructor required - #pragma warning ( disable : 4786 ) // debugger can't handle long symbol names -#endif - - -#include "XMPScanner.hpp" - -#include <cassert> -#include <string> -#include <cstdlib> - -#if DEBUG - #include <iostream> - #include <iomanip> - #include <fstream> -#endif - - -#ifndef UseStringPushBack // VC++ 6.x does not provide push_back for strings! - #define UseStringPushBack 0 -#endif - - -using namespace std; - - -// *** Consider Boyer-Moore style search for "<?xpacket begin=". It isn't an obvious win, the -// *** additional code might be slower than scanning every character. Especially if we will -// *** read every cache line anyway. - - -// ================================================================================================= -// ================================================================================================= -// class PacketMachine -// =================== -// -// This is the packet recognizer state machine. The top of the machine is FindNextPacket, this -// calls the specific state components and handles transitions. The states are described by an -// array of RecognizerInfo records, indexed by the RecognizerKind enumeration. Each RecognizerInfo -// record has a function that does that state's work, the success and failure transition states, -// and a string literal that is passed to the state function. The literal lets a common MatchChar -// or MatchString function be used in several places. -// -// The state functions are responsible for consuming input to recognize their particular state. -// This includes intervening nulls for 16 and 32 bit character forms. For the simplicity, things -// are treated as essentially little endian and the nulls are not actually checked. The opening -// '<' is found with a byte-by-byte search, then the number of bytes per character is determined -// by counting the following nulls. From then on, consuming a character means incrementing the -// buffer pointer by the number of bytes per character. Thus the buffer pointer only points to -// the "real" bytes. This also means that the pointer can go off the end of the buffer by a -// variable amount. The amount of overrun is saved so that the pointer can be positioned at the -// right byte to start the next buffer. -// -// The state functions return a TriState value, eTriYes means the pattern was found, eTriNo means -// the pattern was definitely not found, eTriMaybe means that the end of the buffer was reached -// while working through the pattern. -// -// When eTriYes is returned, the fBufferPtr data member is left pointing to the "real" byte -// following the last actual byte. Which might not be addressable memory! This also means that -// a state function can be entered with nothing available in the buffer. When eTriNo is returned, -// the fBufferPtr data member is left pointing to the byte that caused the failure. The state -// machine starts over from the failure byte. -// -// The state functions must preserve their internal micro-state before returning eTriMaybe, and -// resume processing when called with the next buffer. The fPosition data member is used to denote -// how many actual characters have been consumed. The fNullCount data member is used to denote how -// many nulls are left before the next actual character. - - -// ================================================================================================= -// PacketMachine -// ============= - -XMPScanner::PacketMachine::PacketMachine ( XMP_Int64 bufferOffset, const void * bufferOrigin, XMP_Int64 bufferLength ) : - - // Public members - fPacketStart ( 0 ), - fPacketLength ( 0 ), - fBytesAttr ( -1 ), - fCharForm ( eChar8Bit ), - fAccess ( ' ' ), - fBogusPacket ( false ), - - // Private members - fBufferOffset ( bufferOffset ), - fBufferOrigin ( (const char *) bufferOrigin ), - fBufferPtr ( fBufferOrigin ), - fBufferLimit ( fBufferOrigin + bufferLength ), - fRecognizer ( eLeadInRecognizer ), - fPosition ( 0 ), - fBytesPerChar ( 1 ), - fBufferOverrun ( 0 ), - fQuoteChar ( ' ' ) - -{ - /* - REVIEW NOTES : Should the buffer stuff be in a class? - */ - - assert ( bufferOrigin != NULL ); - assert ( bufferLength != 0 ); - -} // PacketMachine - - -// ================================================================================================= -// ~PacketMachine -// ============== - -XMPScanner::PacketMachine::~PacketMachine () -{ - - // An empty placeholder. - -} // ~PacketMachine - - -// ================================================================================================= -// AssociateBuffer -// =============== - -void -XMPScanner::PacketMachine::AssociateBuffer ( XMP_Int64 bufferOffset, const void * bufferOrigin, XMP_Int64 bufferLength ) -{ - - fBufferOffset = bufferOffset; - fBufferOrigin = (const char *) bufferOrigin; - fBufferPtr = fBufferOrigin + fBufferOverrun; - fBufferLimit = fBufferOrigin + bufferLength; - -} // AssociateBuffer - - -// ================================================================================================= -// ResetMachine -// ============ - -void -XMPScanner::PacketMachine::ResetMachine () -{ - - fRecognizer = eLeadInRecognizer; - fPosition = 0; - fBufferOverrun = 0; - fCharForm = eChar8Bit; - fBytesPerChar = 1; - fAccess = ' '; - fBytesAttr = -1; - fBogusPacket = false; - - fAttrName.erase ( fAttrName.begin(), fAttrName.end() ); - fAttrValue.erase ( fAttrValue.begin(), fAttrValue.end() ); - fEncodingAttr.erase ( fEncodingAttr.begin(), fEncodingAttr.end() ); - -} // ResetMachine - - -// ================================================================================================= -// FindLessThan -// ============ - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::FindLessThan ( PacketMachine * ths, const char * which ) -{ - - if ( *which == 'H' ) { - - // -------------------------------------------------------------------------------- - // We're looking for the '<' of the header. If we fail there is no packet in this - // part of the input, so return eTriNo. - - ths->fCharForm = eChar8Bit; // We might have just failed from a bogus 16 or 32 bit case. - ths->fBytesPerChar = 1; - - while ( ths->fBufferPtr < ths->fBufferLimit ) { // Don't skip nulls for the header's '<'! - if ( *ths->fBufferPtr == '<' ) break; - ths->fBufferPtr++; - } - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriNo; - ths->fBufferPtr++; - return eTriYes; - - } else { - - // -------------------------------------------------------------------------------- - // We're looking for the '<' of the trailer. We're already inside the packet body, - // looking for the trailer. So here if we fail we must return eTriMaybe so that we - // keep looking for the trailer in the next buffer. - - const int bytesPerChar = ths->fBytesPerChar; - - while ( ths->fBufferPtr < ths->fBufferLimit ) { - if ( *ths->fBufferPtr == '<' ) break; - ths->fBufferPtr += bytesPerChar; - } - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - - } - -} // FindLessThan - - -// ================================================================================================= -// MatchString -// =========== - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::MatchString ( PacketMachine * ths, const char * literal ) -{ - const int bytesPerChar = ths->fBytesPerChar; - const char * litPtr = literal + ths->fPosition; - const int charsToGo = strlen ( literal ) - ths->fPosition; - int charsDone = 0; - - while ( (charsDone < charsToGo) && (ths->fBufferPtr < ths->fBufferLimit) ) { - if ( *litPtr != *ths->fBufferPtr ) return eTriNo; - charsDone++; - litPtr++; - ths->fBufferPtr += bytesPerChar; - } - - if ( charsDone == charsToGo ) return eTriYes; - ths->fPosition += charsDone; - return eTriMaybe; - -} // MatchString - - -// ================================================================================================= -// MatchChar -// ========= - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::MatchChar ( PacketMachine * ths, const char * literal ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - if ( currChar != *literal ) return eTriNo; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - -} // MatchChar - - -// ================================================================================================= -// MatchOpenQuote -// ============== - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::MatchOpenQuote ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - if ( (currChar != '\'') && (currChar != '"') ) return eTriNo; - ths->fQuoteChar = currChar; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - -} // MatchOpenQuote - - -// ================================================================================================= -// MatchCloseQuote -// =============== - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::MatchCloseQuote ( PacketMachine * ths, const char * /* unused */ ) -{ - - return MatchChar ( ths, &ths->fQuoteChar ); - -} // MatchCloseQuote - - -// ================================================================================================= -// CaptureAttrName -// =============== - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::CaptureAttrName ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - char currChar; - - if ( ths->fPosition == 0 ) { // Get the first character in the name. - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - currChar = *ths->fBufferPtr; - if ( ths->fAttrName.size() == 0 ) { - if ( ! ( ( ('a' <= currChar) && (currChar <= 'z') ) || - ( ('A' <= currChar) && (currChar <= 'Z') ) || - (currChar == '_') || (currChar == ':') ) ) { - return eTriNo; - } - } - - ths->fAttrName.erase ( ths->fAttrName.begin(), ths->fAttrName.end() ); - #if UseStringPushBack - ths->fAttrName.push_back ( currChar ); - #else - ths->fAttrName.insert ( ths->fAttrName.end(), currChar ); - #endif - ths->fBufferPtr += bytesPerChar; - - } - - while ( ths->fBufferPtr < ths->fBufferLimit ) { // Get the remainder of the name. - - currChar = *ths->fBufferPtr; - if ( ! ( ( ('a' <= currChar) && (currChar <= 'z') ) || - ( ('A' <= currChar) && (currChar <= 'Z') ) || - ( ('0' <= currChar) && (currChar <= '9') ) || - (currChar == '-') || (currChar == '.') || (currChar == '_') || (currChar == ':') ) ) { - break; - } - - #if UseStringPushBack - ths->fAttrName.push_back ( currChar ); - #else - ths->fAttrName.insert ( ths->fAttrName.end(), currChar ); - #endif - ths->fBufferPtr += bytesPerChar; - - } - - if ( ths->fBufferPtr < ths->fBufferLimit ) return eTriYes; - ths->fPosition = ths->fAttrName.size(); // The name might span into the next buffer. - return eTriMaybe; - -} // CaptureAttrName - - -// ================================================================================================= -// CaptureAttrValue -// ================ -// -// Recognize the equal sign and the quoted string value, capture the value along the way. - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::CaptureAttrValue ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - char currChar = 0; - TriState result = eTriMaybe; - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - switch ( ths->fPosition ) { - - case 0 : // The name should haved ended at the '=', nulls already skipped. - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - if ( *ths->fBufferPtr != '=' ) return eTriNo; - ths->fBufferPtr += bytesPerChar; - ths->fPosition = 1; - // fall through OK because MatchOpenQuote will check the buffer limit and nulls ... - - case 1 : // Look for the open quote. - - result = MatchOpenQuote ( ths, NULL ); - if ( result != eTriYes ) return result; - ths->fPosition = 2; - // fall through OK because the buffer limit and nulls are checked below ... - - default : // Look for the close quote, capturing the value along the way. - - assert ( ths->fPosition == 2 ); - - const char quoteChar = ths->fQuoteChar; - - while ( ths->fBufferPtr < ths->fBufferLimit ) { - currChar = *ths->fBufferPtr; - if ( currChar == quoteChar ) break; - #if UseStringPushBack - ths->fAttrValue.push_back ( currChar ); - #else - ths->fAttrValue.insert ( ths->fAttrValue.end(), currChar ); - #endif - ths->fBufferPtr += bytesPerChar; - } - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - assert ( currChar == quoteChar ); - ths->fBufferPtr += bytesPerChar; // Advance past the closing quote. - return eTriYes; - - } - -} // CaptureAttrValue - - -// ================================================================================================= -// RecordStart -// =========== -// -// Note that this routine looks at bytes, not logical characters. It has to figure out how many -// bytes per character there are so that the other recognizers can skip intervening nulls. - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::RecordStart ( PacketMachine * ths, const char * /* unused */ ) -{ - - while ( true ) { - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currByte = *ths->fBufferPtr; - - switch ( ths->fPosition ) { - - case 0 : // Record the length. - assert ( ths->fCharForm == eChar8Bit ); - assert ( ths->fBytesPerChar == 1 ); - ths->fPacketStart = ths->fBufferOffset + ((ths->fBufferPtr - 1) - ths->fBufferOrigin); - ths->fPacketLength = 0; - ths->fPosition = 1; - // ! OK to fall through here, we didn't consume a byte in this step. - - case 1 : // Look for the first null byte. - if ( currByte != 0 ) return eTriYes; // No nulls found. - ths->fCharForm = eChar16BitBig; // Assume 16 bit big endian for now. - ths->fBytesPerChar = 2; - ths->fBufferPtr++; - ths->fPosition = 2; - break; // ! Don't fall through, have to check for the end of the buffer between each byte. - - case 2 : // One null was found, look for a second. - if ( currByte != 0 ) return eTriYes; // Just one null found. - ths->fBufferPtr++; - ths->fPosition = 3; - break; - - case 3 : // Two nulls were found, look for a third. - if ( currByte != 0 ) return eTriNo; // Just two nulls is not valid. - ths->fCharForm = eChar32BitBig; // Assume 32 bit big endian for now. - ths->fBytesPerChar = 4; - ths->fBufferPtr++; - return eTriYes; - break; - - } - - } - -} // RecordStart - - -// ================================================================================================= -// RecognizeBOM -// ============ -// -// Recognizing the byte order marker is a surprisingly messy thing to do. It can't be done by the -// normal string matcher, there are no intervening nulls. There are 4 transitions after the opening -// quote, the closing quote or one of the three encodings. For the actual BOM there are then 1 or 2 -// following bytes that depend on which of the encodings we're in. Not to mention that the buffer -// might end at any point. -// -// The intervening null count done earlier determined 8, 16, or 32 bits per character, but not the -// big or little endian nature for the 16/32 bit cases. The BOM must be present for the 16 and 32 -// bit cases in order to determine the endian mode. There are six possible byte sequences for the -// quoted BOM string, ignoring the differences for quoting with ''' versus '"'. -// -// Keep in mind that for the 16 and 32 bit cases there will be nulls for the quote. In the table -// below the symbol <quote> means just the one byte containing the ''' or '"'. The nulls for the -// quote character are explicitly shown. -// -// <quote> <quote> - 1: No BOM, this must be an 8 bit case. -// <quote> \xEF \xBB \xBF <quote> - 1.12-13: The 8 bit form. -// -// <quote> \xFE \xFF \x00 <quote> - 1.22-23: The 16 bit, big endian form -// <quote> \x00 \xFF \xFE <quote> - 1.32-33: The 16 bit, little endian form. -// -// <quote> \x00 \x00 \xFE \xFF \x00 \x00 \x00 <quote> - 1.32.43-45.56-57: The 32 bit, big endian form. -// <quote> \x00 \x00 \x00 \xFF \xFE \x00 \x00 <quote> - 1.32.43.54-57: The 32 bit, little endian form. - -enum { - eBOM_8_1 = 0xEF, - eBOM_8_2 = 0xBB, - eBOM_8_3 = 0xBF, - eBOM_Big_1 = 0xFE, - eBOM_Big_2 = 0xFF, - eBOM_Little_1 = eBOM_Big_2, - eBOM_Little_2 = eBOM_Big_1 -}; - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::RecognizeBOM ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - while ( true ) { // Handle one character at a time, the micro-state (fPosition) changes for each. - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const unsigned char currChar = *ths->fBufferPtr; // ! The BOM bytes look like integers bigger than 127. - - switch ( ths->fPosition ) { - - case 0 : // Look for the opening quote. - if ( (currChar != '\'') && (currChar != '"') ) return eTriNo; - ths->fQuoteChar = currChar; - ths->fBufferPtr++; - ths->fPosition = 1; - break; // ! Don't fall through, have to check for the end of the buffer between each byte. - - case 1 : // Look at the byte immediately following the opening quote. - if ( currChar == ths->fQuoteChar ) { // Closing quote, no BOM character, must be 8 bit. - if ( ths->fCharForm != eChar8Bit ) return eTriNo; - ths->fBufferPtr += bytesPerChar; // Skip the nulls after the closing quote. - return eTriYes; - } else if ( currChar == eBOM_8_1 ) { // Start of the 8 bit form. - if ( ths->fCharForm != eChar8Bit ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 12; - } else if ( currChar == eBOM_Big_1 ) { // Start of the 16 bit big endian form. - if ( ths->fCharForm != eChar16BitBig ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 22; - } else if ( currChar == 0 ) { // Start of the 16 bit little endian or either 32 bit form. - if ( ths->fCharForm == eChar8Bit ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 32; - } else { - return eTriNo; - } - break; - - case 12 : // Look for the second byte of the 8 bit form. - if ( currChar != eBOM_8_2 ) return eTriNo; - ths->fPosition = 13; - ths->fBufferPtr++; - break; - - case 13 : // Look for the third byte of the 8 bit form. - if ( currChar != eBOM_8_3 ) return eTriNo; - ths->fPosition = 99; - ths->fBufferPtr++; - break; - - case 22 : // Look for the second byte of the 16 bit big endian form. - if ( currChar != eBOM_Big_2 ) return eTriNo; - ths->fPosition = 23; - ths->fBufferPtr++; - break; - - case 23 : // Look for the null before the closing quote of the 16 bit big endian form. - if ( currChar != 0 ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 99; - break; - - case 32 : // Look at the second byte of the 16 bit little endian or either 32 bit form. - if ( currChar == eBOM_Little_1 ) { - ths->fPosition = 33; - } else if ( currChar == 0 ) { - ths->fPosition = 43; - } else { - return eTriNo; - } - ths->fBufferPtr++; - break; - - case 33 : // Look for the third byte of the 16 bit little endian form. - if ( ths->fCharForm != eChar16BitBig ) return eTriNo; // Null count before assumed big endian. - if ( currChar != eBOM_Little_2 ) return eTriNo; - ths->fCharForm = eChar16BitLittle; - ths->fPosition = 99; - ths->fBufferPtr++; - break; - - case 43 : // Look at the third byte of either 32 bit form. - if ( ths->fCharForm != eChar32BitBig ) return eTriNo; // Null count before assumed big endian. - if ( currChar == eBOM_Big_1 ) { - ths->fPosition = 44; - } else if ( currChar == 0 ) { - ths->fPosition = 54; - } else { - return eTriNo; - } - ths->fBufferPtr++; - break; - - case 44 : // Look for the fourth byte of the 32 bit big endian form. - if ( currChar != eBOM_Big_2 ) return eTriNo; - ths->fPosition = 45; - ths->fBufferPtr++; - break; - - case 45 : // Look for the first null before the closing quote of the 32 bit big endian form. - if ( currChar != 0 ) return eTriNo; - ths->fPosition = 56; - ths->fBufferPtr++; - break; - - case 54 : // Look for the fourth byte of the 32 bit little endian form. - ths->fCharForm = eChar32BitLittle; - if ( currChar != eBOM_Little_1 ) return eTriNo; - ths->fPosition = 55; - ths->fBufferPtr++; - break; - - case 55 : // Look for the fifth byte of the 32 bit little endian form. - if ( currChar != eBOM_Little_2 ) return eTriNo; - ths->fPosition = 56; - ths->fBufferPtr++; - break; - - case 56 : // Look for the next to last null before the closing quote of the 32 bit forms. - if ( currChar != 0 ) return eTriNo; - ths->fPosition = 57; - ths->fBufferPtr++; - break; - - case 57 : // Look for the last null before the closing quote of the 32 bit forms. - if ( currChar != 0 ) return eTriNo; - ths->fPosition = 99; - ths->fBufferPtr++; - break; - - default : // Look for the closing quote. - assert ( ths->fPosition == 99 ); - if ( currChar != ths->fQuoteChar ) return eTriNo; - ths->fBufferPtr += bytesPerChar; // Skip the nulls after the closing quote. - return eTriYes; - break; - - } - - } - -} // RecognizeBOM - - -// ================================================================================================= -// RecordHeadAttr -// ============== - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::RecordHeadAttr ( PacketMachine * ths, const char * /* unused */ ) -{ - - if ( ths->fAttrName == "encoding" ) { - - assert ( ths->fEncodingAttr.empty() ); - ths->fEncodingAttr = ths->fAttrValue; - - } else if ( ths->fAttrName == "bytes" ) { - - long value = 0; - int count = ths->fAttrValue.size(); - int i; - - assert ( ths->fBytesAttr == -1 ); - - if ( count > 0 ) { // Allow bytes='' to be the same as no bytes attribute. - - for ( i = 0; i < count; i++ ) { - const char currChar = ths->fAttrValue[i]; - if ( ('0' <= currChar) && (currChar <= '9') ) { - value = (value * 10) + (currChar - '0'); - } else { - ths->fBogusPacket = true; - value = -1; - break; - } - } - ths->fBytesAttr = value; - - if ( CharFormIs16Bit ( ths->fCharForm ) ) { - if ( (ths->fBytesAttr & 1) != 0 ) ths->fBogusPacket = true; - } else if ( CharFormIs32Bit ( ths->fCharForm ) ) { - if ( (ths->fBytesAttr & 3) != 0 ) ths->fBogusPacket = true; - } - - } - - } - - ths->fAttrName.erase ( ths->fAttrName.begin(), ths->fAttrName.end() ); - ths->fAttrValue.erase ( ths->fAttrValue.begin(), ths->fAttrValue.end() ); - - return eTriYes; - -} // RecordHeadAttr - - -// ================================================================================================= -// CaptureAccess -// ============= - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::CaptureAccess ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - while ( true ) { - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - - switch ( ths->fPosition ) { - - case 0 : // Look for the opening quote. - if ( (currChar != '\'') && (currChar != '"') ) return eTriNo; - ths->fQuoteChar = currChar; - ths->fBufferPtr += bytesPerChar; - ths->fPosition = 1; - break; // ! Don't fall through, have to check for the end of the buffer between each byte. - - case 1 : // Look for the 'r' or 'w'. - if ( (currChar != 'r') && (currChar != 'w') ) return eTriNo; - ths->fAccess = currChar; - ths->fBufferPtr += bytesPerChar; - ths->fPosition = 2; - break; - - default : // Look for the closing quote. - assert ( ths->fPosition == 2 ); - if ( currChar != ths->fQuoteChar ) return eTriNo; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - break; - - } - - } - -} // CaptureAccess - - -// ================================================================================================= -// RecordTailAttr -// ============== - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::RecordTailAttr ( PacketMachine * ths, const char * /* unused */ ) -{ - - // There are no known "general" attributes for the packet trailer. - - ths->fAttrName.erase ( ths->fAttrName.begin(), ths->fAttrName.end() ); - ths->fAttrValue.erase ( ths->fAttrValue.begin(), ths->fAttrValue.end() ); - - return eTriYes; - - -} // RecordTailAttr - - -// ================================================================================================= -// CheckPacketEnd -// ============== -// -// Check for trailing padding and record the packet length. We have trailing padding if the bytes -// attribute is present and has a value greater than the current length. - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::CheckPacketEnd ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - if ( ths->fPosition == 0 ) { // First call, decide if there is trailing padding. - - const XMP_Int64 currLen64 = (ths->fBufferOffset + (ths->fBufferPtr - ths->fBufferOrigin)) - ths->fPacketStart; - if ( currLen64 > 0x7FFFFFFF ) throw std::runtime_error ( "Packet length exceeds 2GB-1" ); - const XMP_Int32 currLength = (XMP_Int32)currLen64; - - if ( (ths->fBytesAttr != -1) && (ths->fBytesAttr != currLength) ) { - if ( ths->fBytesAttr < currLength ) { - ths->fBogusPacket = true; // The bytes attribute value is too small. - } else { - ths->fPosition = ths->fBytesAttr - currLength; - if ( (ths->fPosition % ths->fBytesPerChar) != 0 ) { - ths->fBogusPacket = true; // The padding is not a multiple of the character size. - ths->fPosition = (ths->fPosition / ths->fBytesPerChar) * ths->fBytesPerChar; - } - } - } - - } - - while ( ths->fPosition > 0 ) { - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - - if ( (currChar != ' ') && (currChar != '\t') && (currChar != '\n') && (currChar != '\r') ) { - ths->fBogusPacket = true; // The padding is not whitespace. - break; // Stop the packet here. - } - - ths->fPosition -= bytesPerChar; - ths->fBufferPtr += bytesPerChar; - - } - - const XMP_Int64 currLen64 = (ths->fBufferOffset + (ths->fBufferPtr - ths->fBufferOrigin)) - ths->fPacketStart; - if ( currLen64 > 0x7FFFFFFF ) throw std::runtime_error ( "Packet length exceeds 2GB-1" ); - ths->fPacketLength = (XMP_Int32)currLen64; - return eTriYes; - -} // CheckPacketEnd - - -// ================================================================================================= -// CheckFinalNulls -// =============== -// -// Do some special case processing for little endian characters. We have to make sure the presumed -// nulls after the last character actually exist, i.e. that the stream does not end too soon. Note -// that the prior character scanning has moved the buffer pointer to the address following the last -// byte of the last character. I.e. we're already past the presumed nulls, so we can't check their -// content. All we can do is verify that the stream does not end too soon. -// -// Doing this check is simple yet subtle. If we're still in the current buffer then the trailing -// bytes obviously exist. If we're exactly at the end of the buffer then the bytes also exist. -// The only question is when we're actually past this buffer, partly into the next buffer. This is -// when "ths->fBufferPtr > ths->fBufferLimit" on entry. For that case we have to wait until we've -// actually seen enough extra bytes of input. -// -// Since the normal buffer processing is already adjusting for this partial character overrun, all -// that needs to be done here is wait until "ths->fBufferPtr <= ths->fBufferLimit" on entry. In -// other words, if we're presently too far, ths->fBufferPtr will be adjusted by the amount of the -// overflow the next time XMPScanner::Scan is called. This might still be too far, so just keep -// waiting for enough data to pass by. -// -// Note that there is a corresponding special case for big endian characters, we must decrement the -// starting offset by the number of leading nulls. But we don't do that here, we leave it to the -// outer code. This is because the leading nulls might have been at the exact end of a previous -// buffer, in which case we have to also decrement the length of that raw data snip. - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::CheckFinalNulls ( PacketMachine * ths, const char * /* unused */ ) -{ - - if ( (ths->fCharForm != eChar8Bit) && CharFormIsLittleEndian ( ths->fCharForm ) ) { - if ( ths->fBufferPtr > ths->fBufferLimit ) return eTriMaybe; - } - - return eTriYes; - -} // CheckFinalNulls - - -// ================================================================================================= -// SetNextRecognizer -// ================= - -void -XMPScanner::PacketMachine::SetNextRecognizer ( RecognizerKind nextRecognizer ) -{ - - fRecognizer = nextRecognizer; - fPosition = 0; - -} // SetNextRecognizer - - -// ================================================================================================= -// FindNextPacket -// ============== - -// *** When we start validating intervening nulls for 2 and 4 bytes characters, throw an exception -// *** for errors. Don't return eTriNo, that might skip at an optional point. - -XMPScanner::PacketMachine::TriState -XMPScanner::PacketMachine::FindNextPacket () -{ - - TriState status; - - #define kPacketHead "?xpacket begin=" - #define kPacketID "W5M0MpCehiHzreSzNTczkc9d" - #define kPacketTail "?xpacket end=" - - static const RecognizerInfo recognizerTable [eRecognizerCount] = { // ! Would be safer to assign these explicitly. - - // proc successNext failureNext literal - - { NULL, eFailureRecognizer, eFailureRecognizer, NULL}, // eFailureRecognizer - { NULL, eSuccessRecognizer, eSuccessRecognizer, NULL}, // eSuccessRecognizer - - { FindLessThan, eHeadStartRecorder, eFailureRecognizer, "H" }, // eLeadInRecognizer - { RecordStart, eHeadStartRecognizer, eLeadInRecognizer, NULL }, // eHeadStartRecorder - { MatchString, eBOMRecognizer, eLeadInRecognizer, kPacketHead }, // eHeadStartRecognizer - - { RecognizeBOM, eIDTagRecognizer, eLeadInRecognizer, NULL }, // eBOMRecognizer - - { MatchString, eIDOpenRecognizer, eLeadInRecognizer, " id=" }, // eIDTagRecognizer - { MatchOpenQuote, eIDValueRecognizer, eLeadInRecognizer, NULL }, // eIDOpenRecognizer - { MatchString, eIDCloseRecognizer, eLeadInRecognizer, kPacketID }, // eIDValueRecognizer - { MatchCloseQuote, eAttrSpaceRecognizer_1, eLeadInRecognizer, NULL }, // eIDCloseRecognizer - - { MatchChar, eAttrNameRecognizer_1, eHeadEndRecognizer, " " }, // eAttrSpaceRecognizer_1 - { CaptureAttrName, eAttrValueRecognizer_1, eLeadInRecognizer, NULL }, // eAttrNameRecognizer_1 - { CaptureAttrValue, eAttrValueRecorder_1, eLeadInRecognizer, NULL }, // eAttrValueRecognizer_1 - { RecordHeadAttr, eAttrSpaceRecognizer_1, eLeadInRecognizer, NULL }, // eAttrValueRecorder_1 - - { MatchString, eBodyRecognizer, eLeadInRecognizer, "?>" }, // eHeadEndRecognizer - - { FindLessThan, eTailStartRecognizer, eBodyRecognizer, "T"}, // eBodyRecognizer - - { MatchString, eAccessValueRecognizer, eBodyRecognizer, kPacketTail }, // eTailStartRecognizer - { CaptureAccess, eAttrSpaceRecognizer_2, eBodyRecognizer, NULL }, // eAccessValueRecognizer - - { MatchChar, eAttrNameRecognizer_2, eTailEndRecognizer, " " }, // eAttrSpaceRecognizer_2 - { CaptureAttrName, eAttrValueRecognizer_2, eBodyRecognizer, NULL }, // eAttrNameRecognizer_2 - { CaptureAttrValue, eAttrValueRecorder_2, eBodyRecognizer, NULL }, // eAttrValueRecognizer_2 - { RecordTailAttr, eAttrSpaceRecognizer_2, eBodyRecognizer, NULL }, // eAttrValueRecorder_2 - - { MatchString, ePacketEndRecognizer, eBodyRecognizer, "?>" }, // eTailEndRecognizer - { CheckPacketEnd, eCloseOutRecognizer, eBodyRecognizer, "" }, // ePacketEndRecognizer - { CheckFinalNulls, eSuccessRecognizer, eBodyRecognizer, "" } // eCloseOutRecognizer - - }; - - while ( true ) { - - switch ( fRecognizer ) { - - case eFailureRecognizer : - return eTriNo; - - case eSuccessRecognizer : - return eTriYes; - - default : - - // ------------------------------------------------------------------- - // For everything else, the normal cases, use the state machine table. - - const RecognizerInfo * thisState = &recognizerTable [fRecognizer]; - - status = thisState->proc ( this, thisState->literal ); - - switch ( status ) { - - case eTriNo : - SetNextRecognizer ( thisState->failureNext ); - continue; - - case eTriYes : - SetNextRecognizer ( thisState->successNext ); - continue; - - case eTriMaybe : - fBufferOverrun = (unsigned char)(fBufferPtr - fBufferLimit); - return eTriMaybe; // Keep this recognizer intact, to be resumed later. - - } - - } // switch ( fRecognizer ) { ... - - } // while ( true ) { ... - -} // FindNextPacket - - -// ================================================================================================= -// ================================================================================================= -// class InternalSnip -// ================== - - -// ================================================================================================= -// InternalSnip -// ============ - -XMPScanner::InternalSnip::InternalSnip ( XMP_Int64 offset, XMP_Int64 length ) -{ - - fInfo.fOffset = offset; - fInfo.fLength = length; - -} // InternalSnip - - -// ================================================================================================= -// InternalSnip -// ============ - -XMPScanner::InternalSnip::InternalSnip ( const InternalSnip & rhs ) : - fInfo ( rhs.fInfo ), - fMachine ( NULL ) -{ - - assert ( rhs.fMachine.get() == NULL ); // Don't copy a snip with a machine. - assert ( (rhs.fInfo.fEncodingAttr == 0) || (*rhs.fInfo.fEncodingAttr == 0) ); // Don't copy a snip with an encoding. - -} // InternalSnip - - -// ================================================================================================= -// ~InternalSnip -// ============= - -XMPScanner::InternalSnip::~InternalSnip () -{ -} // ~InternalSnip - - - -// ================================================================================================= -// ================================================================================================= -// class XMPScanner -// ================ - - -// ================================================================================================= -// DumpSnipList -// ============ - -#if DEBUG - -static const char * snipStateName [6] = { "not-seen", "pending", "raw-data", "good-packet", "partial", "bad-packet" }; - -void -XMPScanner::DumpSnipList ( const char * title ) -{ - InternalSnipIterator currPos = fInternalSnips.begin(); - InternalSnipIterator endPos = fInternalSnips.end(); - - cout << endl << title << " snip list: " << fInternalSnips.size() << endl; - - for ( ; currPos != endPos; ++currPos ) { - SnipInfo * currSnip = &currPos->fInfo; - cout << '\t' << currSnip << ' ' << snipStateName[currSnip->fState] << ' ' - << currSnip->fOffset << ".." << (currSnip->fOffset + currSnip->fLength - 1) - << ' ' << currSnip->fLength << ' ' << endl; - } -} // DumpSnipList - -#endif - - -// ================================================================================================= -// PrevSnip and NextSnip -// ===================== - -XMPScanner::InternalSnipIterator -XMPScanner::PrevSnip ( InternalSnipIterator snipPos ) -{ - - InternalSnipIterator prev = snipPos; - return --prev; - -} // PrevSnip - -XMPScanner::InternalSnipIterator -XMPScanner::NextSnip ( InternalSnipIterator snipPos ) -{ - - InternalSnipIterator next = snipPos; - return ++next; - -} // NextSnip - - -// ================================================================================================= -// XMPScanner -// ========== -// -// Initialize the scanner object with one "not seen" snip covering the whole stream. - -XMPScanner::XMPScanner ( XMP_Int64 streamLength ) : - - fStreamLength ( streamLength ) - -{ - InternalSnip rootSnip ( 0, streamLength ); - - if ( streamLength > 0 ) fInternalSnips.push_front ( rootSnip ); // Be nice for empty files. - // DumpSnipList ( "New XMPScanner" ); - -} // XMPScanner - - -// ================================================================================================= -// ~XMPScanner -// =========== - -XMPScanner::~XMPScanner() -{ - -} // ~XMPScanner - - -// ================================================================================================= -// GetSnipCount -// ============ - -long -XMPScanner::GetSnipCount () -{ - - return fInternalSnips.size(); - -} // GetSnipCount - - -// ================================================================================================= -// StreamAllScanned -// ================ - -bool -XMPScanner::StreamAllScanned () -{ - InternalSnipIterator currPos = fInternalSnips.begin(); - InternalSnipIterator endPos = fInternalSnips.end(); - - for ( ; currPos != endPos; ++currPos ) { - if ( currPos->fInfo.fState == eNotSeenSnip ) return false; - } - return true; - -} // StreamAllScanned - - -// ================================================================================================= -// SplitInternalSnip -// ================= -// -// Split the given snip into up to 3 pieces. The new pieces are inserted before and after this one -// in the snip list. The relOffset is the first byte to be kept, it is relative to this snip. If -// the preceeding or following snips have the same state as this one, just shift the boundaries. -// I.e. move the contents from one snip to the other, don't create a new snip. - -// *** To be thread safe we ought to lock the entire list during manipulation. Let data scanning -// *** happen in parallel, serialize all mucking with the list. - -void -XMPScanner::SplitInternalSnip ( InternalSnipIterator snipPos, XMP_Int64 relOffset, XMP_Int64 newLength ) -{ - - assert ( (relOffset + newLength) > relOffset ); // Check for overflow. - assert ( (relOffset + newLength) <= snipPos->fInfo.fLength ); - - // ----------------------------------- - // First deal with the low offset end. - - if ( relOffset > 0 ) { - - InternalSnipIterator prevPos; - if ( snipPos != fInternalSnips.begin() ) prevPos = PrevSnip ( snipPos ); - - if ( (snipPos != fInternalSnips.begin()) && (snipPos->fInfo.fState == prevPos->fInfo.fState) ) { - prevPos->fInfo.fLength += relOffset; // Adjust the preceeding snip. - } else { - InternalSnip headExcess ( snipPos->fInfo.fOffset, relOffset ); - headExcess.fInfo.fState = snipPos->fInfo.fState; - headExcess.fInfo.fOutOfOrder = snipPos->fInfo.fOutOfOrder; - fInternalSnips.insert ( snipPos, headExcess ); // Insert the head piece before the middle piece. - } - - snipPos->fInfo.fOffset += relOffset; // Adjust the remainder of this snip. - snipPos->fInfo.fLength -= relOffset; - - } - - // ---------------------------------- - // Now deal with the high offset end. - - if ( newLength < snipPos->fInfo.fLength ) { - - InternalSnipIterator nextPos = NextSnip ( snipPos ); - const XMP_Int64 tailLength = snipPos->fInfo.fLength - newLength; - - if ( (nextPos != fInternalSnips.end()) && (snipPos->fInfo.fState == nextPos->fInfo.fState) ) { - nextPos->fInfo.fOffset -= tailLength; // Adjust the following snip. - nextPos->fInfo.fLength += tailLength; - } else { - InternalSnip tailExcess ( (snipPos->fInfo.fOffset + newLength), tailLength ); - tailExcess.fInfo.fState = snipPos->fInfo.fState; - tailExcess.fInfo.fOutOfOrder = snipPos->fInfo.fOutOfOrder; - fInternalSnips.insert ( nextPos, tailExcess ); // Insert the tail piece after the middle piece. - } - - snipPos->fInfo.fLength = newLength; - - } - -} // SplitInternalSnip - - -// ================================================================================================= -// MergeInternalSnips -// ================== - -XMPScanner::InternalSnipIterator -XMPScanner::MergeInternalSnips ( InternalSnipIterator firstPos, InternalSnipIterator secondPos ) -{ - - firstPos->fInfo.fLength += secondPos->fInfo.fLength; - fInternalSnips.erase ( secondPos ); - return firstPos; - -} // MergeInternalSnips - - -// ================================================================================================= -// Scan -// ==== - -void -XMPScanner::Scan ( const void * bufferOrigin, XMP_Int64 bufferOffset, XMP_Int64 bufferLength ) -{ - XMP_Int64 relOffset; - - #if 0 - cout << "Scan: @ " << bufferOrigin << ", " << bufferOffset << ", " << bufferLength << endl; - #endif - - if ( bufferLength == 0 ) return; - - // ---------------------------------------------------------------- - // These comparisons are carefully done to avoid overflow problems. - - if ( (bufferOffset >= fStreamLength) || - (bufferLength > (fStreamLength - bufferOffset)) || - (bufferOrigin == 0) ) { - throw ScanError ( "Bad origin, offset, or length" ); - } - - // ---------------------------------------------------------------------------------------------- - // This buffer must be within a not-seen snip. Find it and split it. The first snip whose whose - // end is beyond the buffer must be the enclosing one. - - // *** It would be friendly for rescans for out of order problems to accept any buffer postion. - - const XMP_Int64 endOffset = bufferOffset + bufferLength - 1; - InternalSnipIterator snipPos = fInternalSnips.begin(); - - while ( endOffset > (snipPos->fInfo.fOffset + snipPos->fInfo.fLength - 1) ) ++ snipPos; - if ( snipPos->fInfo.fState != eNotSeenSnip ) throw ScanError ( "Already seen" ); - - relOffset = bufferOffset - snipPos->fInfo.fOffset; - if ( (relOffset + bufferLength) > snipPos->fInfo.fLength ) throw ScanError ( "Not within existing snip" ); - - SplitInternalSnip ( snipPos, relOffset, bufferLength ); // *** If sequential & prev is partial, just tack on, - - // -------------------------------------------------------- - // Merge this snip with the preceeding snip if appropriate. - - // *** When out of order I/O is supported we have to do something about buffers who's predecessor is not seen. - - if ( snipPos->fInfo.fOffset > 0 ) { - InternalSnipIterator prevPos = PrevSnip ( snipPos ); - if ( prevPos->fInfo.fState == ePartialPacketSnip ) snipPos = MergeInternalSnips ( prevPos, snipPos ); - } - - // ---------------------------------- - // Look for packets within this snip. - - snipPos->fInfo.fState = ePendingSnip; - PacketMachine* thisMachine = snipPos->fMachine.get(); - // DumpSnipList ( "Before scan" ); - - if ( thisMachine != 0 ) { - thisMachine->AssociateBuffer ( bufferOffset, bufferOrigin, bufferLength ); - } else { - // *** snipPos->fMachine.reset ( new PacketMachine ( bufferOffset, bufferOrigin, bufferLength ) ); VC++ lacks reset - #if 0 - snipPos->fMachine = auto_ptr<PacketMachine> ( new PacketMachine ( bufferOffset, bufferOrigin, bufferLength ) ); - #else - { - // Some versions of gcc complain about the assignment operator above. This avoids the gcc bug. - PacketMachine * pm = new PacketMachine ( bufferOffset, bufferOrigin, bufferLength ); - auto_ptr<PacketMachine> ap ( pm ); - snipPos->fMachine = ap; - } - #endif - thisMachine = snipPos->fMachine.get(); - } - - bool bufferDone = false; - while ( ! bufferDone ) { - - PacketMachine::TriState foundPacket = thisMachine->FindNextPacket(); - - if ( foundPacket == PacketMachine::eTriNo ) { - - // ----------------------------------------------------------------------- - // No packet, mark the snip as raw data and get rid of the packet machine. - // We're done with this buffer. - - snipPos->fInfo.fState = eRawInputSnip; - #if 0 - snipPos->fMachine = auto_ptr<PacketMachine>(); // *** snipPos->fMachine.reset(); VC++ lacks reset - #else - { - // Some versions of gcc complain about the assignment operator above. This avoids the gcc bug. - auto_ptr<PacketMachine> ap ( 0 ); - snipPos->fMachine = ap; - } - #endif - bufferDone = true; - - } else { - - // --------------------------------------------------------------------------------------------- - // Either a full or partial packet. First trim any excess off of the front as a raw input snip. - // If this is a partial packet mark the snip and keep the packet machine to be resumed later. - // We're done with this buffer, the partial packet by definition extends to the end. If this is - // a complete packet first extract the additional information from the packet machine. If there - // is leftover data split the snip and transfer the packet machine to the new trailing snip. - - if ( thisMachine->fPacketStart > snipPos->fInfo.fOffset ) { - - // There is data at the front of the current snip that must be trimmed. - SnipState savedState = snipPos->fInfo.fState; - snipPos->fInfo.fState = eRawInputSnip; // ! So it gets propagated to the trimmed front part. - relOffset = thisMachine->fPacketStart - snipPos->fInfo.fOffset; - SplitInternalSnip ( snipPos, relOffset, (snipPos->fInfo.fLength - relOffset) ); - snipPos->fInfo.fState = savedState; - - } - - if ( foundPacket == PacketMachine::eTriMaybe ) { - - // We have only found a partial packet. - snipPos->fInfo.fState = ePartialPacketSnip; - bufferDone = true; - - } else { - - // We have found a complete packet. Extract all the info for it and split any trailing data. - - InternalSnipIterator packetSnip = snipPos; - SnipState packetState = eValidPacketSnip; - - if ( thisMachine->fBogusPacket ) packetState = eBadPacketSnip; - - packetSnip->fInfo.fAccess = thisMachine->fAccess; - packetSnip->fInfo.fCharForm = thisMachine->fCharForm; - packetSnip->fInfo.fBytesAttr = thisMachine->fBytesAttr; - packetSnip->fInfo.fEncodingAttr = thisMachine->fEncodingAttr.c_str(); - thisMachine->fEncodingAttr.erase ( thisMachine->fEncodingAttr.begin(), thisMachine->fEncodingAttr.end() ); - - if ( (thisMachine->fCharForm != eChar8Bit) && CharFormIsBigEndian ( thisMachine->fCharForm ) ) { - - // ------------------------------------------------------------------------------ - // Handle a special case for big endian characters. The packet machine works as - // though things were little endian. The packet starting offset points to the - // byte containing the opening '<', and the length includes presumed nulls that - // follow the last "real" byte. If the characters are big endian we now have to - // decrement the starting offset of the packet, and also decrement the length of - // the previous snip. - // - // Note that we can't do this before the head trimming above in general. The - // nulls might have been exactly at the end of a buffer and already in the - // previous snip. We are doing this before trimming the tail from the raw snip - // containing the packet. We adjust the raw snip's size because it ends with - // the input buffer. We don't adjust the packet's size, it is already correct. - // - // The raw snip (the one before the packet) might entirely disappear. A simple - // example of this is when the packet is at the start of the file. - - assert ( packetSnip != fInternalSnips.begin() ); // Leading nulls were trimmed! - - if ( packetSnip != fInternalSnips.begin() ) { // ... but let's program defensibly. - - InternalSnipIterator prevSnip = PrevSnip ( packetSnip ); - const unsigned int nullsToAdd = ( CharFormIs16Bit ( thisMachine->fCharForm ) ? 1 : 3 ); - - assert ( nullsToAdd <= prevSnip->fInfo.fLength ); - prevSnip->fInfo.fLength -= nullsToAdd; - if ( prevSnip->fInfo.fLength == 0 ) (void) fInternalSnips.erase ( prevSnip ); - - packetSnip->fInfo.fOffset -= nullsToAdd; - packetSnip->fInfo.fLength += nullsToAdd; - thisMachine->fPacketStart -= nullsToAdd; - - } - - } - - if ( thisMachine->fPacketLength == snipPos->fInfo.fLength ) { - - // This packet ends exactly at the end of the current snip. - #if 0 - snipPos->fMachine = auto_ptr<PacketMachine>(); // *** snipPos->fMachine.reset(); VC++ lacks reset - #else - { - // Some versions of gcc complain about the assignment operator above. This avoids the gcc bug. - auto_ptr<PacketMachine> ap ( 0 ); - snipPos->fMachine = ap; - } - #endif - bufferDone = true; - - } else { - - // There is trailing data to split from the just found packet. - SplitInternalSnip ( snipPos, 0, thisMachine->fPacketLength ); - - InternalSnipIterator tailPos = NextSnip ( snipPos ); - - tailPos->fMachine = snipPos->fMachine; // auto_ptr assignment - taking ownership - thisMachine->ResetMachine (); - - snipPos = tailPos; - - } - - packetSnip->fInfo.fState = packetState; // Do this last to avoid messing up the tail split. - // DumpSnipList ( "Found a packet" ); - - - } - - } - - } - - // -------------------------------------------------------- - // Merge this snip with the preceeding snip if appropriate. - - // *** When out of order I/O is supported we have to check the following snip too. - - if ( (snipPos->fInfo.fOffset > 0) && (snipPos->fInfo.fState == eRawInputSnip) ) { - InternalSnipIterator prevPos = PrevSnip ( snipPos ); - if ( prevPos->fInfo.fState == eRawInputSnip ) snipPos = MergeInternalSnips ( prevPos, snipPos ); - } - - // DumpSnipList ( "After scan" ); - -} // Scan - - -// ================================================================================================= -// Report -// ====== - -void -XMPScanner::Report ( SnipInfoVector& snips ) -{ - const int count = fInternalSnips.size(); - InternalSnipIterator snipPos = fInternalSnips.begin(); - - int s; - - // DumpSnipList ( "Report" ); - - snips.erase ( snips.begin(), snips.end() ); // ! Should use snips.clear, but VC++ doesn't have it. - snips.reserve ( count ); - - for ( s = 0; s < count; s += 1 ) { - snips.push_back ( SnipInfo ( snipPos->fInfo.fState, snipPos->fInfo.fOffset, snipPos->fInfo.fLength ) ); - snips[s] = snipPos->fInfo; // Pick up all of the fields. - ++ snipPos; - } - -} // Report diff --git a/samples/source/XMPScanner.hpp b/samples/source/XMPScanner.hpp deleted file mode 100644 index 2c8b6fa..0000000 --- a/samples/source/XMPScanner.hpp +++ /dev/null @@ -1,330 +0,0 @@ -#ifndef __XMPScanner_hpp__ -#define __XMPScanner_hpp__ - -// ================================================================================================= -// Copyright 2002-2007 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// -// Adobe patent application tracking #P435, entitled 'Unique markers to simplify embedding data of -// one format in a file with a different format', inventors: Sean Parent, Greg Gilley. -// ================================================================================================= - -#include "XMP_Environment.h" // ! This must be the first include. - -#include <list> -#include <vector> -#include <string> -#include <memory> -#include <stdexcept> - -#include "XMP_Const.h" - -// ================================================================================================= -// The XMPScanner class is used to scan a stream of input for XMP packets. A scanner object is -// constructed then fed the input through a series of calls to Scan. Report may be called at any -// time to get the current knowledge of the input. -// -// A packet starts when a valid header is found and ends when a valid trailer is found. If the -// header contains a "bytes" attribute, additional whitespace must follow. -// -// *** RESTRICTIONS: The current implementation of the scanner has the the following restrictions: -// - The input must be presented in order. -// - Not fully thread safe, don't make concurrent calls to the same XMPScanner object. -// ================================================================================================= - -class XMPScanner { -public: - - // ============================================================================================= - // The entire input stream is represented as a series of snips. Each snip defines one portion - // of the input stream that either has not been seen, has been seen and contains no packets, is - // exactly one packet, or contains the start of an unfinished packet. Adjacent snips with the - // same state are merged, so the number of snips is always minimal. - // - // A newly constructed XMPScanner object has one snip covering the whole input with a state - // of "not seen". A block of input that contains a full XMP packet is split into 3 parts: a - // (possibly empty) raw input snip, the packet, and another (possibly empty) raw input snip. A - // block of input that contains the start of an XMP packet is split into two snips, a (possibly - // empty) raw input snip and the packet start; the following snip must be a "not seen" snip. - // - // It is possible to have ill-formed packets. These have a syntactically valid header and - // trailer, but some semantic error. For example, if the "bytes" attribute length does not span - // to the end of the trailer, or if the following packet begins within trailing padding. - - enum { - eNotSeenSnip, // This snip has not been seen yet. - ePendingSnip, // This snip is an input buffer being processed. - eRawInputSnip, // This snip is raw input, it doesn't contain any part of an XMP packet. - eValidPacketSnip, // This snip is a complete, valid XMP packet. - ePartialPacketSnip, // This snip contains the start of a possible XMP packet. - eBadPacketSnip // This snip contains a complete, but semantically incorrect XMP packet. - }; - typedef XMP_Uns8 SnipState; - - enum { // The values allow easy testing for 16/32 bit and big/little endian. - eChar8Bit = 0, - eChar16BitBig = 2, - eChar16BitLittle = 3, - eChar32BitBig = 4, - eChar32BitLittle = 5 - }; - typedef XMP_Uns8 CharacterForm; - - enum { - eChar16BitMask = 2, // These constant shouldn't be used directly, they are mainly - eChar32BitMask = 4, // for the CharFormIsXyz macros below. - eCharLittleEndianMask = 1 - }; - - #define CharFormIs16Bit(f) ( ((int)(f) & XMPScanner::eChar16BitMask) != 0 ) - #define CharFormIs32Bit(f) ( ((int)(f) & XMPScanner::eChar32BitMask) != 0 ) - - #define CharFormIsBigEndian(f) ( ((int)(f) & XMPScanner::eCharLittleEndianMask) == 0 ) - #define CharFormIsLittleEndian(f) ( ((int)(f) & XMPScanner::eCharLittleEndianMask) != 0 ) - - struct SnipInfo { - - XMP_Int64 fOffset; // The byte offset of this snip within the input stream. - XMP_Int64 fLength; // The length in bytes of this snip. - SnipState fState; // The state of this snip. - bool fOutOfOrder; // If true, this snip was seen before the one in front of it. - char fAccess; // The read-only/read-write access from the end attribute. - CharacterForm fCharForm; // How the packet is divided into characters. - const char * fEncodingAttr; // The value of the encoding attribute, if any, with nulls removed. - XMP_Int64 fBytesAttr; // The value of the bytes attribute, -1 if not present. - - SnipInfo() : - fOffset ( 0 ), - fLength ( 0 ), - fState ( eNotSeenSnip ), - fOutOfOrder ( false ), - fAccess ( ' ' ), - fCharForm ( eChar8Bit ), - fEncodingAttr ( "" ), - fBytesAttr( -1 ) - { } - - SnipInfo ( SnipState state, XMP_Int64 offset, XMP_Int64 length ) : - fOffset ( offset ), - fLength ( length ), - fState ( state ), - fOutOfOrder ( false ), - fAccess ( ' ' ), - fCharForm ( eChar8Bit ), - fEncodingAttr ( "" ), - fBytesAttr( -1 ) - { } - - }; - - typedef std::vector<SnipInfo> SnipInfoVector; - - XMPScanner ( XMP_Int64 streamLength ); - // Constructs a new XMPScanner object for a stream with the given length. - - ~XMPScanner(); - - long GetSnipCount(); - // Returns the number of snips that the stream has been divided into. - - bool StreamAllScanned(); - // Returns true if all of the stream has been seen. - - void Scan ( const void * bufferOrigin, XMP_Int64 bufferOffset, XMP_Int64 bufferLength ); - // Scans the given part of the input, incorporating it in to the known snips. - // The bufferOffset is the offset of this block of input relative to the entire stream. - // The bufferLength is the length in bytes of this block of input. - - void Report ( SnipInfoVector & snips ); - // Produces a report of what is known about the input stream. - - class ScanError : public std::logic_error { - public: - ScanError() throw() : std::logic_error ( "" ) {} - explicit ScanError ( const char * message ) throw() : std::logic_error ( message ) {} - virtual ~ScanError() throw() {} - }; - -private: // XMPScanner - - class PacketMachine; - - class InternalSnip { - public: - - SnipInfo fInfo; // The public info about this snip. - std::auto_ptr<PacketMachine> fMachine; // The state machine for "active" snips. - - InternalSnip ( XMP_Int64 offset, XMP_Int64 length ); - InternalSnip ( const InternalSnip & ); - ~InternalSnip (); - - }; // InternalSnip - - typedef std::list<InternalSnip> InternalSnipList; - typedef InternalSnipList::iterator InternalSnipIterator; - - class PacketMachine { - public: - - XMP_Int64 fPacketStart; // Byte offset relative to the entire stream. - XMP_Int32 fPacketLength; // Length in bytes to the end of the trailer processing instruction. - XMP_Int32 fBytesAttr; // The value of the bytes attribute, -1 if not present. - std::string fEncodingAttr; // The value of the encoding attribute, if any, with nulls removed. - CharacterForm fCharForm; // How the packet is divided into characters. - char fAccess; // The read-only/read-write access from the end attribute. - bool fBogusPacket; // True if the packet has an error such as a bad "bytes" attribute value. - - void ResetMachine(); - - enum TriState { - eTriNo, - eTriMaybe, - eTriYes - }; - - TriState FindNextPacket(); - - void AssociateBuffer ( XMP_Int64 bufferOffset, const void * bufferOrigin, XMP_Int64 bufferLength ); - - PacketMachine ( XMP_Int64 bufferOffset, const void * bufferOrigin, XMP_Int64 bufferLength ); - ~PacketMachine(); - - private: // PacketMachine - - PacketMachine() {}; // ! Hide the default constructor. - - enum RecognizerKind { - - eFailureRecognizer, // Not really recognizers, special states to end one buffer's processing. - eSuccessRecognizer, - - eLeadInRecognizer, // Anything up to the next '<'. - eHeadStartRecorder, // Save the starting offset, count intervening nulls. - eHeadStartRecognizer, // The literal string "?xpacket begin=". - - eBOMRecognizer, // Recognize and record the quoted byte order marker. - - eIDTagRecognizer, // The literal string " id=". - eIDOpenRecognizer, // The opening quote for the ID. - eIDValueRecognizer, // The literal string "W5M0MpCehiHzreSzNTczkc9d". - eIDCloseRecognizer, // The closing quote for the ID. - - eAttrSpaceRecognizer_1, // The space before an attribute. - eAttrNameRecognizer_1, // The name of an attribute. - eAttrValueRecognizer_1, // The equal sign and quoted string value for an attribute. - eAttrValueRecorder_1, // Record the value of an attribute. - - eHeadEndRecognizer, // The string literal "?>". - - eBodyRecognizer, // The packet body, anything up to the next '<'. - - eTailStartRecognizer, // The string literal "?xpacket end=". - eAccessValueRecognizer, // Recognize and record the quoted r/w access mode. - - eAttrSpaceRecognizer_2, // The space before an attribute. - eAttrNameRecognizer_2, // The name of an attribute. - eAttrValueRecognizer_2, // The equal sign and quoted string value for an attribute. - eAttrValueRecorder_2, // Record the value of an attribute. - - eTailEndRecognizer, // The string literal "?>". - ePacketEndRecognizer, // Look for trailing padding, check and record the packet size. - eCloseOutRecognizer, // Look for final nulls for little endian multibyte characters. - - eRecognizerCount - - }; - - XMP_Int64 fBufferOffset; // The offset of the data buffer within the input stream. - const char * fBufferOrigin; // The starting address of the data buffer for this snip. - const char * fBufferPtr; // The current postion in the data buffer. - const char * fBufferLimit; // The address one past the last byte in the data buffer. - - RecognizerKind fRecognizer; // Which recognizer is currently active. - signed long fPosition; // The internal position within a string literal, etc. - unsigned char fBytesPerChar; // The number of bytes per logical character, 1, 2, or 4. - unsigned char fBufferOverrun; // Non-zero if suspended while skipping intervening nulls. - char fQuoteChar; // The kind of quote seen at the start of a quoted value. - std::string fAttrName; // The name for an arbitrary attribute (other than "begin" and "id"). - std::string fAttrValue; // The value for an arbitrary attribute (other than "begin" and "id"). - - void SetNextRecognizer ( RecognizerKind nextRecognizer ); - - typedef TriState (* RecognizerProc) ( PacketMachine *, const char * ); - - static TriState - FindLessThan ( PacketMachine * ths, const char * which ); - - static TriState - MatchString ( PacketMachine * ths, const char * literal ); - - static TriState - MatchChar ( PacketMachine * ths, const char * literal ); - - static TriState - MatchOpenQuote ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - MatchCloseQuote ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CaptureAttrName ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CaptureAttrValue ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecordStart ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecognizeBOM ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecordHeadAttr ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CaptureAccess ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecordTailAttr ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CheckPacketEnd ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CheckFinalNulls ( PacketMachine * ths, const char * /* unused */ ); - - struct RecognizerInfo { - RecognizerProc proc; - RecognizerKind successNext; - RecognizerKind failureNext; - const char * literal; - }; - - }; // PacketMachine - - XMP_Int64 fStreamLength; - InternalSnipList fInternalSnips; - - void - SplitInternalSnip ( InternalSnipIterator snipPos, XMP_Int64 relOffset, XMP_Int64 newLength ); - - InternalSnipIterator - MergeInternalSnips ( InternalSnipIterator firstPos, InternalSnipIterator secondPos ); - - InternalSnipIterator - PrevSnip ( InternalSnipIterator snipPos ); - - InternalSnipIterator - NextSnip ( InternalSnipIterator snipPos ); - - #if DEBUG - void DumpSnipList ( const char * title ); - #endif - -}; // XMPScanner - -#endif // __XMPScanner_hpp__ diff --git a/samples/source/XmpIterations.cpp b/samples/source/XmpIterations.cpp deleted file mode 100644 index 855bb71..0000000 --- a/samples/source/XmpIterations.cpp +++ /dev/null @@ -1,318 +0,0 @@ -// ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// ================================================================================================= - -/** -* Demonstrates how to use the iteration utility in the XMPCore component to walk through property trees. -*/ - -#include <string> - -// Must be defined to instantiate template classes -#define TXMP_STRING_TYPE std::string - -// Must be defined to give access to XMPFiles -#define XMP_INCLUDE_XMPFILES 1 - -// Ensure XMP templates are instantiated -#include "XMP.incl_cpp" - -// Provide access to the API -#include "XMP.hpp" - -#include <iostream> - -using namespace std; - -// Provide some custom XMP -static const char * rdf = - "<rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'>" - " <rdf:Description rdf:about='' xmlns:xmpTest='http://ns.adobe.com/xmpTest/'>" - "" - " <xmpTest:MySimpleProp rdf:parseType='Resource'>" - " <rdf:value>A Value</rdf:value>" - " <xmpTest:MyQual>Qual Value</xmpTest:MyQual>" - " </xmpTest:MySimpleProp>" - "" - " <xmpTest:MyTopStruct rdf:parseType='Resource'>" - " <xmpTest:MySecondStruct rdf:parseType='Resource'>" - " <xmpTest:MyThirdStruct rdf:parseType='Resource'>" - " <xmpTest:MyThirdStructField>Field Value 3</xmpTest:MyThirdStructField>" - " </xmpTest:MyThirdStruct>" - " <xmpTest:MySecondStructField>Field Value 2</xmpTest:MySecondStructField>" - " </xmpTest:MySecondStruct>" - " <xmpTest:MyTopStructField>Field Value 1</xmpTest:MyTopStructField>" - " </xmpTest:MyTopStruct>" - - " <xmpTest:MyArrayWithNestedArray>" - " <rdf:Bag>" - " <rdf:li>" - " <rdf:Seq>" - " <rdf:li>Item 1</rdf:li>" - " <rdf:li>Item 2</rdf:li>" - " </rdf:Seq>" - " </rdf:li>" - " </rdf:Bag>" - " </xmpTest:MyArrayWithNestedArray>" - - " <xmpTest:MyArrayWithStructures>" - " <rdf:Seq>" - " <rdf:li rdf:parseType='Resource'>" - " <rdf:value>Field Value 1</rdf:value>" - " <xmpTest:FirstQual>Qual Value 1</xmpTest:FirstQual>" - " <xmpTest:SecondQual>Qual Value 2</xmpTest:SecondQual>" - " </rdf:li>" - " <rdf:li rdf:parseType='Resource'>" - " <rdf:value>Field Value 2</rdf:value>" - " <xmpTest:FirstQual>Qual Value 3</xmpTest:FirstQual>" - " <xmpTest:SecondQual>Qual Value 4</xmpTest:SecondQual>" - " </rdf:li>" - " </rdf:Seq>" - " </xmpTest:MyArrayWithStructures>" - "" - " <xmpTest:MyStructureWithArray rdf:parseType='Resource'>" - " <xmpTest:NestedArray>" - " <rdf:Bag>" - " <rdf:li>Item 3</rdf:li>" - " <rdf:li>Item 4</rdf:li>" - " <rdf:li>Item 5</rdf:li>" - " <rdf:li>Item 6</rdf:li>" - " </rdf:Bag>" - " </xmpTest:NestedArray>" - " <xmpTest:NestedArray2>" - " <rdf:Bag>" - " <rdf:li>Item 66</rdf:li>" - " <rdf:li>Item 46</rdf:li>" - " <rdf:li>Item 56</rdf:li>" - " <rdf:li>Item 66</rdf:li>" - " </rdf:Bag>" - " </xmpTest:NestedArray2>" - " </xmpTest:MyStructureWithArray>" - "" - " </rdf:Description>" - "</rdf:RDF>"; - -// The namespace to be used. This will be automatically registered -// when the RDF is parsed. -const XMP_StringPtr kXMP_NS_SDK = "http://ns.adobe.com/xmpTest/"; - -/** -* Reads some metadata from a file and appends some custom XMP to it. Then does several -* iterations, using various iterators. Each iteration is displayed in the console window. -*/ -int main() -{ - if(SXMPMeta::Initialize()) - { - SXMPFiles::Initialize(); - - bool ok; - SXMPFiles myFile; - - XMP_OptionBits opts = kXMPFiles_OpenForRead | kXMPFiles_OpenUseSmartHandler; - ok = myFile.OpenFile("../../../testfiles/Image1.jpg", kXMP_UnknownFile, opts); - if(ok) - { - SXMPMeta xmp; - myFile.GetXMP(&xmp); - - // Add some custom metadata to the XMP object - SXMPMeta custXMP(rdf, (XMP_StringLen) strlen(rdf)); - SXMPUtils::AppendProperties(custXMP, &xmp); - - // Store any details from the iter.Next() call - string schemaNS, propPath, propVal; - - // Only visit the immediate children that are leaf properties of the Dublin Core schema - SXMPIterator dcLeafIter(xmp, kXMP_NS_DC, (kXMP_IterJustChildren | kXMP_IterJustLeafNodes)); - while(dcLeafIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit one property from the XMP Basic schema - SXMPIterator xmpKeywordsIter(xmp, kXMP_NS_XMP, "Keywords", kXMP_IterJustLeafNodes); - while(xmpKeywordsIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit the Dublin Core schema, omit any quailifiers and only - // show the leaf properties - SXMPIterator dcIter(xmp, kXMP_NS_DC, (kXMP_IterOmitQualifiers | kXMP_IterJustLeafNodes)); - while(dcIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit the Dublin Core schema, omit any quailifiers, - // show the leaf properties but only return the leaf name and not the full path - SXMPIterator dcIter2(xmp, kXMP_NS_DC, (kXMP_IterOmitQualifiers | kXMP_IterJustLeafNodes | kXMP_IterJustLeafName)); - while(dcIter2.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Iterate over a single namespace. Show all properties within - // the Photoshop schema - SXMPIterator exifIter(xmp, kXMP_NS_Photoshop); - while(exifIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Just visit the leaf nodes of EXIF properties. That is just - // properties that may have values. - SXMPIterator exifLeafIter(xmp, kXMP_NS_EXIF, kXMP_IterJustLeafNodes); - while(exifLeafIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Iterate over all properties but skip the EXIF schema and skip the custom schema - // and continue visiting nodes - SXMPIterator skipExifIter (xmp); - while(skipExifIter.Next(&schemaNS, &propPath, &propVal)) - { - if(schemaNS == kXMP_NS_EXIF || schemaNS == kXMP_NS_SDK) - { - skipExifIter.Skip(kXMP_IterSkipSubtree); - } - else - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - } - - cout << "----------------------------------" << endl; - - // Iterate over all properties but skip the EXIF schema - // and any remaining siblings of the current node. - SXMPIterator stopAfterExifIter ( xmp ); - while(stopAfterExifIter.Next(&schemaNS, &propPath, &propVal)) - { - if(schemaNS == kXMP_NS_EXIF || schemaNS == kXMP_NS_SDK) - { - stopAfterExifIter.Skip(kXMP_IterSkipSiblings); - } - else - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - } - - cout << "----------------------------------" << endl; - - ////////////////////////////////////////////////////////////////////////////////////// - - // Iterate over the custom XMP - - // Visit the immediate children of this node. - // No qualifiers are visisted as they are below the property being visisted. - SXMPIterator justChildrenIter(xmp, kXMP_NS_SDK, kXMP_IterJustChildren); - while(justChildrenIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit the immediate children of this node but only those that may have values. - // No qualifiers are visisted as they are below the property being visisted. - SXMPIterator justChildrenAndLeafIter(xmp, kXMP_NS_SDK, (kXMP_IterJustChildren | kXMP_IterJustLeafNodes)); - while(justChildrenAndLeafIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit the leaf nodes of TopStructProperty - SXMPIterator myTopStructIter(xmp, kXMP_NS_SDK, "MyTopStruct", kXMP_IterJustLeafNodes); - while(myTopStructIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit the leaf nodes of the TopStructProperty but only return the names for - // the leaf components and not the full path - SXMPIterator xmyTopStructIterShortNames(xmp, kXMP_NS_SDK, "MyTopStruct", (kXMP_IterJustLeafNodes | kXMP_IterJustLeafName)); - while(xmyTopStructIterShortNames.Next(&schemaNS, &propPath, &propVal)) - { - cout << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit a property and all of the qualifiers - SXMPIterator iterArrayProp (xmp, kXMP_NS_SDK, "ArrayWithStructures", kXMP_IterJustLeafNodes ); - while(iterArrayProp.Next(&schemaNS, &propPath, &propVal)) - { - cout << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Visit a property and omit all of the qualifiers - SXMPIterator iterArrayPropNoQual (xmp, kXMP_NS_SDK, "ArrayWithStructures", (kXMP_IterJustLeafNodes | kXMP_IterOmitQualifiers)); - while(iterArrayPropNoQual.Next(&schemaNS, &propPath, &propVal)) - { - cout << propPath << " = " << propVal << endl; - } - - cout << "----------------------------------" << endl; - - // Skip a subtree and continue onwards. Once 'Item 4' is found then the we can skip all of the - // siblings of the current node. If the the current node were a top level node the iteration - // would be complete as all siblings would be skipped. However, when 'Item 4' is found the current - // node is not at the top level so there are other nodes further up the tree that still need to be - // visited. - SXMPIterator skipIter (xmp, kXMP_NS_SDK, (kXMP_IterJustLeafNodes | kXMP_IterOmitQualifiers | kXMP_IterJustLeafName)); - while(skipIter.Next(&schemaNS, &propPath, &propVal)) - { - if(propVal == "Item 4") - { - skipIter.Skip(kXMP_IterSkipSiblings); - } - else - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - } - - /* - // Visit all properties and qualifiers - SXMPIterator allPropsIter(xmp); - while(allPropsIter.Next(&schemaNS, &propPath, &propVal)) - { - cout << schemaNS << " " << propPath << " = " << propVal << endl; - } - */ - } - - } - - SXMPFiles::Terminate(); - SXMPMeta::Terminate(); - - return 0; -} - diff --git a/samples/source/common/DumpFile.cpp b/samples/source/common/DumpFile.cpp index 64aa548..2b8c834 100644 --- a/samples/source/common/DumpFile.cpp +++ b/samples/source/common/DumpFile.cpp @@ -1,6 +1,6 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated -// All Rights Reserved. +// Copyright 2008 Adobe Systems Incorporated +// All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms // of the Adobe license agreement accompanying it. @@ -8,27 +8,25 @@ // ================================================================================================= // // DumpFile is suitable for both dumping an entire file structure to screen _as well as_ access to -// specific to specific legacy fields (as much required for auto-testing). +// specific to specific legacy fields (as much required for auto-testing). // // Currently supports // - JPEG -// - TIFF +// - TIFF // - PHOTOSHOP // - JPEG2K // - WMAV (ASF/WMA/WMV) -// - AVI -// - WAV +// - RIFF (AVI/WAV) // - PNG // - InDesign // - MP3 // - MOV (Quicktime) -// - UCF (done, including comment zips, zip64 (>4GB)) -// - SWF (in progress) +// - UCF (done, including commented zips and zip64 (>4GB)) +// - SWF // - FLV // - MPEG-4 // -// DumpFile does depend on XMPCore, it does not depend on XMPFiles. Thus not suitable for any form -// of XMPCore testing. Suitable for XMPFiles in particular reconciliation issues. +// DumpFile does depend on XMPCore and the packetscanner from XMPFiles. #include <stdarg.h> #include "globals.h" @@ -236,9 +234,9 @@ using namespace std; #include "XMP_Const.h" //scanning routines - needed by PacketScan(...) routine for unknown files -#include "QEScanner.hpp" +#include "XMPScanner.hpp" -// QE related +// needed for logging #include "Log.h" //disabled warning (take-over) @@ -255,6 +253,10 @@ using namespace std; // resp. queried (testrunner) static TagTree* tree; +// specifc 'state machine' for QT/MPEG4 dumping +// * false by default (set in DumpISO() stub) +static bool TimeCodeTrack; + // ================================================================================================= long kOne = 1; @@ -262,6 +264,7 @@ char firstByte = *((char*)&kOne); const bool sBigEndianHost = (firstByte == 0); const bool sLittleEndianHost = (firstByte == 1); +static bool beTIFF; typedef const char * ChPtr; #define CheckBytes(left,right,len) (memcmp (((ChPtr)(left)), ((ChPtr)(right)), len) == 0) @@ -293,10 +296,10 @@ static XMP_Uns64 GetUns64LE ( const void * addr ); // ================================================================================================= // ahead declarations -static void DumpTIFF ( XMP_Uns8 * tiffContent, XMP_Uns32 tiffLen, XMP_Uns32 fileOffset, const char * label ); +static void DumpTIFF ( XMP_Uns8 * tiffContent, XMP_Uns32 tiffLen, XMP_Uns32 fileOffset, const char * label, std::string path ); static void DumpIPTC ( XMP_Uns8 * iptcOrigin, XMP_Uns32 iptcLen, XMP_Uns32 fileOffset, const char * label ); static void DumpImageResources ( XMP_Uns8 * psirOrigin, XMP_Uns32 psirLen, XMP_Uns32 fileOffset, const char * label ); -static void DumpIFDChain ( XMP_Uns8 * startPtr, XMP_Uns8 * endPtr, XMP_Uns8 * tiffContent, XMP_Uns32 fileOffset, const char * label ); +static void DumpIFDChain ( XMP_Uns8 * startPtr, XMP_Uns8 * endPtr, XMP_Uns8 * tiffContent, XMP_Uns32 fileOffset, const char * label, std::string path ); // ================================================================================================= @@ -345,6 +348,8 @@ enum { // IPTC DataSet IDs kIPTC_Instructions = 40, kIPTC_DateCreated = 55, kIPTC_TimeCreated = 60, + kIPTC_DigitalCreationDate = 62 , + kIPTC_DigitalCreationTime = 63, kIPTC_Creator = 80, kIPTC_CreatorJobtitle = 85, kIPTC_City = 90, @@ -377,6 +382,8 @@ static const DataSetInfo kDataSetNames[] = { kIPTC_Instructions, "Instructions" }, { kIPTC_DateCreated, "Date Created" }, { kIPTC_TimeCreated, "Time Created" }, +{ kIPTC_DigitalCreationDate, "Digital Creation Date" }, +{ kIPTC_DigitalCreationTime, "Digital Creation Time" }, { kIPTC_Creator, "Creator" }, { kIPTC_CreatorJobtitle, "Creator Jobtitle" }, { kIPTC_City, "City" }, @@ -795,9 +802,9 @@ void static Rewind(LFA_FileRef file) LFA_Seek (file, 0, SEEK_SET); } -XMP_Uns32 static Peek32u(LFA_FileRef file) +XMP_Uns32 static Peek32u(LFA_FileRef file, bool bigEndian = false ) { - XMP_Uns32 value = tree->digest32u(file); + XMP_Uns32 value = tree->digest32u(file, "", bigEndian ); Rewind(file, 4); return value; } @@ -888,98 +895,29 @@ CaptureXMP ( const XMP_Uns8 * xmpPtr, const XMP_Uns32 xmpLen, XMP_Int64 fileOffs // ------------------------------------------------------------------------------------------------- -static void PrintOnlyASCII_8 ( XMP_Uns8 * strPtr, XMP_Uns32 strLen, const char * /*label = 0*/, bool /*stopOnNUL*/ = true ) +static void PrintOnlyASCII_8 ( XMP_Uns8 * strPtr, XMP_Uns32 strLen, bool stopOnNUL = true ) { //wrapping to QEBuginese // - NB: remainder (zero termination earlier then length) is catered for... - tree->addComment ( convert8Bit ( strPtr, true, strLen ) ); + tree->addComment ( convert8Bit ( strPtr, stopOnNUL, strLen ) ); } // ------------------------------------------------------------------------------------------------- -static void PrintOnlyASCII_16BE ( XMP_Uns16 * u16Ptr, XMP_Uns32 u16Bytes, - const char * label = 0, bool stopOnNUL = true ) +// this wrap and the LE counterpart can only be inferior, since +// its always added as a comment, even if value was more appropriate. +// ==> callers should make use of convert16Bit directly. +static void PrintOnlyASCII_16BE ( XMP_Uns16 * u16Ptr, XMP_Uns32 u16Bytes, bool stopOnNUL = true ) { - //TODO: wrap to QEBuginese just like 8bit flavor above... - - bool prevBig = false; - size_t remainder = 0; - XMP_Uns32 u16Count = u16Bytes/2; - - if ( (label != 0) && (*label != 0) ) tree->addComment ( "%s \"", label ); - - for ( XMP_Uns32 i = 0; i < u16Count; ++i ) { - - XMP_Uns16 u16 = u16Ptr[i]; - u16 = GetUns16BE (&u16); - - if ( (0x20 <= u16) && (u16 <= 0x7E) ) { - if (prevBig) tree->addComment ( "\xA9" ); - tree->addComment ( "%c", u16 ); - prevBig = false; - } else { - if ( ! prevBig ) tree->addComment ( "\xA9" ); - tree->addComment ( "%.4X", u16 ); - prevBig = true; - if ( (u16 == 0) && stopOnNUL ) { - remainder = 2 * (u16Count - (i + 1)); - break; - } - } - - } - - if ( prevBig ) tree->addComment ( "\xA9" ); - - if ( (label != 0) && (*label != 0) ) { - tree->addComment ( "\"" ); - if (remainder != 0) tree->addComment ( "** remainder %d bytes **", remainder ); - - } - + tree->addComment ( convert16Bit( true, (XMP_Uns8*) u16Ptr, stopOnNUL, u16Bytes ) ); } // PrintOnlyASCII_16BE // ------------------------------------------------------------------------------------------------- -static void PrintOnlyASCII_16LE ( XMP_Uns16 * u16Ptr, XMP_Uns32 u16Bytes, - const char * label = 0, bool stopOnNUL = true ) +static void PrintOnlyASCII_16LE ( XMP_Uns16 * u16Ptr, XMP_Uns32 u16Bytes, bool stopOnNUL = true ) { - bool prevBig = false; - size_t remainder = 0; - XMP_Uns32 u16Count = u16Bytes/2; - - if ( (label != 0) && (*label != 0) ) tree->addComment ( "%s \"", label ); - - for ( XMP_Uns32 i = 0; i < u16Count; ++i ) { - - XMP_Uns16 u16 = u16Ptr[i]; - u16 = GetUns16LE (&u16); - - if ( (0x20 <= u16) && (u16 <= 0x7E) ) { - if ( prevBig ) tree->addComment ( "\xA9" ); - tree->addComment ( "%c", u16 ); - prevBig = false; - } else { - if ( ! prevBig ) tree->addComment ( "\xA9" ); - tree->addComment ( "%.4X", u16 ); - prevBig = true; - if ( (u16 == 0) && stopOnNUL ) { - remainder = 2 * (u16Count - (i + 1)); - break; - } - } - - } - - if ( prevBig ) tree->addComment ( "\xA9" ); - - if ( (label != 0) && (*label != 0) ) { - tree->addComment ( "\"" ); - if ( remainder != 0 ) tree->addComment ( "** remainder %d bytes **", remainder ); - - } - + tree->addComment ( convert16Bit( false, (XMP_Uns8*) u16Ptr, stopOnNUL, u16Bytes ) ); } // PrintOnlyASCII_16LE // ================================================================================================= @@ -1009,16 +947,7 @@ static const XMP_Int64 kInDesignMinSize = 2 * kINDD_PageSize; // Two master page static const XMP_Int64 kISOMediaMinSize = 16; // At least a minimal file type box. static const XMP_Uns8 kISOMediaFTyp[] = { 0x66, 0x74, 0x79, 0x70 }; // "ftyp" - static const XMP_Uns32 kISOTag_ftyp = 0x66747970UL; -static const XMP_Uns32 kISOTag_uuid = 0x75756964UL; -static const XMP_Uns32 kISOTag_moov = 0x6D6F6F76UL; -static const XMP_Uns32 kISOTag_mvhd = 0x6D766864UL; -static const XMP_Uns32 kISOTag_udta = 0x75647461UL; -static const XMP_Uns32 kISOTag_trak = 0x7472616BUL; -static const XMP_Uns32 kISOTag_meta = 0x6D657461UL; -static const XMP_Uns32 kISOTag_ilst = 0x696C7374UL; - static const XMP_Uns32 kISOBrand_mp41 = 0x6D703431UL; static const XMP_Uns32 kISOBrand_mp42 = 0x6D703432UL; static const XMP_Uns32 kISOBrand_f4v = 0x66347620UL; @@ -1120,7 +1049,7 @@ CheckFileFormat ( const char * filePath, XMP_Uns8 * fileContent, XMP_Int64 fileS static void DumpXMP (XMP_Uns8 * xmpPtr, XMP_Uns32 xmpLen, XMP_Int64 xmpOffset, const char * label) { - if (xmpOffset <= 0xFFFFFFFFUL) { + if (xmpOffset <= 0xFFFFFFFFL) { tree->pushNode("XMP"); tree->addComment("from %s, offset %u (0x%X), size %d", label, (XMP_Uns32)xmpOffset, (XMP_Uns32)xmpOffset, xmpLen); @@ -1300,12 +1229,12 @@ DumpImageResources (XMP_Uns8 * psirOrigin, XMP_Uns32 psirLen, XMP_Uns32 fileOffs tree->setKeyValue("PSIR:Exif-1",""); exif1Ptr = irPtr; exif1Len = irLen; - DumpTIFF (exif1Ptr, exif1Len, (fileOffset + (exif1Ptr - psirOrigin)), "PSIR #1058 (Exif 1)"); + DumpTIFF (exif1Ptr, exif1Len, (fileOffset + (exif1Ptr - psirOrigin)), "PSIR #1058 (Exif 1)", "PSIR:Exif-1"); } else if (irID == kPSIR_Exif_3) { //**************** tree->setKeyValue("PSIR:Exif-3",""); exif3Ptr = irPtr; exif3Len = irLen; - if (exif3Ptr != 0) DumpTIFF (exif3Ptr, exif3Len, (fileOffset + (exif3Ptr - psirOrigin)), "PSIR #1059 (Exif 3)"); + if (exif3Ptr != 0) DumpTIFF (exif3Ptr, exif3Len, (fileOffset + (exif3Ptr - psirOrigin)), "PSIR #1059 (Exif 3)", "PSIR:Exif-3"); } else if (irID == kPSIR_IPTC_Digest) { tree->setKeyValue("PSIR:IPTC digest", fromArgs("%.8X-%.8X-%.8X-%.8X", @@ -1347,7 +1276,7 @@ DumpImageResources (XMP_Uns8 * psirOrigin, XMP_Uns32 psirLen, XMP_Uns32 fileOffs static void DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, - XMP_Uns8 * tiffContent, XMP_Uns32 fileOffset, const char * label) + XMP_Uns8 * tiffContent, XMP_Uns32 fileOffset, const char * label, std::string path) { XMP_Uns8 * exifPtr = 0; XMP_Uns8 * gpsPtr = 0; @@ -1377,6 +1306,12 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, XMP_Uns16 prevTag = 0; XMP_Uns8 * fieldPtr = tiffContent+ifdOffset+2; + + if (!path.empty()) + path.append("/"); + + path.append(fromArgs("IFD%d", ifdIndex)); + for (int i = 0; i < fieldCount; ++i, fieldPtr += 12) { XMP_Uns16 fieldTag = TIFF_GetUns16 (fieldPtr); @@ -1390,7 +1325,7 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, if (valueLen <= 4) valuePtr = fieldPtr + 8; //===================== adding key here - tree->setKeyValue( fromArgs("TIFF:%d", fieldTag) ); + tree->setKeyValue( fromArgs("%s/TIFF:%d",path.c_str(), fieldTag) ); if ((fieldType < 1) || (fieldType >= kTIFF_TypeEnd)) { tree->addComment("type %d", fieldType); @@ -1438,9 +1373,9 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, xmpPtr = valuePtr; xmpLen = valueLen; } else { - for (int i = 0; sTIFF_TagNames[i].tag != 0; ++i) { - if (sTIFF_TagNames[i].tag == fieldTag) { - tree->addComment("%s", sTIFF_TagNames[i].name); + for (int j = 0; sTIFF_TagNames[j].tag != 0; ++j) { + if (sTIFF_TagNames[j].tag == fieldTag) { + tree->addComment("%s", sTIFF_TagNames[j].name); break; } } @@ -1449,8 +1384,10 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, XMP_Uns8 value8; XMP_Uns16 value16; XMP_Uns32 value32; + XMP_Uns32 denom; std::string tempStr; - + char cs[31]; + switch (fieldType) { case kTIFF_Uns8 : @@ -1482,11 +1419,19 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, break; case kTIFF_URational : - if (valueCount == 1) { - value32 = TIFF_GetUns32 (valuePtr); - XMP_Uns32 denom = TIFF_GetUns32 (valuePtr+4); - tree->changeValue( "%u/%u" , value32,denom); + for( unsigned int j = 0; j < valueCount ; j++) { + value32 = TIFF_GetUns32 (valuePtr+(j*8)); + denom = TIFF_GetUns32 (valuePtr+(j*8)+4); + snprintf(cs,30,"%u",value32); + tempStr += cs; + tempStr += "/"; + snprintf(cs,30,"%u",denom); + tempStr += cs; + if(j < valueCount-1) + tempStr += ";"; } + if(tempStr.length() > 0) + tree->changeValue( tempStr ); break; case kTIFF_Int8 : @@ -1499,12 +1444,47 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, break; case kTIFF_Undef8 : - if (valueCount == 1) { + if ( valueCount == 1 ) { value8 = *valuePtr; - tree->changeValue("0x%.2X", value8); + tree->changeValue( "0x%.2X", value8 ); + } else if ( fieldTag == 36864 ) { // ExifVersion + tree->changeValue( "%.*s", valueCount, valuePtr ); + } else if ( fieldTag == 37510 ) { // UserComment + XMP_Uns8 * encPtr = valuePtr; + valuePtr += 8; + valueCount -= 8; + sprintf(cs,"encoding = %.8s", encPtr ); + tempStr += cs; + if ( ! CheckBytes ( encPtr, "UNICODE\0", 8 ) ) { + tree->changeValue( convert8Bit ( valuePtr, false, valueCount ) ); + } else { + bool doBE = beTIFF; + if ( CheckBytes ( valuePtr, "\xFE\xFF", 2 ) ) { + doBE = true; + valuePtr += 2; + valueCount -= 2; + tempStr += ", BE BOM"; + } + if ( CheckBytes ( valuePtr, "\xFF\xFE", 2 ) ) { + doBE = false; + valuePtr += 2; + valueCount -= 2; + tempStr += ", LE BOM"; + } + if ( doBE ) { + tree->changeValue( convert16Bit( true, (XMP_Uns8*) valuePtr, false, valueCount ) ); + //PrintOnlyASCII_16BE ( (XMP_Uns16*)valuePtr, valueCount, ", value =", false /* ! stopOnNUL */ ); + } else { + tree->changeValue( convert16Bit( false, (XMP_Uns8*) valuePtr, false, valueCount ) ); + //PrintOnlyASCII_16LE ( (XMP_Uns16*)valuePtr, valueCount, ", value =", false /* ! stopOnNUL */ ); + } + } } + + if(tempStr.length() > 0) + tree->addComment( tempStr ); break; - + case kTIFF_Int16 : if (valueCount == 1) { value16 = TIFF_GetUns16 (valuePtr); @@ -1522,7 +1502,7 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, case kTIFF_SRational : if (valueCount == 1) { value32 = TIFF_GetUns32 (valuePtr); - XMP_Uns32 denom = TIFF_GetUns32 (valuePtr+4); + denom = TIFF_GetUns32 (valuePtr+4); tree->changeValue("%d/%d", *((XMP_Int32*)&value32), *((XMP_Int32*)&denom)); } break; @@ -1551,22 +1531,22 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, if (exifPtr != 0) { DumpIFDChain (exifPtr, endPtr, tiffContent, - (fileOffset + (exifPtr - tiffContent)), "TIFF tag #34665 (Exif IFD)"); + (fileOffset + (exifPtr - tiffContent)), "TIFF tag #34665 (Exif IFD)", path+"/TIFF:34665"); } if (gpsPtr != 0) { DumpIFDChain (gpsPtr, endPtr, tiffContent, - (fileOffset + (gpsPtr - tiffContent)), "TIFF tag #34853 (GPS Info IFD)"); + (fileOffset + (gpsPtr - tiffContent)), "TIFF tag #34853 (GPS Info IFD)", path+"/TIFF:34853"); } if (interopPtr != 0) { DumpIFDChain (interopPtr, endPtr, tiffContent, - (fileOffset + (interopPtr - tiffContent)), "TIFF tag #40965 (Interoperability IFD)"); + (fileOffset + (interopPtr - tiffContent)), "TIFF tag #40965 (Interoperability IFD)", path+"/TIFF:40965"); } if (makerNotePtr != 0) { DumpIFDChain (makerNotePtr, endPtr, tiffContent, - (fileOffset + (makerNotePtr - tiffContent)), "TIFF tag #37500 (Maker Note)"); + (fileOffset + (makerNotePtr - tiffContent)), "TIFF tag #37500 (Maker Note)", path+"/TIFF:37500"); } if (iptcPtr != 0) { @@ -1589,7 +1569,7 @@ DumpOneIFD (int ifdIndex, XMP_Uns8 * ifdPtr, XMP_Uns8 * endPtr, static void DumpIFDChain (XMP_Uns8 * startPtr, XMP_Uns8 * endPtr, - XMP_Uns8 * tiffContent, XMP_Uns32 fileOrigin, const char * label) + XMP_Uns8 * tiffContent, XMP_Uns32 fileOrigin, const char * label, std::string path) { XMP_Uns8 * ifdPtr = startPtr; XMP_Uns32 ifdOffset = startPtr - tiffContent; @@ -1603,7 +1583,7 @@ DumpIFDChain (XMP_Uns8 * startPtr, XMP_Uns8 * endPtr, } XMP_Uns16 fieldCount = TIFF_GetUns16 (ifdPtr); - DumpOneIFD (ifdIndex, ifdPtr, endPtr, tiffContent, fileOrigin, label); + DumpOneIFD (ifdIndex, ifdPtr, endPtr, tiffContent, fileOrigin, label, path); ifdOffset = TIFF_GetUns32 (ifdPtr+2+(12*fieldCount)); ifdPtr = tiffContent + ifdOffset; @@ -1614,7 +1594,7 @@ DumpIFDChain (XMP_Uns8 * startPtr, XMP_Uns8 * endPtr, // ================================================================================================= static void -DumpTIFF (XMP_Uns8 * tiffContent, XMP_Uns32 tiffLen, XMP_Uns32 fileOffset, const char * label) +DumpTIFF (XMP_Uns8 * tiffContent, XMP_Uns32 tiffLen, XMP_Uns32 fileOffset, const char * label, std::string path) { // ! TIFF can be nested because of the Photoshop 6 weiredness. Save and restore the procs. GetUns16_Proc save_GetUns16 = TIFF_GetUns16; @@ -1622,11 +1602,13 @@ DumpTIFF (XMP_Uns8 * tiffContent, XMP_Uns32 tiffLen, XMP_Uns32 fileOffset, const GetUns64_Proc save_GetUns64 = TIFF_GetUns64; if (CheckBytes(tiffContent,"II\x2A\x00",4)) { + beTIFF = false; TIFF_GetUns16 = GetUns16LE; TIFF_GetUns32 = GetUns32LE; TIFF_GetUns64 = GetUns64LE; tree->addComment("Little endian "); } else if (CheckBytes(tiffContent,"MM\x00\x2A",4)) { + beTIFF = true; TIFF_GetUns16 = GetUns16BE; TIFF_GetUns32 = GetUns32BE; TIFF_GetUns64 = GetUns64BE; @@ -1639,7 +1621,7 @@ DumpTIFF (XMP_Uns8 * tiffContent, XMP_Uns32 tiffLen, XMP_Uns32 fileOffset, const tree->addComment("TIFF from %s, offset %d (0x%X), size %d", label, fileOffset, fileOffset, tiffLen); XMP_Uns32 ifdOffset = TIFF_GetUns32 (tiffContent+4); - DumpIFDChain (tiffContent+ifdOffset, tiffContent+tiffLen, tiffContent, fileOffset, label); + DumpIFDChain (tiffContent+ifdOffset, tiffContent+tiffLen, tiffContent, fileOffset, label, path); TIFF_GetUns16 = save_GetUns16; TIFF_GetUns32 = save_GetUns32; @@ -1731,15 +1713,18 @@ DumpJPEG (XMP_Uns8 * jpegContent, XMP_Uns32 jpegLen) tree->addComment(fromArgs("size %d, APP%d, \"%s\"", segLen, minorKind, segName)); if ((minorKind == 1) && ((memcmp(segName,"Exif\0\0",6) == 0) || (memcmp(segName,"Exif\0\xFF",6) == 0))) { - tree->addComment("Exif"); + tree->addComment("EXIF"); + tree->changeValue("EXIF"); exifPtr = segPtr + 4 + 6; exifLen = segLen - 2 - 6; } else if ((minorKind == 13) && (strcmp(segName,"Photoshop 3.0") == 0)) { tree->addComment("PSIR"); + tree->changeValue("PSIR"); psirPtr = segPtr + 4 + strlen(segName) + 1; psirLen = (XMP_Uns16)(segLen - 2 - strlen(segName) - 1); } else if ((minorKind == 1) && (strcmp(segName,"http://ns.adobe.com/xap/1.0/") == 0)) { tree->addComment("XMP"); + tree->changeValue("XMP"); xmpPtr = segPtr + 4 + strlen(segName) + 1; xmpLen = (XMP_Uns16)(segLen - 2 - strlen(segName) - 1); } @@ -1824,7 +1809,7 @@ DumpJPEG (XMP_Uns8 * jpegContent, XMP_Uns32 jpegLen) )); } - if (exifPtr != 0) DumpTIFF (exifPtr, exifLen, (exifPtr - jpegContent), "JPEG Exif APP1"); + if (exifPtr != 0) DumpTIFF (exifPtr, exifLen, (exifPtr - jpegContent), "JPEG Exif APP1", "JPEG:APP1"); if (psirPtr != 0) DumpImageResources (psirPtr, psirLen, (psirPtr - jpegContent), "JPEG Photoshop APP13"); if (xmpPtr != 0) DumpXMP (xmpPtr, xmpLen, (xmpPtr - jpegContent), "JPEG XMP APP1"); @@ -1832,40 +1817,6 @@ DumpJPEG (XMP_Uns8 * jpegContent, XMP_Uns32 jpegLen) // ================================================================================================= -static XMP_Uns32 -DumpISOFileType (XMP_Uns8 * ftypPtr, XMP_Uns32 ftypLimit) -{ - XMP_Uns32 ftypLen = GetUns32BE (ftypPtr); - XMP_Uns32 ftypTag = GetUns32BE (ftypPtr+4); - - if (ftypLen == 0) ftypLen = ftypLimit; - - if ((ftypLen < 16) || (ftypLen > ftypLimit) || (ftypTag != kISOTag_ftyp)) { - tree->addComment( "** Invalid File Type box: tag = '%.4s' (0x%.8X), size = %d **", - ftypPtr+4, ftypTag, ftypLen); - return ftypLimit; // Force caller to skip the rest of the file. - } - - XMP_Uns32 ftypVersion = GetUns32BE (ftypPtr+12); - - tree->comment("Major Brand '%.4s', version %d",ftypPtr+8, ftypVersion); - tree->comment("Compatible brands:"); - - XMP_Uns8 * compatPtr = ftypPtr + 16; - XMP_Uns8 * compatEnd = ftypPtr + ftypLen; - - if (compatPtr >= compatEnd) - tree->addComment("<none>"); - - for (; compatPtr < compatEnd; compatPtr += 4) { - tree->addComment("'%.4s'", compatPtr); - } - - return ftypLen; -} // DumpISOFileType - -// ================================================================================================= - static const XMP_Uns8 kUUID_XMP[16] = { 0xBE, 0x7A, 0xCF, 0xCB, 0x97, 0xA9, 0x42, 0xE8, 0x9C, 0x71, 0x99, 0x94, 0x91, 0xE3, 0xAF, 0xAC }; static const XMP_Uns8 kUUID_Exif[16] = @@ -1875,314 +1826,443 @@ static const XMP_Uns8 kUUID_IPTC[16] = static const XMP_Uns8 kUUID_PSIR[16] = { 0x2C, 0x4C, 0x01, 0x00, 0x85, 0x04, 0x40, 0xB9, 0xA0, 0x3E, 0x56, 0x21, 0x48, 0xD6, 0xDF, 0xEB }; -static XMP_Uns8 * sISOPtr_XMP = 0; // ! Rather ugly, but what the heck. -static XMP_Uns32 sISOLen_XMP = 0; -static XMP_Int32 sISOPos_XMP = 0; // *** Should be 64 bits. - -static XMP_Uns8 * sISOPtr_Exif = 0; -static XMP_Uns32 sISOLen_Exif = 0; -static XMP_Uns32 sISOPos_Exif = 0; - -static XMP_Uns8 * sISOPtr_IPTC = 0; -static XMP_Uns32 sISOLen_IPTC = 0; -static XMP_Uns32 sISOPos_IPTC = 0; - -static XMP_Uns8 * sISOPtr_PSIR = 0; -static XMP_Uns32 sISOLen_PSIR = 0; -static XMP_Uns32 sISOPos_PSIR = 0; - // ------------------------------------------------------------------------------------------------- +/** + * helper routine to get past the version and flags field... + */ static void -CleanISOLegacy() +digestISOFullBoxExtension ( LFA_FileRef file, std::string isoPath, XMP_Int64& remainingSize, XMP_Uns8& version, XMP_Uns32& flags ) { - if (sISOPtr_XMP != 0) free (sISOPtr_XMP); - sISOPtr_XMP = 0; - sISOLen_XMP = 0; - sISOPos_XMP = 0; - - if (sISOPtr_Exif != 0) free (sISOPtr_Exif); - sISOPtr_Exif = 0; - sISOLen_Exif = 0; - sISOPos_Exif = 0; - - if (sISOPtr_IPTC != 0) free (sISOPtr_IPTC); - sISOPtr_IPTC = 0; - sISOLen_IPTC = 0; - sISOPos_IPTC = 0; + version = LFA_ReadUns8( file ); + flags = 0; - if (sISOPtr_PSIR != 0) free (sISOPtr_PSIR); - sISOPtr_PSIR = 0; - sISOLen_PSIR = 0; - sISOPos_PSIR = 0; + LFA_Read( file, &flags, 3, true ); // read only 3 byte! + flags = flags >> 8; // (move to bit 0-23) + remainingSize -= 4; -} // CleanISOLegacy - -// ================================================================================================= + tree->setKeyValue( isoPath + "version", fromArgs( "%d", version) ); + tree->setKeyValue( isoPath + "flags", fromArgs( "0x%.8X", flags) ); +} -static XMP_Uns32 -DumpISOBoxInfo ( LFA_FileRef file, XMP_Uns32 parent, XMP_Uns32 boxOffset, XMP_Uns32 boxLimit, - bool isQT = false, bool doPop = true ) +static void +digestInternationalTextSequence ( LFA_FileRef file, std::string isoPath, XMP_Int64* remainingSize ) { - XMP_Uns32 boxLen, boxTag, contentOffset; - XMP_Int64 bigLen = 0; - - if ( (boxOffset > boxLimit) || ((boxLimit-boxOffset) < 8) ) { - tree->comment ( "** No room for box in %d..%d, %d **", boxOffset, boxLimit, (boxLimit-boxOffset) ); - return boxLimit; // Off track, stop looking for boxes. - } - - LFA_Seek ( file, boxOffset, SEEK_SET ); - LFA_Read ( file, &boxLen, 4, true); - LFA_Read ( file, &boxTag, 4, true); - - boxLen = GetUns32BE ( &boxLen ); - contentOffset = 8; - - if ( boxLen == 0 ) { - boxLen = boxLimit; - } else if ( boxLen == 1 ) { - LFA_Read ( file, &bigLen, 8, true); - bigLen = GetUns64BE ( &bigLen ); - boxLen = Low32(bigLen); // ! HACK! - contentOffset = 16; - } + XMP_Int64 miniBoxStringSize = tree->digest16u(file,isoPath+"size",true,true); + tree->digest16u(file,isoPath+"language code",true,true); + (*remainingSize) -= 4; + if ( (*remainingSize) != miniBoxStringSize ) + tree->addComment("WARNING: boxSize and miniBoxSize differ!"); + tree->digestString( file, isoPath+"value", miniBoxStringSize, false ); +} - if ( boxLen < 8 ) { - tree->comment ( "** Bad box size of %d **", boxLen ); - return boxLimit; // Off track, stop looking for boxes. - } - - tree->pushNode ( fromArgs("'%.4s'", &boxTag) ); +/** +* dumps one *or several* (ohter while loop) ISO Boxes within the indicated space: +* +* maxBoxLen is :== fileLen on top-level, otherwise available length of outer box (exluding header size naturally) +* +* (NB: reading (and displaying) box types, compat brands and other 4-letter stuff +* as LE is somehow easier (might need adjustment for PPC though) +* +* practices: +* compensate endianess using MakeUns32BE() prior to use as string, NOT for numeric compare +*/ +static void +DumpISOBoxes ( LFA_FileRef file, XMP_Uns32 maxBoxLen, std::string _isoPath ) +{ + XMP_Int64 endOfThisLevel = LFA_Tell(file) + maxBoxLen; + std::string origIsoPath( _isoPath ); + std::string isoPath( _isoPath ); - boxTag = GetUns32BE ( &boxTag ); // ! After string print before hex print. needed below, too, to make comparisons work. - if ( bigLen == 0 ) { - tree->addComment( fromArgs( "hex value: %X, offset %d (0x%X), size %d", boxTag, boxOffset, boxOffset, boxLen ) ); - } else { - tree->addComment( fromArgs( "hex value: %X, offset %d (0x%X), size %llu (0x %X_%.8X)",boxTag, boxOffset, boxOffset, bigLen, High32(bigLen), Low32(bigLen) )); - } + while ( LFA_Tell(file) < endOfThisLevel ) + { + XMP_Int64 boxHeaderSize = 8; - // only go deeper into boxes that matter: - if ( (boxTag == kISOTag_moov) || (boxTag == kISOTag_udta) || - (boxTag == kISOTag_trak) || (boxTag == kISOTag_meta) ) { - - XMP_Uns32 childPos, childLen; - XMP_Uns32 childLimit = boxOffset + boxLen; - if ( boxTag == kISOTag_meta ) { - // reasoning here: - // meta boxes with parent udta have versions - // (as any ISO-part-12 file should, thus "!isQT or ...") - // other meta boxes (i.e. under 'trak') do not - if ( (! isQT) || (parent == kISOTag_udta) ) { - contentOffset += 4; // Skip the version and flags. - } else { - tree->addComment ( "\t** QT hack for 'meta' with no version/flags **" ); - } - } + //assertMsg("No space for ISO box header", boxHeaderSize <= maxBoxLen ); - for ( childPos = (boxOffset + contentOffset); childPos < childLimit; childPos += childLen ) { - childLen = DumpISOBoxInfo ( file, boxTag, childPos, childLimit, isQT ); + //// certainly not enough room for another box? + // could be a 32bit zero trailing a udta + // or, uhm, something garbage-ish... + if ( LFA_Tell(file) + boxHeaderSize > endOfThisLevel ) + { + XMP_Int64 numUnusedBytes = (LFA_Tell(file) + boxHeaderSize - endOfThisLevel); + tree->digestString( file, isoPath+"unused", numUnusedBytes, false ); + tree->addComment( "'free' since too small for a box" ); + + bool ok; + LFA_Seek( file, maxBoxLen, SEEK_CUR, &ok ); + assertMsg("skippind to-small space failed (truncated file?)", ok ); + continue; // could just as well: return } - } else if ( boxTag == kISOTag_ilst ) { + XMP_Int64 boxPos= LFA_Tell( file ); // store here, output below + XMP_Int64 boxSize = tree->digest32u( file, "", true ); // NB: 32bit <- 64bit + XMP_Uns32 boxType = tree->digest32u( file, "", false ); - // The Apple 'ilst' metadata appears to have a simple 2 level structure. 'Appears to' because - // we do not have official documentation. The 'ilst' box is a simple container Box (no - // version/flags), as are the boxes that it contains. It looks like those subboxes are the - // metadata items. The sub-box type is the metadata item name, the sub-box contains just a - // 'data' box with the value. - - // *** The inner loop is hardwired to assume the sub-box has an 8 byte header. - - XMP_Uns32 downOnePos, downOneLen; - XMP_Uns32 downOneLimit = boxOffset + boxLen; - - for ( downOnePos = (boxOffset + contentOffset); downOnePos < downOneLimit; downOnePos += downOneLen ) { - - downOneLen = DumpISOBoxInfo ( file, boxTag, downOnePos, downOneLimit, isQT, false /* no pop */ ); - - XMP_Uns32 downTwoPos, downTwoLen; - XMP_Uns32 downTwoLimit = downOnePos + downOneLen; + switch (boxSize) + { + case 0: + // A value of zero says that the box extends to the end of the file. + boxSize = ( maxBoxLen - boxPos ); // *** could be errorneous below top-level + break; + case 1: + // A value of 1 says that a 64-bit big endian size is written after the box type field, the data follows. + boxSize = LFA_ReadUns64_BE( file ); + boxHeaderSize += 8; + break; + default: + break; + } - for ( downTwoPos = (downOnePos + 8); downTwoPos < downTwoLimit; downTwoPos += downTwoLen ) { - downTwoLen = DumpISOBoxInfo ( file, boxTag, downTwoPos, downTwoLimit, isQT ); - } - - tree->popNode(); + std::string boxString( fromArgs( "%.4s" , &boxType ) ); + // substitute mac-copyright signs with an easier-to-handle "(c)" + if ( boxString.at(0) == 0xA9 ) + boxString = std::string("(c)") + boxString.substr(1); + isoPath = origIsoPath + boxString + "/"; - } + // TEMP + // Log::info("pushing %s, endOfThisLevel: 0x%X", isoPath.c_str(), endOfThisLevel ); - } else if ( boxTag == kQTTag_XMP_ ) { + tree->pushNode( isoPath ); + tree->addComment("offset 0x%I64X, size 0x%I64X", boxPos , boxSize); - tree->addComment ( "QuickTime XMP" ); - if ( sISOPtr_XMP != 0 ) tree->addComment ( "** Redundant QuickTime XMP **" ); - - sISOLen_XMP = boxLen - 8; - sISOPtr_XMP = (XMP_Uns8*) malloc ( sISOLen_XMP ); - sISOPos_XMP = LFA_Tell ( file ); - LFA_Read ( file, sISOPtr_XMP, sISOLen_XMP, true ); + // endOfBoxPos saves the hassle of keeping the remainingSize up-to-date + // (which is only needed and only done, if usefull for the specific box) + XMP_Int64 remainingSize = boxSize - boxHeaderSize; + XMP_Int64 endOfBoxPos = LFA_Tell(file) + remainingSize; - } else if ( (boxTag == kISOTag_uuid) && (boxLen >= 24) ) { + // --------------------------------------------- + // for FullBoxes: + XMP_Uns8 version = 255; + XMP_Uns32 flags = 0xFFFFFF; - XMP_Uns8 uuid[16]; - LFA_Read ( file, &uuid, 16, true); + switch ( boxType ) + { + // container boxes (FULL), that contain (relevant) boxes: + case 0x6174656D: // meta, FULLBOX + if ( isoPath != "moov/udta/meta/" ) + break; //no area of interest (and navigate around some malformed files) - tree->addComment ( "UUID" ); - for ( int i = 0; i < 16; ++i ) { - if ( (i & 3) == 0 ) tree->addComment ( " " ); - tree->addComment ( "%.2X", uuid[i] ); - } + digestISOFullBoxExtension( file, isoPath, remainingSize, version, flags ); + // intentionally: no break. - if ( CheckBytes ( uuid, kUUID_XMP, 16 ) ) { + // container boxes (all non-FULL), that contain (relevant) boxes: + case 0x666E696D: // minf - "simple container, no direct content" + if ( boxString == "minf" && isoPath != "moov/trak/mdia/minf/" ) + break; + case 0x6C627473: // stbl is a simple container, no direct content + { + TimeCodeTrack = false; // assume until we known better by a relevant + // moov/trak/mdia/minf/stbl/stsd/ of format tmcd + if ( boxString == "stbl" && isoPath != "moov/trak/mdia/minf/stbl/") + break; + } + case 0x766F6F6D: // moov + case 0x6169646D: // mdia + case 0x61746475: // udta - user data + case 0x6B617274: // trak - track + case 0x74736C69: // ilst (contains cprt box) + { + // store and restore current position to not depend + // on sub-level mischief... + bool ok; + XMP_Int64 keep = LFA_Tell( file ); + DumpISOBoxes( file, remainingSize, isoPath ); + LFA_Seek( file, keep, SEEK_SET, &ok ); + assertMsg( "seek failed", ok ); + } + break; - tree->addComment ( "ISO Base Media XMP" ); - if ( sISOPtr_XMP != 0 ) tree->addComment ( "** Redundant ISO Base Media XMP **" ); + // known boxes, that need content extraction + case 0x70797466: // ftyp - file type + { + XMP_Uns32 majorBrand = LFA_ReadUns32_LE( file ); + XMP_Uns32 minorVersion = LFA_ReadUns32_LE( file ); + + //Log::info( fromArgs( "major Brand: '%.4s' (0x%.8X)" , &majorBrand, MakeUns32BE(majorBrand) )); + //Log::info( fromArgs( "minor Version: 0x%.8X" , MakeUns32BE(minorVersion) ) ); + tree->setKeyValue( isoPath + "majorBrand", + fromArgs( "%.4s", &majorBrand ), + fromArgs( "0x%.8X" , MakeUns32BE(majorBrand) ) ); + tree->setKeyValue( isoPath + "minorVersion", + fromArgs( "0x%.8X", MakeUns32BE(minorVersion) ) ); + + remainingSize -= 4 + 4; + //Log::info( fromArgs( "remaining Size: %d" , remainingSize ) ); + + while ( remainingSize >= 4 ) + { + LFA_ReadUns32_LE( file ); + // TODO: Concatenate for KeyValue... + //XMP_Uns32 compatVersion = LFA_ReadUns32_LE( file ); + //Log::info( fromArgs( "compatible brand: '%.4s' (0x%.8X)" , &compatVersion, MakeUns32BE(compatVersion) )); + remainingSize -= 4; + } - sISOLen_XMP = boxLen - 24; - sISOPtr_XMP = (XMP_Uns8*) malloc ( sISOLen_XMP ); - sISOPos_XMP = LFA_Tell ( file ); - LFA_Read ( file, sISOPtr_XMP, sISOLen_XMP, true ); + // odd bytes left? + if ( remainingSize > 0 ) + tree->addComment( "WARNING: %d bytes left, considering FREE", remainingSize ); - } else if ( CheckBytes ( uuid, kUUID_Exif, 16 ) ) { + } + break; + case 0x61746164: // data (within itunes Metadata) + // all data atoms start with two common fields: a type indicator, and a locale indicator. + // each of these fields is four bytes long: + tree->digest32u(file, isoPath+ "type", true, true); + tree->digest32u(file, isoPath+ "locale", true, true); + remainingSize -= 8; + // rest is actual contents: + tree->digestString( file, isoPath + "value", remainingSize, false ); + break; + case 0x64697575: // uuid + XMP_Uns8 uid[16]; + tree->digest( file, isoPath + "uuidValue" , uid , 16 ); + if ( ! strncmp( (const char*) kUUID_XMP, (const char*) uid, 16 )) + tree->addComment(" - the XMP UUID !"); - tree->addComment ( "ISO Base Media Exif" ); - if ( sISOPtr_Exif != 0 ) tree->addComment ( "** Redundant ISO Base Media Exif **" ); + break; - sISOLen_Exif = boxLen - 24; - sISOPtr_Exif = (XMP_Uns8*) malloc ( sISOLen_Exif ); - sISOPos_Exif = LFA_Tell ( file ); - LFA_Read ( file, sISOPtr_Exif, sISOLen_Exif, true ); + case 0x65657266: // free + tree->addComment( "free space" ); + break; - } else if ( CheckBytes ( uuid, kUUID_IPTC, 16 ) ) { + // FULL BOXES (w/o container boxes, above) ********************************** + case 0x6468766D: // mvhd, FULLBOX, movie-header-box + { + digestISOFullBoxExtension( file, isoPath, remainingSize, version, flags ); + if ( version == 1 ) + { + tree->digest64u(file,isoPath+ "creation_time",true,true); + tree->digest64u(file,isoPath+ "modification_time",true,true); + tree->digest32u(file,isoPath+ "timescale",true,true); + tree->digest64u(file,isoPath+ "duration",true,true); + } + else if ( version == 0 ) + { + tree->digest32u(file,isoPath+ "creation_time",true,true); + tree->digest32u(file,isoPath+ "modification_time",true,true); + tree->digest32u(file,isoPath+ "timescale",true,true); + tree->digest32u(file,isoPath+ "duration",true,true); + } else + { + tree->addComment("WARNING: unknown mvhd version!"); + } + // COULDDO more fields, but not needed right now. + } + break; + case 0x726C6468: // hdlr - handler reference + { + if ( isoPath != "moov/trak/mdia/hdlr/") + break; + if ( remainingSize < 4*4 ) + break; // box too small... - tree->addComment ( "ISO Base Media IPTC" ); - if ( sISOPtr_IPTC != 0 ) tree->addComment ( "** Redundant ISO Base Media IPTC **" ); + digestISOFullBoxExtension( file, isoPath, remainingSize, version, flags ); - sISOLen_IPTC = boxLen - 24; - sISOPtr_IPTC = (XMP_Uns8*) malloc ( sISOLen_IPTC ); - sISOPos_IPTC = LFA_Tell ( file ); - LFA_Read ( file, sISOPtr_IPTC, sISOLen_IPTC, true ); + tree->digestString( file, isoPath+"quickTimeType", 4, false ); // expecting: 'mhlr' - media handler + tree->digestString( file, isoPath+"subType", 4, false ); // expecting: 'tmcd' - timecode + tree->digestString( file, isoPath+"manufacturer", 4, false ); // e.g. 'appl' - } else if ( CheckBytes ( uuid, kUUID_PSIR, 16 ) ) { + break; // rest doesn't bother us... + } + case 0x64737473: // stsd - timecode sample description table + { + if ( isoPath != "moov/trak/mdia/minf/stbl/stsd/") + break; - tree->addComment ( "ISO Base Media PSIR" ); - if ( sISOPtr_PSIR != 0 ) tree->addComment ( "** Redundant ISO Base Media PSIR **" ); + // version (1 byte), flags (3 byte) - must be 0 + assertMsg("stbl: version and flags must be zero", 0 == tree->digest32u( file, "", true ) ); + // entryCount - must be 0 + assertMsg("stbl: at least one entry needed", 1 <= tree->digest32u( file, "", true ) ); + remainingSize -= 8; + + // only dump first occurence + // ensure there's enough bytes for at least on stbl: + if ( remainingSize < 29 ) + break; // MPEG4/al_sbr_twi_22_1_fsaac22.mv4 has a box that is smaller than this... + // hence can only break, not throw + XMP_Int64 entrySize = (XMP_Int64) tree->digest32u( file, "", true, true ); + std::string format = tree->digestString( file, "", 4, false ); // must be 'tmcd' + tree->pushNode( isoPath + format + "/" ); + if ( format != "tmcd" ) + { + tree->addComment("irrelevant node"); + tree->popNode(); + break; // different party... + } + + TimeCodeTrack = true; // we're in the right track + + Skip( file, 6 ); // [6] reserved bytes + tree->digest16u( file, isoPath+"dataReferenceIndex", true, true ); // (ignored) + tree->addOffset( file ); + Skip( file, 4 ); // uint32 reserved + tree->digest32u( file, isoPath+"flags", true, true ); + tree->digest32u( file, isoPath+"timeScale", true, true ); + tree->digest32u( file, isoPath+"frameDuration", true, true ); + Skip( file, 2 ); // skip ignored frame count, reserved + + // ////////////////////////////// dig out 'trailing boxes' + // comparing "atom remains" vs. "entry" (probably must be '>=' + + // deduct the already digested... + entrySize -= 34; // ( 4+4+6+2+4+4+4+4+1+1 ) + remainingSize -= 34; // (the atom-level value) + + assertMsg( "entry Size must be 0 or positive", entrySize >= 0 ); + assertMsg( "must not overreach atom", entrySize <= remainingSize ); + + XMP_Int64 endOfTrailingBoxes = LFA_Tell(file) + remainingSize; + while ( LFA_Tell(file) < endOfTrailingBoxes ) + { + LFA_Tell( file ); + DumpISOBoxes( file, entrySize, isoPath ); + LFA_Tell( file ); + } - sISOLen_PSIR = boxLen - 24; - sISOPtr_PSIR = (XMP_Uns8*) malloc ( sISOLen_PSIR ); - sISOPos_PSIR = LFA_Tell ( file ); - LFA_Read ( file, sISOPtr_PSIR, sISOLen_PSIR, true ); - } + assertMsg( "did not boil down to zero", LFA_Tell(file) == endOfTrailingBoxes ); - } + tree->popNode(); + break; + } - if ( doPop ) tree->popNode(); + case 0x63737473: // stsc - timecode sample description table + { + if ( isoPath != "moov/trak/mdia/minf/stbl/stsc/") + break; + if ( !TimeCodeTrack ) + { + tree->addComment("not tcmd -> not of interest"); + break; // not of interest + } - return boxLen; + // version (1 byte), flags (3 byte) - must be 0 + assertMsg("stbl: version and flags must be zero", 0 == tree->digest32u( file, "", true ) ); + // entryCount - must be 0 + assertMsg("stbl: at least one entry needed", 1 <= tree->digest32u( file, "", true ) ); + remainingSize -= 8; -} // DumpISOBoxInfo + LFA_Tell( file ); -// ================================================================================================= + tree->digest32u( file, isoPath+"firstChunkNo", true, true ); + tree->digest32u( file, isoPath+"numSamplesPerChunk", true, true ); + tree->digest32u( file, isoPath+"sampleDescriptionID", true, true ); + + break; + } -static void -DumpJPEG2000 (LFA_FileRef file, XMP_Uns32 jp2kLen) -{ - // JPEG 2000 files are ISO base media file format, - // ISO 14496-12, with an initial 12 byte signature. + case 0x6F637473: // stco - timecode sample description table + { + if ( isoPath != "moov/trak/mdia/minf/stbl/stco/") + break; + if ( !TimeCodeTrack ) + break; // not of interest + + // version (1 byte), flags (3 byte) - must be 0 + assertMsg("stbl: version and flags must be zero", 0 == tree->digest32u( file, "", true ) ); + // entryCount - must be 0 + assertMsg("stbl: at least one entry needed", 1 <= tree->digest32u( file, "", true ) ); + remainingSize -= 8; + + XMP_Int64 absOffset = tree->digest32u( file, isoPath+"absFileOffset32", true, true ); + + // recklessly navigate to that timecode media sample, grab value, return to old position... + XMP_Int64 oldPos = LFA_Tell( file ); + LFA_Seek( file, absOffset, SEEK_SET, 0 ); + tree->digest32u( file, isoPath+"timecodeMediaSample", true, true ); + LFA_Seek( file, oldPos, SEEK_SET, 0 ); - XMP_Uns8 buffer[4096]; - size_t bufLen; + } + case 0x34366F63: // co64 - timecode sample description table -> 64 bit offset + { + if ( isoPath != "moov/trak/mdia/minf/stbl/co64/") + break; - LFA_Seek ( file, 12, SEEK_SET ); - bufLen = LFA_Read ( file, buffer, sizeof(buffer), false); + tree->digest64u( file, isoPath+"absFileOffset64", true, true ); + break; + } - XMP_Uns32 ftypLen = DumpISOFileType ( buffer, bufLen ); + case 0x74727063: // cprt, FULLBOX + if ( isoPath == "moov/udta/cprt/") + { - XMP_Uns32 boxPos, boxLen; + digestISOFullBoxExtension( file, isoPath, remainingSize, version, flags ); - for ( boxPos = ftypLen+12; boxPos < jp2kLen; boxPos += boxLen ) { - boxLen = DumpISOBoxInfo ( file, 0, boxPos, jp2kLen ); - } + // 1/8 byte ISO language padding = 1-bit value set to 0 + // 1 7/8 bytes content language = 3 * 5-bits ISO 639-2 language code less 0x60 + // - example code for english = 0x15C7 + tree->digest16u(file,isoPath+"content language",true,true); + tree->addComment("(0x15C7 == english)"); + // zero-terminated, actual string: + tree->digestString( file, isoPath+"value", 0 ); + } else + { + // ISO - copyright (?) + // a container box, hunt for 'data' atom by recursion: + bool ok; + XMP_Int64 keep = LFA_Tell( file ); + DumpISOBoxes( file, remainingSize, isoPath ); + LFA_Seek( file, keep, SEEK_SET, &ok ); + assertMsg( "seek failed", ok ); + } + break; + case 0x64686B74: // tkhd, FULLBOX + case 0x6E61656D: // mean, FULLBOX + case 0x656D616E: // name, FULLBOX + { + if ( isoPath == "moov/trak/mdia/minf/stbl/stsd/name/" ) // this regrettably is a diffrent animal (international text sequence) + digestInternationalTextSequence( file, isoPath, &remainingSize ); + else + digestISOFullBoxExtension( file, isoPath, remainingSize, version, flags ); + } + break; - if ( sISOPtr_Exif != 0 ) { - DumpTIFF ( sISOPtr_Exif, sISOLen_Exif, sISOPos_Exif, "JPEG 2000 Exif" ); - } + // (c)-style quicktime boxes and boxes of no interest: + default: + if ( (boxType & 0xA9) == 0xA9) // (c)something + { + if ( 0 == isoPath.compare( 0 , 20, "moov/udta/meta/ilst/")) + { // => iTunes metadata (hunt for data childs) + // a container box, hunt for 'data' atom by recursion: + bool ok; + XMP_Int64 keep = LFA_Tell( file ); + DumpISOBoxes( file, remainingSize, isoPath ); + LFA_Seek( file, keep, SEEK_SET, &ok ); + assertMsg( "seek failed", ok ); + } + else if ( 0 == isoPath.compare( 0 , 10, "moov/udta/" )) + { // => Quicktime metadata "international text sequence" ( size, language code, value ) + digestInternationalTextSequence( file, isoPath, &remainingSize ); + } else + { + tree->addComment("WARNING: unknown flavor of (c)*** boxes, neither QT nor iTunes"); + } + break; + } + //boxes of no interest: - if ( sISOPtr_IPTC != 0) { - DumpIPTC ( sISOPtr_IPTC, sISOLen_IPTC, sISOPos_IPTC, "JPEG 2000 IPTC" ); - } + break; + } - if ( sISOPtr_PSIR != 0 ) { - DumpImageResources ( sISOPtr_PSIR, sISOLen_PSIR, sISOPos_PSIR, "JPEG 2000 PSIR" ); - } + bool ok; + LFA_Seek( file, endOfBoxPos, SEEK_SET, &ok ); + assertMsg("End-of-Box Seek failed (truncated file?)", ok ); - if ( sISOPtr_XMP != 0 ) { - DumpXMP ( sISOPtr_XMP, sISOLen_XMP, sISOPos_XMP, "JPEG 2000 XMP" ); - } + tree->popNode(); - CleanISOLegacy(); -} // DumpJPEG2000 + } // while -// ================================================================================================= +} // DumpISOBoxes() // attempt to combine dumping of mpeg-4 and quicktime (mov) into one routine... static void -DumpISO ( LFA_FileRef file, XMP_Uns32 fileLen, bool isQT ) +DumpISO ( LFA_FileRef file, XMP_Uns32 fileLen ) { - // MPEG-4 - // uses the ISO base media file format, ISO 14496-12. The base structure is a sequence of - // boxes, a box can have nested boxes. The box structure is: - // A big endian UInt32 box size, the offset to the following box - // A big endian UInt32 box type, generally 4 ASCII characters - // The box data - // A box size of 0 means the last box in the file. If the box size is 1 a big endian UInt64 size - // is present between the box type and data. A common subtype is a UUID box. These have a type - // of uuid, followed by a 16 byte UUID, followed by other box data. - // - // The XMP is in a top level UUID box with the UUID BE7ACFCB 97A942E8 9C719994 91E3AFAC. - - // MOV - // A QuickTime file happens to be identical to the ISO base media file format at the raw level. - // * The chunks are called "atoms" instead of "boxes" - // * A leading 'ftyp' atom is not required. - - XMP_Uns32 boxPos = 0; //assume null (presence of ftyp box will correct) - XMP_Uns32 boxLen; - - // TODO: make obsolete - XMP_Uns8 buffer[4096]; - size_t bufLen; - LFA_Seek ( file, 0, SEEK_SET ); - bufLen = LFA_Read ( file, buffer, sizeof(buffer) ); - - //peek at first 4 bytes - Rewind(file); - Skip(file,4); - - XMP_Uns32 temp = tree->digest32u( file , "" , true, true ); - bool hasFtyp = ( kISOTag_ftyp == temp); - Rewind(file,4); - - //in Quicktime, a leading 'ftyp' atom is optional, - //in MPEG-4 it is mandatory: - if (! isQT) - assertMsg("MPEG-4 File must have a first 'ftyp' box!", hasFtyp ); - - // if existing, go parse it: - if( hasFtyp ) - boxPos = DumpISOFileType ( buffer, bufLen ); - - // iterate rest of boxes - for ( ; boxPos < fileLen; boxPos += boxLen ) { - boxLen = DumpISOBoxInfo ( file, 0, boxPos, fileLen, isQT); - } - - // final memory test - assertMsg("Unexpected Exif (somewhere in file)", sISOPtr_Exif == 0); - assertMsg("Unexpected IPTC (somewhere in file)", sISOPtr_IPTC == 0); - assertMsg("Unexpected PSIR (somewhere in file)", sISOPtr_PSIR == 0); + TimeCodeTrack = false; - CleanISOLegacy(); //free's some memory + // see specificition at https://zerowing.corp.adobe.com/display/XMP/Embedding+Spec+MPEG4 + DumpISOBoxes(file, fileLen, "" ); + assertMsg("truncated file/last box reached beyond end?", LFA_Tell(file) == fileLen ); } // ================================================================================================= @@ -2397,12 +2477,12 @@ static void DumpASFContentEncryption (LFA_FileRef file, XMP_Uns32 objOffset, XMP LFA_Read ( file, &fieldSize, 4, true); fieldSize = GetUns32LE (&fieldSize); CaptureFileData (file, 0, fieldSize); - PrintOnlyASCII_8 (sDataPtr, fieldSize, "Key ID"); + PrintOnlyASCII_8 (sDataPtr, fieldSize); LFA_Read ( file, &fieldSize, 4, true); fieldSize = GetUns32LE (&fieldSize); CaptureFileData (file, 0, fieldSize); - PrintOnlyASCII_8 (sDataPtr, fieldSize, "License URL"); + PrintOnlyASCII_8 (sDataPtr, fieldSize); } // DumpASFContentEncryption @@ -2830,171 +2910,314 @@ DumpUCF(LFA_FileRef file, XMP_Int64 len) tree->setKeyValue("file data","",fromArgs("skipping %u bytes",compressed_size)); Skip(file,compressed_size); } - - - - tree->popNode(); } - ///////////////////////////////////////////////////////////////////// - // : tree->pushNode(""); tree->popNode(); - - ///////////////////////////////////////////////////////////////////// - // : - - ///////////////////////////////////////////////////////////////////// - // : - - ///////////////////////////////////////////////////////////////////// - // : - - } // DumpUCF // ================================================================================================= -static const char * kRIFF_Lists[] = { "INFO", "hdrl", "Tdat", "Cdat", 0 }; - -static const char * kRIFF_LegacyText[] = +// AVI and WAV files are RIFF based. Although they have usage differences, we can have a common +// dumper. This might need changes for other RIFF-based files, e.g. for specific legacy +// metadata. RIFF is a chunky format. AVI and WAV have an outermost RIFF chunk. The XMP is in a +// top level "_PMX" chunk. Legacy metadata for WAV is in a top level LIST/INFO chunk. Legacy +// metadata for AVI is in a variety of places, don't have specs at present. Free space can be +// JUNK or JUNQ. +// +// A RIFF chunk contains: +// - A 4 byte chunk ID, typically ASCII letters +// - A little endian UInt32 size of the chunk data +// - The chunk data +// - Pad byte if the chunk data length is odd (added on 2007-03-22) + +// The ID is written in "reading order", e.g. the 'R' in "RIFF" is first in the file. Chunks +// must start on even offsets. A pad byte of 0 is written after the data if necessary. The size +// does not include the pad, nor the ID and size fields. Some chunks contain nested chunks, +// notably the RIFF and LIST chunks do. These have the layout: +// - A 4 byte chunk ID, typically ASCII letters +// - A little endian UInt32 size of the chunk data +// - A 4 byte usage ID, typically ASCII letters +// - The nested chunks + +// reads maxSize bytes from file (not "up to", exactly fullSize) +// puts it into a string, sets respective tree property +static void setFixedBEXTField ( LFA_FileRef file, std::string propName, XMP_Int64 fullSize ) { - // INFO list items, taken from Exif 2.2, ASCII with terminating nul: - "IARL", "IART", "ICMS", "ICMT", "ICOP", "ICRD", "ICRP", "IDIM", "IDPI", "IENG", "IGNR", - "IKEY", "ILGT", "IMED", "INAM", "IPLT", "IPRD", "ISBJ", "ISFT", "ISHP", "ISRC", "ISRF", "ITCH", - // AVI legacy items, taken from AVI handler code: - "ISMT", "tc_O", "tc_A", "rn_O", "rn_A", "cmnt", - // WAV legacy items, taken from WAV handler code: - "DISP", - // Final sentinel - 0 -}; - -// ================================================================================================= + char* descriptionBuffer = new char[ fullSize + 2 ]; + LFA_Read( file, descriptionBuffer, fullSize, true ); + descriptionBuffer[fullSize]='\0'; // tack on, in case not contained + // parse till first \0 + std::string description( descriptionBuffer ); + tree->setKeyValue( propName, description) ; + delete[] descriptionBuffer; +} -static bool -IsKnownRIFFList ( const char * listID ) +static void +DumpRIFFChunk ( LFA_FileRef file, XMP_Int64 parentEnd, std::string origChunkPath ) { - for ( size_t i = 0; kRIFF_Lists[i] != 0; ++i ) { - if ( strncmp ( listID, kRIFF_Lists[i], 4 ) == 0 ) return true; - } - - return false; - -} // IsKnownRIFFList -// ================================================================================================= + while ( LFA_Tell(file) < parentEnd ) + { + bool isOutermost = origChunkPath.empty(); -static bool -IsKnownRIFFLegacy ( const char * chunkID ) -{ - for ( size_t i = 0; kRIFF_LegacyText[i] != 0; ++i ) { - if ( strncmp ( chunkID, kRIFF_LegacyText[i], 4 ) == 0 ) return true; - } + XMP_Int64 chunkPos= LFA_Tell( file ); + XMP_Uns32 chunkID = tree->digest32u( file, "", false ); + std::string idString( fromArgs( "%.4s" , &chunkID ) ); + XMP_Int64 chunkSize = tree->digest32u( file, "", false ) // NB: XMPInt64 <- XMPUns32 + + 8; //adding size of id and length field itself - return false; + // calculate size if size field seems broken + if (chunkSize > parentEnd) + chunkSize = parentEnd - chunkPos; -} // IsKnownRIFFLegacy + std::string chunkPath = isOutermost ? ( idString ) : (origChunkPath + "/" + idString); + + // check special case of trailing bytes not in a valid RIFF structure + if ( isOutermost && idString != "RIFF") + { + //dump undefined bytes till the end of the file + tree->pushNode("** unknown bytes **"); + tree->addOffset( file ); + chunkSize = parentEnd - chunkPos; // get size through calculation (and not from size bytes) + tree->addComment("size: 0x%llX", chunkSize ); + Skip( file, chunkSize-8 ); // Already read the 8 byte header. + tree->popNode(); + } + else + { + // only RIFF and LIST contain subchunks... + bool hasSubChunks = (idString == "RIFF") || (idString == "LIST"); + bool skipper=false; + if (hasSubChunks) + { + //get inner ID 'type' as in 'listType', 'fileType', ... + XMP_Uns32 chunkType = tree->digest32u( file ); + std::string typeString( fromArgs( "%.4s" , &chunkType ) ); + + if ( isOutermost ) + assertMsg("level-0 chunk must be AVI, AVIX or WAVE", + ( typeString == "AVI " ) || ( typeString == "AVIX" ) || ( typeString == "WAVE" ) ); + + chunkPath = chunkPath + ":" + typeString; + + tree->pushNode( chunkPath ); + tree->addComment("offset 0x%llX, size 0x%llX", chunkPos , chunkSize); + + if ( ( idString + ":" + typeString == "LIST:INFO" ) || + ( idString + ":" + typeString == "LIST:Tdat" ) || + ( idString + ":" + typeString == "RIFF:AVI " ) || + ( idString + ":" + typeString == "RIFF:AVIX" ) || + ( idString + ":" + typeString == "RIFF:WAVE" ) || + ( idString + ":" + typeString == "LIST:hdrl" ) || + ( idString + ":" + typeString == "LIST:strl" ) || + ( idString + ":" + typeString == "LIST:movi" ) + ) + { + DumpRIFFChunk( file, LFA_Tell(file) + chunkSize - 12, chunkPath ); // recurse! + } + else + { + Skip( file, chunkSize - 12 ); // skip it ! + } + tree->popNode(); + } + else if (idString.length() == 4) // check that we got a valid idString + { + + // now that LIST:movi gets dumped, + // skip some very frequent, irrelevant chunks, + // otherwise the dump becomes unusably long... + std::string firstTwo = idString.substr(0,2); + std::string secondTwo = idString.substr(2,2); + if ( secondTwo == "db" || secondTwo == "dc" || secondTwo == "wb" ) // nb: _could_ colidde, requiring additional numeric test + { + skipper = true; + } + + if ( ! skipper ) + { + tree->pushNode( chunkPath ); + //Log::info( chunkPath ); + tree->addComment("offset 0x%llX, size 0x%llX", chunkPos , chunkSize); + } -// ================================================================================================= + // tackle chunks of interest ////////////////////////////////////////////// + bool isListInfo = + ( (origChunkPath == "RIFF:WAVE/LIST:INFO" || origChunkPath == "RIFF:AVI /LIST:INFO" ) + && idString.at(0) == 'I' ); // so far all mapping relevant props begin with "I" -static XMP_Uns32 -DumpRIFFChunk ( LFA_FileRef file, XMP_Uns32 chunkOffset, XMP_Uns32 parentEnd, size_t nesting ) -{ - // AVI and WAV files are RIFF based. Although they have usage differences, we can have a common - // dumper. This might need changes for other RIFF-based files, e.g. for specific legacy - // metadata. RIFF is a chunky format. AVI and WAV have an outermost RIFF chunk. The XMP is in a - // top level "_PMX" chunk. Legacy metadata for WAV is in a top level LIST/INFO chunk. Legacy - // metadata for AVI is in a variety of places, don't have specs at present. Free space can be - // JUNK or JUNQ. - // - // A RIFF chunk contains: - // - A 4 byte chunk ID, typically ASCII letters - // - A little endian UInt32 size of the chunk data - // - The chunk data - // - Pad byte if the chunk data length is odd (added on 2007-03-22) - - // The ID is written in "reading order", e.g. the 'R' in "RIFF" is first in the file. Chunks - // must start on even offsets. A pad byte of 0 is written after the data if necessary. The size - // does not include the pad, nor the ID and size fields. Some chunks contain nested chunks, - // notably the RIFF and LIST chunks do. These have the layout: - // - A 4 byte chunk ID, typically ASCII letters - // - A little endian UInt32 size of the chunk data - // - A 4 byte usage ID, typically ASCII letters - // - The nested chunks + bool isListTdat = (origChunkPath == "RIFF:WAVE/LIST:Tdat" || origChunkPath == "RIFF:AVI /LIST:Tdat") + && idString.at(0) != 'J' ; // just exclude JUNK/Q - LFA_Seek ( file, chunkOffset, SEEK_SET ); - char chunkID[4]; - LFA_Read ( file, chunkID, 4, true ); + bool isDispChunk = + ( ( origChunkPath == "RIFF:WAVE" || origChunkPath == "RIFF:AVI ") + && idString == "DISP" ); - XMP_Uns32 chunkLen = tree->digest32u(file); + bool isBextChunk = + ( ( origChunkPath == "RIFF:WAVE" || origChunkPath == "RIFF:AVI ") + && idString == "bext" ); - bool isRIFF = CheckBytes ( chunkID, "RIFF", 4 ); - bool isLIST = CheckBytes ( chunkID, "LIST", 4 ); + bool isXMPchunk = false; //assume beforehand + if ( idString == "_PMX" ) + { // detour first, to detect xmp in wrong places + assertMsg( "XMP packet found in wrong place!", + ( origChunkPath == "RIFF:WAVE" || "RIFF:AVI" || "RIFF:AVIX" ) ); //be very linient here. + isXMPchunk = true; + } - if ( isRIFF | isLIST ) { - char subID[4]; - LFA_Read ( file, subID, 4, true ); + // deal with chunks of interest ///////////////////////////////////////////// + // a little prelude for disp chunk + if ( isDispChunk ) + { + XMP_Uns32 dispChunkType = LFA_ReadUns32_LE(file); + // only dispChunks starting with a 0x0001 are of interest to us. + // others do exist and are not an error - bool knownList = false; - if ( isLIST ) knownList = IsKnownRIFFList ( subID ); + if ( dispChunkType != 0x0001 ) + isDispChunk = false; - tree->pushNode ( "%.4s:%.4s" , chunkID, subID ); - tree->addComment ( "offset 0x%X, size %d%s", - chunkOffset, chunkLen, (((chunkLen & 1) == 0) ? "" : " (has pad byte)") ); + chunkSize -= 4; + } - if ( (chunkOffset + chunkLen) > parentEnd ) { - tree->addComment ( "Chunk too long, extends beyond parent" ); - } + if ( isListInfo || isListTdat || isDispChunk ) + { + // dump that string: + std::string value; - if ( (isRIFF && (nesting == 0)) || (knownList && (nesting == 1)) ) { + if ( chunkSize > 8 ) // aka skip for empty chunks + { + // first check if the string is zero terminated + LFA_Seek( file , chunkSize - 8 - 1, SEEK_CUR ); // jump to last char + bool zeroTerm = (LFA_ReadUns8( file ) == 0); + LFA_Seek( file , -(chunkSize - 8 ), SEEK_CUR ); //jump back + // some strings are zero-terminated (so despite initial length they are "c-strings" + // others are not ( "pascal strings" if you will. + // must cater to both: zero-terminated-ness should not affect resulting value. + + // Samy: also dump out the zero termination. Needed for testing + + if (zeroTerm) + { + // read string without zero (last char) + value = tree->digestString( file, "" , chunkSize - 8 - 1, false ); + tree->addComment(" zero terminated"); + LFA_ReadUns8( file ); // skip the zero + } + else + { + // read string including last char + value = tree->digestString( file, "" , chunkSize - 8 , false ); + tree->addComment(" not zero terminated"); + } + tree->changeValue( value ); + } - XMP_Uns32 childLimit = chunkOffset + 8 + chunkLen; - XMP_Uns32 childOffset, childLen; + tree->changeValue( value ); + } + else if ( isXMPchunk ) + { + tree->pushNode("XMP packet"); + + tree->addOffset( file ); + tree->addComment("packet size: 0x%llX", chunkSize - 8 ); + Skip( file, chunkSize - 8 ); + tree->addComment("packet end: 0x%llX", LFA_Tell( file ) ); + + tree->popNode(); + } + else if ( isBextChunk ) + { + tree->pushNode("bext chunk"); + tree->addOffset( file ); + tree->addComment("packet size: 0x%llX", chunkSize - 8 ); + + // I assume that the minimum BEXT chunk size is 602: + // > 8 + ( 256+32+32+10+8+4+4+2+64+190+0 ) + // ans = 610 + const XMP_Int64 MIN_BEXT_SIZE = 610; + assertMsg("minimum Berx Chunk Size", chunkSize >= MIN_BEXT_SIZE ); + XMP_Int64 BEXT_CodingHistorySize = chunkSize - MIN_BEXT_SIZE; + + setFixedBEXTField ( file, chunkPath+".Description" , 256 ); + setFixedBEXTField ( file, chunkPath+".Originator" , 32 ); + setFixedBEXTField ( file, chunkPath+".OriginatorReference" , 32 ); + setFixedBEXTField ( file, chunkPath+".OriginationDate" , 10 ); + setFixedBEXTField ( file, chunkPath+".OriginationTime" , 8 ); + + tree->digest32u( file, chunkPath+".TimeReferenceLow", false, true ); // DWORD == 32 Bit + tree->digest32u( file, chunkPath+".TimeReferenceHigh", false, true ); // DWORD == 32 Bit + + tree->digest16u( file, chunkPath+".Version", false, true ); + + // UMID has 64 bytes: + tree->digestString(file, chunkPath+".UMID",64); + //tree->digest32u( file, chunkPath+".UMID_0-4", false, true ); + //tree->setKeyValue( "UMID_5-59" ); + //Skip( file, 64 - 4 - 4 ); + //tree->digest32u( file, chunkPath+".UMID_60-63", false, true ); + + tree->setKeyValue( chunkPath+".Reserved" ); + Skip( file, 190 ); + + if ( BEXT_CodingHistorySize ) + { + setFixedBEXTField ( file, chunkPath+".CodingHistory" , BEXT_CodingHistorySize ); + + //tree->setKeyValue( chunkPath+".CodingHistory" ); // not bothering details. + tree->addComment( "( 0x%llx bytes ) ", BEXT_CodingHistorySize ); + //Skip( file, BEXT_CodingHistorySize ); + } - for ( childOffset = (chunkOffset + 12); childOffset < childLimit; childOffset += (8 + childLen) ) { - childLen = DumpRIFFChunk ( file, childOffset, childLimit, nesting+1 ); + tree->addComment("packet end: 0x%llX", LFA_Tell( file ) ); + tree->popNode(); + } + else + { + Skip( file, chunkSize - 8 ); // skip remainder of chunk ( id, length already digested ) + assertMsg( fromArgs( "inner chunk size too big, curPos:0x%llx, parentEnd:0x%llx", + LFA_Tell(file), + parentEnd ), + LFA_Tell(file) <= parentEnd ); + } + + if ( ! skipper ) + tree->popNode(); } - - if ( childOffset != childLimit ) { - tree->addComment ( "Unexpected end to nested chunks at offset %u (0x%X)", childOffset, childOffset ); + else + { + //dump undefined bytes in LIST + tree->pushNode("** unknown bytes **"); + tree->addOffset( file ); + tree->addComment("size: 0x%llX", chunkSize ); + Skip( file, chunkSize - 8 ); + tree->popNode(); } - } - tree->popNode(); - } else { - tree->pushNode ( "%.4s",chunkID ); - tree->addComment ( "offset 0x%X, size %d%s", - chunkOffset, chunkLen, (((chunkLen & 1) == 0) ? "" : " (has pad byte)") ); - - if ( (chunkLen > 0) && IsKnownRIFFLegacy (chunkID) ) { - CaptureFileData ( file, (chunkOffset + 8), chunkLen ); - std::string legacyField = convert8Bit ( sDataPtr, false, chunkLen ); - tree->changeValue ( legacyField.substr(0,30) + ((legacyField.size()>30)?"...":"") ); - } - if ( (chunkLen > 0) && CheckBytes ( chunkID, "_PMX", 4 ) ) { - tree->pushNode( "XMP packet" ); + if ( LFA_Tell(file) % 2 == 1 ) // if odd file position, add pad byte. { - tree->addComment( "xmp at offset 0x%X, length 0x%X", (chunkOffset + 8), chunkLen ); + XMP_Uns8 padByte = LFA_ReadUns8( file ); + if (!skipper) + { + if ( 0 != padByte ) + tree->addComment(" (non-zero pad byte!)"); + else + tree->addComment(" (pad byte)"); + // Samy: add pad byte comment + } } - tree->popNode(); } - tree->popNode(); - } + } // while - return (chunkLen + (chunkLen & 1)); // Account for the pad byte. } // DumpRIFFChunk // ================================================================================================= static void -DumpRIFF ( LFA_FileRef file, XMP_Uns32 fileLen ) +DumpRIFF ( LFA_FileRef file, XMP_Int64 fileLen ) { - XMP_Uns32 riffLen = DumpRIFFChunk ( file, 0, fileLen, 0 ); - riffLen += 8; // Include the chunk header. - - assertMsg( fromArgs("Unexpected end to RIFF at offset %d (0x%X)", riffLen, riffLen), riffLen == fileLen ); + DumpRIFFChunk ( file, fileLen, "" ); } // ================================================================================================= @@ -3096,7 +3319,7 @@ DumpPNGChunk ( LFA_FileRef file, XMP_Uns32 pngLen, XMP_Uns32 chunkOffset ) XMP_Uns8 * keywordPtr = sDataPtr; size_t keywordLen = strlen ((char*)keywordPtr); - PrintOnlyASCII_8 ( keywordPtr, keywordLen, "," ); + PrintOnlyASCII_8 ( keywordPtr, keywordLen); if ( (chunkType == kPNG_iTXt) && (keywordLen == 17) && CheckBytes (keywordPtr, "XML:com.adobe.xmp", 18) ) { @@ -3501,7 +3724,14 @@ static XMP_Uns32 DumpScriptDataValue ( LFA_FileRef file, XMP_Uns32 sdOffset, boo ReadSDValue ( 2 ); u16 = GetUns16BE ( &buffer[0] ); ReadSDValue ( u16 ); - tree->addComment ( "string (%d) = \"%.*s\"", u16, u16, buffer ); + //Samy: restrict to string under 4k. This was crashing for large XMP packets before, because addComment + // uses a fixed buffer with the size 4096 + if (int(u16) < 4096 ) + { + tree->addComment ( "string (%d) = \"%.*s\"", u16, u16, buffer ); + } else { + tree->addComment ( "string (%d) ", u16 ); + } if ( buffer[u16-1] != 0 ) tree->addComment ( "value lacks trailing nul" ); if ( isOnXMP ) CaptureXMPF ( file, (sdOffset+1+2), u16 ); return (sdOffset + 1 + 2 + u16); @@ -3920,29 +4150,25 @@ PrintID3Encoding ( XMP_Uns8 encoding, XMP_Uns8 * strPtr ) { bool bigEndian = true; - tree->addComment ( "encoding %d", encoding ); - switch ( encoding ) { case 0 : - tree->addComment ( " (ISO Latin-1)" ); + tree->addComment ( "Latin1" ); break; case 1 : if ( *strPtr == 0xFF ) bigEndian = false; - tree->addComment ( " (UTF-16 with BOM, %s endian)", (bigEndian ? "big" : "little") ); + tree->addComment ( "UTF-16 with BOM, %s)", (bigEndian ? "BE" : "LE") ); break; case 2 : - tree->addComment ( " (UTF-16 big endian)" ); + tree->addComment ( "UTF-16 BE" ); break; case 3 : - tree->addComment ( " (UTF-8)" ); + tree->addComment ( "UTF-8" ); break; default : - tree->addComment ( " (** unknown **)" ); + tree->addComment ( "** unknown **" ); break; } - return bigEndian; - } // PrintID3Encoding // ================================================================================================= @@ -3975,19 +4201,44 @@ static void DumpMP3 (LFA_FileRef file, XMP_Uns32 /*mp3Len*/) { // ** We're ignoring the effects of the unsync flag, and not checking the CRC (if present). - assert (sizeof(ID3_Header) == 10); assert (sizeof(ID3_FrameHeader) == 10); - // Dump the basic ID3 header. + // Detect ID3v1 header: + if ( LFA_Measure( file ) > 128 ) + { + LFA_Seek( file, -128, SEEK_END ); + XMP_Uns32 tagID = 0xFFFFFF00 & LFA_ReadUns32_BE( file ); + if ( tagID == 0x54414700 ) // must be "TAG" + { + // Dump ID3v1 header: + tree->pushNode("ID3v1"); + Rewind( file, 1 ); // read one byte too many... + + tree->digestString( file, "ID3v1:title", 30, false, true ); + tree->digestString( file, "ID3v1:artist", 30, false, true ); + tree->digestString( file, "ID3v1:album", 30, false, true ); + tree->digestString( file, "ID3v1:year", 4, false, true ); + tree->digestString( file, "ID3v1:comment", 30, false, true ); + tree->digest( file, "ID3v1:genreNo", 0, 1 ); + + // ID3v1.1 trackNo byte dance: + Rewind( file, 3); + LFA_Tell( file ); + if ( LFA_ReadUns8( file ) == 0 ) + tree->digest( file, "ID3v1:trackNo", 0, 1 ); - ID3_Header id3Head; + tree->popNode(); + } + } + // Dump ID3v2 header: + ID3_Header id3Head; LFA_Seek (file, 0, SEEK_SET); LFA_Read ( file, &id3Head, sizeof(id3Head), true); if (! CheckBytes (id3Head.id3, "ID3", 3)) { - tree->addComment(" No ID3v2 section"); + tree->setKeyValue("No ID3v2 tag"); return; } @@ -3995,8 +4246,8 @@ DumpMP3 (LFA_FileRef file, XMP_Uns32 /*mp3Len*/) XMP_Uns32 framePos = sizeof(id3Head); // The offset of the next (first) ID3 frame. XMP_Uns32 frameEnd = framePos + id3Len; - tree->addComment(" ID3v2.%d.%d, size %d, flags 0x%.2X", - id3Head.vMajor, id3Head.vMinor, id3Len, id3Head.flags); + tree->pushNode("ID3v2.%d.%d, size %d, flags 0x%.2X",id3Head.vMajor, id3Head.vMinor, id3Len, id3Head.flags); + if (id3Head.flags != 0) { tree->addComment("%s%s%s%s", ((id3Head.flags & 0x80) ? ", unsync" : ""), @@ -4005,22 +4256,21 @@ DumpMP3 (LFA_FileRef file, XMP_Uns32 /*mp3Len*/) ((id3Head.flags & 0x10) ? ", has footer" : "")); } - if ((id3Head.vMajor != 3) && (id3Head.vMajor != 4)) { tree->addComment(" ** Unrecognized major version tree."); + tree->popNode(); return; } bool hasExtHeader = ((id3Head.flags & 0x40) != 0); // Dump the extended header if present. - if (hasExtHeader) { - XMP_Uns32 extHeaderLen; - LFA_Read ( file, &extHeaderLen, 4, true ); - extHeaderLen = GetID3Size (id3Head.vMajor, &extHeaderLen); + + extHeaderLen = tree->digest32u(file); + extHeaderLen = GetID3Size (id3Head.vMajor, &extHeaderLen); framePos += (4 + extHeaderLen); if (id3Head.vMajor == 3) { @@ -4035,81 +4285,94 @@ DumpMP3 (LFA_FileRef file, XMP_Uns32 /*mp3Len*/) frameEnd -= padLen; - tree->addComment(" Extended header size %d, flags 0x%.4X, pad size %d", + tree->pushNode("Extended header MajorV3 size %d, flags 0x%.4X, pad size %d", extHeaderLen, extHeaderFlags, padLen); if (extHeaderFlags & 0x8000) { XMP_Uns32 crc; LFA_Read ( file, &crc, 4, true ); crc = GetUns32BE (&crc); - tree->addComment("CRC 0x%.8X", crc); + tree->setKeyValue( "CRC" , fromArgs("0x%.8X", crc) ); } + tree->popNode(); } else if (id3Head.vMajor == 4) { XMP_Uns8 flagCount; LFA_Read ( file, &flagCount, 1, true ); - tree->addComment(" Extended header size %d, flag count %d", extHeaderLen, flagCount); + tree->pushNode("Extended header MajorV4 size %d, flag count %d", extHeaderLen, flagCount); for (size_t i = 0; i < flagCount; ++i) { XMP_Uns8 flag; LFA_Read ( file, &flag, 1, true ); - tree->addComment("0x%.2X", flag); + tree->setKeyValue( fromArgs( "Flag %.2d", flag ) , fromArgs( "0x%.2X", flag ) ); } - + tree->popNode(); + } else + { // ADDED + tree->addComment("unknown major version !"); } } - // Dump the ID3 frames. - + //////////////////////////////////////////////////// + // Dump the ID3 frames while (framePos < frameEnd) { - if ((frameEnd - framePos) < 10) break; // Assume into ID3v2.4 padding. ID3_FrameHeader frameHead; LFA_Seek (file, framePos, SEEK_SET); LFA_Read ( file, &frameHead, sizeof(frameHead), true); - if (CheckBytes (frameHead.id, "\x0\x0\x0\x0", 4)) break; // Assume into ID3v2.4 padding. + if (CheckBytes (frameHead.id, "\x0", 1)) + break; // Assume into ID3v2.4 padding. + // FIXED: there could be just 1 or 2 or 3 bytes of padding total !! frameHead.size = GetID3Size (id3Head.vMajor, &frameHead.size); frameHead.flags = GetUns16BE (&frameHead.flags); - tree->addComment(" '%.4s', offset %d (0x%X), size %d, flags 0x%.2X", - frameHead.id, framePos, framePos, frameHead.size, frameHead.flags); - - if (frameHead.id[0] == 'T') { + tree->setKeyValue( + fromArgs( "ID3v2:%.4s",frameHead.id), + "", //no value yet, tree->changeValue() below + fromArgs( "offset %d (0x%X), size %d, flags 0x%.2X", + framePos, framePos, frameHead.size, frameHead.flags) + ); + if ( frameHead.size == 0 ) + { + // NOTHING TO DO HERE. + // i.e. on 0-byte frames, including known ones... + // ( i.e. the testcase of a (errorneous) TCON 0 byte frame ) + } + else if ( (frameHead.id[0] == 'T') ||(frameHead.id[0] == 'W')) { // Text and URL fields CaptureFileData (file, 0, frameHead.size); - XMP_Uns8 encoding = sDataPtr[0]; - - bool bigEndian = PrintID3Encoding (encoding, (sDataPtr + 1)); + XMP_Uns8 encoding = 0; + XMP_Uns8 skip = 0; + if ( frameHead.id[0] == 'T' ) // URL field has no encoding byte + { + encoding = sDataPtr[0]; + skip = 1; + } + bool bigEndian = PrintID3Encoding (encoding, (sDataPtr + skip)); if ((encoding == 0) || (encoding == 3)) { - PrintOnlyASCII_8 ((sDataPtr + 1), (frameHead.size - 1), " "); + tree->changeValue( convert8Bit( sDataPtr + skip, false, frameHead.size-skip ) ); } else if ((encoding == 1) || (encoding == 2)) { - XMP_Uns16 * valuePtr = (XMP_Uns16*) (sDataPtr + 1); - if (bigEndian) { - PrintOnlyASCII_16BE (valuePtr, (frameHead.size - 1), " "); - } else { - PrintOnlyASCII_16LE (valuePtr, (frameHead.size - 1), " "); - } + tree->changeValue( convert16Bit( bigEndian, sDataPtr + skip, false, (frameHead.size - skip) ) ); } } else if (CheckBytes (frameHead.id, "PRIV", 4) && (frameHead.size >= 4)) { - - CaptureFileData (file, 0, frameHead.size); - PrintOnlyASCII_8 (sDataPtr, strlen((char*)sDataPtr), ","); + // checking on the XMP packet + CaptureFileData (file, 0, frameHead.size); //NB: has side effect: sDataLen, sDataMax, sDataPtr + tree->changeValue( convert8Bit(sDataPtr,false,strlen((char*)sDataPtr) )); if (CheckBytes (sDataPtr, "XMP\x0", 4)) { CaptureXMPF (file, (framePos + 10 + 4), (frameHead.size - 4)); } - } else if (CheckBytes (frameHead.id, "COMM", 4)) { - + } else if (CheckBytes (frameHead.id, "COMM", 4) || CheckBytes (frameHead.id, "USLT", 4)) { CaptureFileData (file, 0, frameHead.size); - XMP_Uns8 * frameEnd = sDataPtr + frameHead.size; + XMP_Uns8 * frameEnd2 = sDataPtr + frameHead.size; XMP_Uns8 encoding = sDataPtr[0]; char * lang = (char*) (sDataPtr + 1); @@ -4118,49 +4381,50 @@ DumpMP3 (LFA_FileRef file, XMP_Uns32 /*mp3Len*/) bool bigEndian = PrintID3Encoding (encoding, (sDataPtr + 4)); if ((encoding == 0) || (encoding == 3)) { - XMP_Uns8 * descrPtr = sDataPtr + 4; XMP_Uns8 * valuePtr = descrPtr; - while (*valuePtr != 0) ++valuePtr; + + while (*valuePtr != 0) + { + ++valuePtr; + } ++valuePtr; - PrintOnlyASCII_8 (descrPtr, (valuePtr - descrPtr - 1), " "); - PrintOnlyASCII_8 (valuePtr, (frameEnd - valuePtr), " "); + size_t descrBytes = valuePtr - descrPtr - 1; + tree->changeValue( convert8Bit( valuePtr, false, frameEnd2 - valuePtr ) ); + if (CheckBytes (frameHead.id, "COMM", 4)) { + tree->setKeyValue("ID3v2:COMM-descr", convert8Bit( descrPtr, false, descrBytes ).c_str() ); + } } else if ((encoding == 1) || (encoding == 2)) { XMP_Uns16 * descrPtr = (XMP_Uns16*) (sDataPtr + 4); XMP_Uns16 * valuePtr = descrPtr; - while (*valuePtr != 0) ++valuePtr; + while (*valuePtr != 0) + { + ++valuePtr; + } ++valuePtr; size_t descrBytes = 2 * (valuePtr - descrPtr - 1); - size_t valueBytes = 2 * ((XMP_Uns16*)frameEnd - valuePtr); - - if (bigEndian) { - PrintOnlyASCII_16BE (descrPtr, descrBytes, " "); - PrintOnlyASCII_16BE (valuePtr, valueBytes, " "); - } else { - PrintOnlyASCII_16LE (descrPtr, descrBytes, " "); - PrintOnlyASCII_16LE (valuePtr, valueBytes, " "); - } + size_t valueBytes = 2 * ((XMP_Uns16*)frameEnd2 - valuePtr); + tree->changeValue( convert16Bit( bigEndian, (XMP_Uns8*) valuePtr, false, valueBytes ) ); + tree->setKeyValue("ID3v2:COMM-descr", convert16Bit( bigEndian, (XMP_Uns8*) descrPtr, false, descrBytes ).c_str() ); } - } - framePos += (sizeof(ID3_FrameHeader) + frameHead.size); - } if (framePos < frameEnd) { - tree->addComment(" Padding assumed, offset %d (0x%X), size %d", - framePos, framePos, (frameEnd - framePos)); + tree->setKeyValue("","", + fromArgs("Padding assumed, offset %d (0x%X), size %d",framePos, framePos, (frameEnd - framePos)) + ); } if (sXMPPtr != 0) DumpXMP ("ID3 'PRIV' \"XMP\" frame"); - + tree->popNode(); } // DumpMP3 // ================================================================================================= @@ -4170,7 +4434,7 @@ PacketScan (LFA_FileRef file, XMP_Int64 fileLen) { try { - QEScanner scanner (fileLen); + XMPScanner scanner (fileLen); LFA_Seek (file, 0, SEEK_SET); XMP_Uns8 buffer [64*1024]; @@ -4183,13 +4447,13 @@ PacketScan (LFA_FileRef file, XMP_Int64 fileLen) } size_t snipCount = scanner.GetSnipCount(); - QEScanner::SnipInfoVector snips (snipCount); + XMPScanner::SnipInfoVector snips (snipCount); scanner.Report (snips); size_t packetCount = 0; for (size_t s = 0; s < snipCount; ++s) { - if (snips[s].fState == QEScanner::eValidPacketSnip) { + if (snips[s].fState == XMPScanner::eValidPacketSnip) { ++packetCount; CaptureXMPF (file, (XMP_Uns32)snips[s].fOffset, (XMP_Uns32)snips[s].fLength); DumpXMP ("packet scan"); @@ -4288,9 +4552,15 @@ namespace DumpFile_NS { // ------------------------------------------------------------------------------------------------- -void DumpFile::Scan (std::string filename, TagTree &tagTree) +void DumpFile::Scan (std::string filename, TagTree &tagTree, bool resetTree) { DumpFile_NS::selfTest(); //calls selftest (will happen only once per runtime, optimization done) + + if( resetTree ) + { + tagTree.reset(); + } + tree = &tagTree; // static "global" helper to avoid looping throug 'tree' 24x7 // Read the first 4K of the file into a local buffer and determine the file format. @@ -4326,7 +4596,7 @@ void DumpFile::Scan (std::string filename, TagTree &tagTree) if ( format == kXMP_JPEGFile ) { tagTree.pushNode ( "Dumping JPEG file" ); - tagTree.addComment ( "size %d(0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); { fileContent = (XMP_Uns8*) malloc(fileLen); LFA_Seek ( fileRef, 0, SEEK_SET ); @@ -4338,7 +4608,7 @@ void DumpFile::Scan (std::string filename, TagTree &tagTree) } else if ( format == kXMP_PhotoshopFile ) { tagTree.pushNode ( "Dumping Photoshop file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); { fileContent = (XMP_Uns8*) malloc(fileLen); LFA_Seek ( fileRef, 0, SEEK_SET ); @@ -4350,91 +4620,84 @@ void DumpFile::Scan (std::string filename, TagTree &tagTree) } else if ( format == kXMP_TIFFFile ) { tagTree.pushNode ( "Dumping TIFF file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); { fileContent = (XMP_Uns8*) malloc(fileLen); LFA_Seek ( fileRef, 0, SEEK_SET ); LFA_Read ( fileRef, fileContent, fileLen, true); - DumpTIFF ( fileContent, fileLen, 0, "TIFF file" ); + DumpTIFF ( fileContent, fileLen, 0, "TIFF file" , ""); } tagTree.popNode(); - } else if ( format == kXMP_JPEG2KFile ) { - - tagTree.pushNode ( "Dumping JPEG 2000 file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); - DumpJPEG2000 (fileRef, fileLen ); - tagTree.popNode(); - } else if ( format == kXMP_WMAVFile ) { tagTree.pushNode ( "Dumping ASF (WMA/WMV) file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpASF ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_AVIFile ) { tagTree.pushNode ( "Dumping AVI file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpRIFF ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_WAVFile ) { tagTree.pushNode ( "Dumping WAV file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpRIFF ( fileRef, fileLen ); tagTree.popNode(); - } else if ( format == kXMP_MPEG4File || format == kXMP_MOVFile ) { - - bool isQT = (format == kXMP_MOVFile); - - tagTree.pushNode ( "Dumping ISO (MPEG-4 or Quicktime) file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); - DumpISO( fileRef, fileLen, isQT ); + } else if ( format == kXMP_MPEG4File || format == kXMP_MOVFile || format == kXMP_JPEG2KFile ) { + // all ISO formats ( MPEG4, MOV, JPEG2000) handled jointly, + // - no longer relying on any advance "isQT" flagging + tagTree.pushNode ( "ISO file" ); + Log::info("size: %d", fileLen ); + tagTree.addComment ( "size %I64d (0x%I64X)", fileLen, fileLen ); + DumpISO( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_PNGFile ) { tagTree.pushNode ( "Dumping PNG file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpPNG ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_InDesignFile ) { tagTree.pushNode ( "Dumping InDesign file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpInDesign ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_SWFFile ) { tagTree.pushNode ( "Dumping SWF file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpSWF ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_FLVFile ) { tagTree.pushNode ( "Dumping FLV file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpFLV ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_MP3File ) { tagTree.pushNode ( "Dumping MP3 file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpMP3 ( fileRef, fileLen ); tagTree.popNode(); } else if ( format == kXMP_UCFFile ) { tagTree.pushNode ( "Dumping UCF (Universal Container Format) file" ); - tagTree.addComment ( "size %d (0x%X)", fileLen, fileLen ); + tagTree.addComment ( "size %lld (0x%llx)", fileLen, fileLen ); DumpUCF ( fileRef, fileLen ); tagTree.popNode(); @@ -4456,3 +4719,12 @@ void DumpFile::Scan (std::string filename, TagTree &tagTree) LFA_Close(fileRef); } // DumpFile + + +void DumpFile::dumpFile( std::string filename ) +{ + TagTree localTree; + DumpFile::Scan( filename , localTree ); // (important test in itself for validity) + localTree.dumpTree(); +} + diff --git a/samples/source/common/DumpFile.h b/samples/source/common/DumpFile.h index 05857f8..d5ef41f 100644 --- a/samples/source/common/DumpFile.h +++ b/samples/source/common/DumpFile.h @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2002-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -14,7 +14,10 @@ class DumpFile { public: - static void Scan(std::string filename, TagTree &tagTree); + static void Scan( std::string filename, TagTree &tagTree, bool resetTree = true ); + + /* dumps file to output, no strings attached Log::info() */ + static void dumpFile( std::string filename ); }; #endif diff --git a/samples/source/common/Log.cpp b/samples/source/common/Log.cpp index fe132ba..d75336e 100644 --- a/samples/source/common/Log.cpp +++ b/samples/source/common/Log.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -35,6 +35,12 @@ FILE* Log::logfile=NULL; bool Log::noErrorsOrWarnings=true; bool Log::mute=false; bool Log::muteWarnings=false; + +// strings for skipped test information +std::string Log::skippedTestsAll; +std::string Log::skippedTestsWin; +std::string Log::skippedTestsMac; +std::string Log::skippedTestsUnix; ////////////////////////////////////////////////////////////////////////////////////// /* standard log, use filename=NULL or filename="" or do not construct at all @@ -184,6 +190,43 @@ void Log::important(std::string s1) info(s1.c_str()); } +void Log::error( std::string s1 ) +{ + error( s1.c_str() ); +} + +/////////////////////////////////////////////////////////////////////////////// +// skipped tests +void Log::printSkippedTest(const char* filename) +{ + std::string completeString = "<html><head><title>Skipped Tests:</title><head><body>"; + if (!skippedTestsAll.empty()) + completeString.append("<br><br><table border=\"1\"><caption>all Plattforms:</caption><tr><th>Path to Test</th><th>comment</th></tr>"+skippedTestsAll + "</table>"); + if (!skippedTestsWin.empty()) + completeString.append("<br><br><table border=\"1\"><caption>Windows:</caption><tr><th>Path to Test</th><th>comment</th></tr>"+skippedTestsWin+ "</table>"); + if (!skippedTestsMac.empty()) + completeString.append("<br><br><table border=\"1\"><caption>Macintosh:</caption><tr><th>Path to Test</th><th>comment</th></tr>"+skippedTestsMac+ "</table>"); + if (!skippedTestsUnix.empty()) + completeString.append("<br><br><table border=\"1\"><caption>UNIX:</caption><tr><th>Path to Test</th><th>comment</th></tr>"+skippedTestsUnix+ "</table>"); + + completeString.append("</body></html>"); + // print to console if mute is off + /*if (!Log::mute) + printf(completeString.c_str());*/ + + //print result to file + FILE* outputFile=fopen(filename,"wt"); + if (!outputFile) { + printf("could not open output file %s for writing",filename); //do in addition to make sure it's output + throw LOG_Exception("could not open output file for writing"); + } + fprintf( outputFile, "%s", completeString.c_str()); + if(outputFile) { + fclose(outputFile); + } + +} + ///////////////////////////////////////////////////////////////////////////////// // mute vs. verbose @@ -206,4 +249,3 @@ void Log::setVerboseWarnings() { Log::muteWarnings = false; } - diff --git a/samples/source/common/Log.h b/samples/source/common/Log.h index 3bd370c..db8d35c 100644 --- a/samples/source/common/Log.h +++ b/samples/source/common/Log.h @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -60,6 +60,7 @@ class Log { //convenience overloads: static void info(std::string s1); static void important(std::string s1); + static void error( std::string s1 ); // mute vs. verbose static void setMute(); @@ -72,7 +73,16 @@ class Log { // NB: public on intend, such that i.e. copy commands can decide to // send their output to nirvana accordingly on mute==true; // (getter/setter on a static boolean var is fairly pointless) - + + // strings to store the skipped test for different plattforms + static std::string skippedTestsAll; + static std::string skippedTestsWin; + static std::string skippedTestsMac; + static std::string skippedTestsUnix; + + // function to print the skipped tests on console and into a log file + static void printSkippedTest(const char* filename); + private: static bool singletonInstanciated; static FILE* logfile; @@ -80,6 +90,8 @@ class Log { //should only be temporarily used, i.e. on selftests that //inevitable yield (during selftet: bogus) warnings static bool muteWarnings; + + }; #endif diff --git a/samples/source/common/OutputUtils.cpp b/samples/source/common/OutputUtils.cpp deleted file mode 100644 index 0c59339..0000000 --- a/samples/source/common/OutputUtils.cpp +++ /dev/null @@ -1,227 +0,0 @@ -// ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// -// ================================================================================================= - -#include <string> -#include <stdarg.h> - -#include "OutputUtils.h" -#include "Log.h" - - -using namespace std; - -namespace Utils { - /* hand over the fileFormat and get a fileFormatName (2-3 Letters ie. PDF) - * and a Description (i.e. "Portable Document Format") back. - * see also public/include/XMP_Const.h - */ - void fileFormatNameToString ( XMP_FileFormat fileFormat, std::string & fileFormatName, std::string & fileFormatDesc ) - { - fileFormatName.erase(); - fileFormatDesc.erase(); - switch(fileFormat) { - case kXMP_PDFFile: - fileFormatName = "PDF ";fileFormatDesc = "Portable Document Format";break; - case kXMP_PostScriptFile: - fileFormatName = "PS ";fileFormatDesc = "Post Script";break; - case kXMP_EPSFile: - fileFormatName = "EPS ";fileFormatDesc = "Encapsulated Post Script";break; - case kXMP_JPEGFile: - fileFormatName = "JPEG";fileFormatDesc = "Joint Photographic Experts Group";break; - case kXMP_JPEG2KFile: - fileFormatName = "JPX ";fileFormatDesc = "JPEG 2000";break; - case kXMP_TIFFFile: - fileFormatName = "TIFF";fileFormatDesc = "Tagged Image File Format";break; - case kXMP_GIFFile: - fileFormatName = "GIF ";fileFormatDesc = "Graphics Interchange Format";break; - case kXMP_PNGFile: - fileFormatName = "PNG ";fileFormatDesc = "Portable Network Graphic";break; - case kXMP_MOVFile: - fileFormatName = "MOV ";fileFormatDesc = "Quicktime";break; - case kXMP_AVIFile: - fileFormatName = "AVI ";fileFormatDesc = "Quicktime";break; - case kXMP_CINFile: - fileFormatName = "CIN ";fileFormatDesc = "Cineon";break; - case kXMP_WAVFile: - fileFormatName = "WAV ";fileFormatDesc = "WAVE Form Audio Format";break; - case kXMP_MP3File: - fileFormatName = "MP3 ";fileFormatDesc = "MPEG-1 Audio Layer 3";break; - case kXMP_SESFile: - fileFormatName = "SES ";fileFormatDesc = "Audition session";break; - case kXMP_CELFile: - fileFormatName = "CEL ";fileFormatDesc = "Audition loop";break; - case kXMP_MPEGFile: - fileFormatName = "MPEG";fileFormatDesc = "Motion Pictures Experts Group";break; - case kXMP_MPEG2File: - fileFormatName = "MP2 ";fileFormatDesc = "MPEG-2";break; - case kXMP_WMAVFile: - fileFormatName = "WMAV";fileFormatDesc = "Windows Media Audio and Video";break; - case kXMP_HTMLFile: - fileFormatName = "HTML";fileFormatDesc = "HyperText Markup Language";break; - case kXMP_XMLFile: - fileFormatName = "XML ";fileFormatDesc = "Extensible Markup Language";break; - case kXMP_TextFile: - fileFormatName = "TXT ";fileFormatDesc = "text";break; - case kXMP_PhotoshopFile: - fileFormatName = "PSD ";fileFormatDesc = "Photoshop Document";break; - case kXMP_IllustratorFile: - fileFormatName = "AI ";fileFormatDesc = "Adobe Illustrator";break; - case kXMP_InDesignFile: - fileFormatName = "INDD";fileFormatDesc = "Adobe InDesign";break; - case kXMP_AEProjectFile: - fileFormatName = "AEP ";fileFormatDesc = "AfterEffects Project";break; - case kXMP_AEFilterPresetFile: - fileFormatName = "FFX ";fileFormatDesc = "AfterEffects Filter Preset";break; - case kXMP_EncoreProjectFile: - fileFormatName = "NCOR";fileFormatDesc = "Encore Project";break; - case kXMP_PremiereProjectFile: - fileFormatName = "PPRJ";fileFormatDesc = "Premier Project";break; - case kXMP_SWFFile: - fileFormatName = "SWF ";fileFormatDesc = "Shockwave Flash";break; - case kXMP_PremiereTitleFile: - fileFormatName = "PRTL";fileFormatDesc = "Premier Title";break; - default: - fileFormatName = " ";fileFormatDesc = "Unkown file format";break; - } - } //fileFormatNameToString - - /** - * GetFormatInfo-Flags (aka Handler-Flags) - * find this in XMP_Const.h under "Options for GetFormatInfo" - */ - std::string XMPFiles_FormatInfoToString ( const XMP_OptionBits options) { - std::string outString; - - if( options & kXMPFiles_CanInjectXMP ) - outString.append(" CanInjectXMP"); - if( options & kXMPFiles_CanExpand ) - outString.append(" CanExpand"); - if( options & kXMPFiles_CanRewrite ) - outString.append(" CanRewrite"); - if( options & kXMPFiles_PrefersInPlace ) - outString.append(" PrefersInPlace"); - if( options & kXMPFiles_CanReconcile ) - outString.append(" CanReconcile"); - if( options & kXMPFiles_AllowsOnlyXMP ) - outString.append(" AllowsOnlyXMP"); - if( options & kXMPFiles_ReturnsRawPacket ) - outString.append(" ReturnsRawPacket"); - if( options & kXMPFiles_ReturnsTNail ) - outString.append(" ReturnsTNail"); - if( options & kXMPFiles_HandlerOwnsFile ) - outString.append(" HandlerOwnsFile"); - if( options & kXMPFiles_AllowsSafeUpdate ) - outString.append(" AllowsSafeUpdate"); - - if (outString.empty()) outString=" (none)"; - return outString; - } - - /** - * openOptions to String, find this in - * XMP_Const.h under "Options for OpenFile" - */ - std::string XMPFiles_OpenOptionsToString ( const XMP_OptionBits options) { - std::string outString; - if( options & kXMPFiles_OpenForRead) - outString.append(" OpenForRead"); - if( options & kXMPFiles_OpenForUpdate) - outString.append(" OpenForUpdate"); - if( options & kXMPFiles_OpenOnlyXMP) - outString.append(" OpenOnlyXMP"); - if( options & kXMPFiles_OpenCacheTNail) - outString.append(" OpenCacheTNail"); - if( options & kXMPFiles_OpenStrictly) - outString.append(" OpenStrictly"); - if( options & kXMPFiles_OpenUseSmartHandler) - outString.append(" OpenUseSmartHandler"); - if( options & kXMPFiles_OpenUsePacketScanning) - outString.append(" OpenUsePacketScanning"); - if( options & kXMPFiles_OpenLimitedScanning) - outString.append(" OpenLimitedScanning"); - - if (outString.empty()) outString=" (none)"; - return outString; - } - - - - std::string fromArgs(const char* format, ...) - { - //note: format and ... are somehow "used up", i.e. dumping them - // via vsprintf _and_ via printf brought up errors on Mac (only) - // i.e. %d %X stuff looking odd (roughly like signed vs unsigned...) - // buffer reuse is fine, just dont use format/... twice. - - char buffer[4096]; //should be big enough but no guarantees.. - va_list args; - va_start(args, format); - vsprintf(buffer, format, args); - va_end(args); - - return std::string(buffer); - } - - - - //just the callback-Function - XMP_Status DumpToString( void * refCon, XMP_StringPtr outStr, XMP_StringLen outLen ) - { - XMP_Status status = 0; - std::string* dumpString = static_cast < std::string* > ( refCon ); - dumpString->append (outStr, outLen); // less effective: ( std::string(outStr).substr(0,outLen) ); - return status; - } - - //full output to Log - void dumpXMP(SXMPMeta* xmpMeta){ - std::string dump; - xmpMeta->Sort(); - xmpMeta->DumpObject(DumpToString, &dump); - Log::info("%s",dump.c_str()); - } - - //same but with Comment - void dumpXMP(std::string comment, SXMPMeta* xmpMeta){ - Log::info("%s",comment.c_str()); - std::string dump; - xmpMeta->Sort(); - xmpMeta->DumpObject(DumpToString, &dump); - Log::info("%s",dump.c_str()); - } - - // dump Aliases **************************** - //full output to Log - void dumpAliases() - { - std::string dump; - SXMPMeta::DumpAliases(DumpToString,&dump); - Log::info("List of aliases:"); - Log::info("%s",dump.c_str()); - } - - void DumpXMP_DateTime( const XMP_DateTime & date ) { - static std::string tempStr; - tempStr.clear(); - SXMPUtils::ConvertFromDate( date, &tempStr ); - Log::info("%s\nYear: '%d' Month: '%d' Day: '%d' Hour: '%d' Min: '%d' Sec: '%d' Nanos: '%d' TZSign: '%d' TZHour: '%d' TZMinute: '%d'\n", tempStr.c_str(), date.year, date.month, date.day, date.hour, date.minute, date.second, date.nanoSecond, date.tzSign, date.tzHour, date.tzMinute ); - } - - - - - void removeNLines(std::string* s, int n){ - for(int i=0;i<n;i++) - s->erase(0,s->find_first_of("\n")+1); - } - - - -} /*namespace Utils*/ - diff --git a/samples/source/common/OutputUtils.h b/samples/source/common/OutputUtils.h deleted file mode 100644 index 6597ed8..0000000 --- a/samples/source/common/OutputUtils.h +++ /dev/null @@ -1,50 +0,0 @@ -// ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// -// ================================================================================================= - -#ifndef __XMPQE_QEOUTPUTUTILS_h__ -#define __XMPQE_QEOUTPUTUTILS_h__ 1 - -//XMP related -#define TXMP_STRING_TYPE std::string - -#include "XMP.hpp" - -namespace Utils{ - - - void fileFormatNameToString( XMP_FileFormat fileFormat, std::string &fileFormatName, std::string &fileFormatDesc ); - - std::string XMPFiles_FormatInfoToString ( const XMP_OptionBits options); - std::string XMPFiles_OpenOptionsToString ( const XMP_OptionBits options); - - std::string fromArgs(const char* format, ...); - - //just the DumpToString Callback routine - // - (taken over from dfranzen resp. DevTestProto) - // - /*outLen*/ ==> avoid unused warning - XMP_Status DumpToString ( void * refCon, XMP_StringPtr outStr, XMP_StringLen /*outLen*/ ); - - // dumps xmp (using Log::info()) - // based on dfranzen's conglomerate of dump-functions in XMPTestCase.cpp - // no return value, will Log::error() on problems - void dumpXMP(SXMPMeta* xmpMeta); - void dumpXMP(std::string comment, SXMPMeta* xmpMeta); - void dumpAliases(); - - //remove n first lines from a string - // handy i.e. for first line of xmp-rdf packets prior to KGO, - // since firt line may contain -date- or actual date, - // (debug) or 7 spaces, etc... - void removeNLines(std::string* s, int n=1); - - void DumpXMP_DateTime( const XMP_DateTime & date ); - -} - -#endif diff --git a/samples/source/common/QEScanner.cpp b/samples/source/common/QEScanner.cpp deleted file mode 100644 index 5553495..0000000 --- a/samples/source/common/QEScanner.cpp +++ /dev/null @@ -1,1469 +0,0 @@ -// ================================================================================================= -// Copyright 2002-2006 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// -// Adobe patent application tracking #P435, entitled 'Unique markers to simplify embedding data of -// one format in a file with a different format', inventors: Sean Parent, Greg Gilley. -// ================================================================================================= - -#if WIN_ENV - #pragma warning ( disable : 4786 ) // The VC++ debugger can't handle long symbol names. - #pragma warning ( disable : 4127 ) // conditional expression is constant -#endif - - -#include "QEScanner.hpp" - -#include <cassert> -#include <string> -#include <iostream> -#include <iomanip> -#include <fstream> -#include <cstdlib> - - -#ifndef UseStringPushBack // VC++ 6.x does not provide push_back for strings! - #define UseStringPushBack 0 -#endif - - -using namespace std; - - -// *** Consider Boyer-Moore style search for "<?xpacket begin=". It isn't an obvious win, the -// *** additional code might be slower than scanning every character. Especially if we will -// *** read every cache line anyway. - - -// ================================================================================================= -// ================================================================================================= -// class PacketMachine -// =================== -// -// This is the packet recognizer state machine. The top of the machine is FindNextPacket, this -// calls the specific state components and handles transitions. The states are described by an -// array of RecognizerInfo records, indexed by the RecognizerKind enumeration. Each RecognizerInfo -// record has a function that does that state's work, the success and failure transition states, -// and a string literal that is passed to the state function. The literal lets a common MatchChar -// or MatchString function be used in several places. -// -// The state functions are responsible for consuming input to recognize their particular state. -// This includes intervening nulls for 16 and 32 bit character forms. For the simplicity, things -// are treated as essentially little endian and the nulls are not actually checked. The opening -// '<' is found with a byte-by-byte search, then the number of bytes per character is determined -// by counting the following nulls. From then on, consuming a character means incrementing the -// buffer pointer by the number of bytes per character. Thus the buffer pointer only points to -// the "real" bytes. This also means that the pointer can go off the end of the buffer by a -// variable amount. The amount of overrun is saved so that the pointer can be positioned at the -// right byte to start the next buffer. -// -// The state functions return a TriState value, eTriYes means the pattern was found, eTriNo means -// the pattern was definitely not found, eTriMaybe means that the end of the buffer was reached -// while working through the pattern. -// -// When eTriYes is returned, the fBufferPtr data member is left pointing to the "real" byte -// following the last actual byte. Which might not be addressable memory! This also means that -// a state function can be entered with nothing available in the buffer. When eTriNo is returned, -// the fBufferPtr data member is left pointing to the byte that caused the failure. The state -// machine starts over from the failure byte. -// -// The state functions must preserve their internal micro-state before returning eTriMaybe, and -// resume processing when called with the next buffer. The fPosition data member is used to denote -// how many actual characters have been consumed. The fNullCount data member is used to denote how -// many nulls are left before the next actual character. - - -// ================================================================================================= -// PacketMachine -// ============= - -QEScanner::PacketMachine::PacketMachine ( SInt64 bufferOffset, const void * bufferOrigin, SInt64 bufferLength ) : - - // Public members - fPacketStart ( 0 ), - fPacketLength ( 0 ), - fBytesAttr ( -1 ), - fCharForm ( eChar8Bit ), - fAccess ( ' ' ), - fBogusPacket ( false ), - - // Private members - fBufferOffset ( bufferOffset ), - fBufferOrigin ( (const char *) bufferOrigin ), - fBufferPtr ( fBufferOrigin ), - fBufferLimit ( fBufferOrigin + bufferLength ), - fRecognizer ( eLeadInRecognizer ), - fPosition ( 0 ), - fBytesPerChar ( 1 ), - fBufferOverrun ( 0 ), - fQuoteChar ( ' ' ) - -{ - /* - REVIEW NOTES : Should the buffer stuff be in a class? - */ - - assert ( bufferOrigin != NULL ); - assert ( bufferLength != 0 ); - -} // PacketMachine - - -// ================================================================================================= -// ~PacketMachine -// ============== - -QEScanner::PacketMachine::~PacketMachine () -{ - - // An empty placeholder. - -} // ~PacketMachine - - -// ================================================================================================= -// AssociateBuffer -// =============== - -void -QEScanner::PacketMachine::AssociateBuffer ( SInt64 bufferOffset, const void * bufferOrigin, SInt64 bufferLength ) -{ - - fBufferOffset = bufferOffset; - fBufferOrigin = (const char *) bufferOrigin; - fBufferPtr = fBufferOrigin + fBufferOverrun; - fBufferLimit = fBufferOrigin + bufferLength; - -} // AssociateBuffer - - -// ================================================================================================= -// ResetMachine -// ============ - -void -QEScanner::PacketMachine::ResetMachine () -{ - - fRecognizer = eLeadInRecognizer; - fPosition = 0; - fBufferOverrun = 0; - fCharForm = eChar8Bit; - fBytesPerChar = 1; - fAccess = ' '; - fBytesAttr = -1; - fBogusPacket = false; - - fAttrName.erase ( fAttrName.begin(), fAttrName.end() ); - fAttrValue.erase ( fAttrValue.begin(), fAttrValue.end() ); - fEncodingAttr.erase ( fEncodingAttr.begin(), fEncodingAttr.end() ); - -} // ResetMachine - - -// ================================================================================================= -// FindLessThan -// ============ - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::FindLessThan ( PacketMachine * ths, const char * which ) -{ - - if ( *which == 'H' ) { - - // -------------------------------------------------------------------------------- - // We're looking for the '<' of the header. If we fail there is no packet in this - // part of the input, so return eTriNo. - - ths->fCharForm = eChar8Bit; // We might have just failed from a bogus 16 or 32 bit case. - ths->fBytesPerChar = 1; - - while ( ths->fBufferPtr < ths->fBufferLimit ) { // Don't skip nulls for the header's '<'! - if ( *ths->fBufferPtr == '<' ) break; - ths->fBufferPtr++; - } - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriNo; - ths->fBufferPtr++; - return eTriYes; - - } else { - - // -------------------------------------------------------------------------------- - // We're looking for the '<' of the trailer. We're already inside the packet body, - // looking for the trailer. So here if we fail we must return eTriMaybe so that we - // keep looking for the trailer in the next buffer. - - const int bytesPerChar = ths->fBytesPerChar; - - while ( ths->fBufferPtr < ths->fBufferLimit ) { - if ( *ths->fBufferPtr == '<' ) break; - ths->fBufferPtr += bytesPerChar; - } - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - - } - -} // FindLessThan - - -// ================================================================================================= -// MatchString -// =========== - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::MatchString ( PacketMachine * ths, const char * literal ) -{ - const int bytesPerChar = ths->fBytesPerChar; - const char * litPtr = literal + ths->fPosition; - const size_t charsToGo = strlen ( literal ) - ths->fPosition; - size_t charsDone = 0; - - while ( (charsDone < charsToGo) && (ths->fBufferPtr < ths->fBufferLimit) ) { - if ( *litPtr != *ths->fBufferPtr ) return eTriNo; - charsDone++; - litPtr++; - ths->fBufferPtr += bytesPerChar; - } - - if ( charsDone == charsToGo ) return eTriYes; - ths->fPosition += charsDone; - return eTriMaybe; - -} // MatchString - - -// ================================================================================================= -// MatchChar -// ========= - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::MatchChar ( PacketMachine * ths, const char * literal ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - if ( currChar != *literal ) return eTriNo; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - -} // MatchChar - - -// ================================================================================================= -// MatchOpenQuote -// ============== - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::MatchOpenQuote ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - if ( (currChar != '\'') && (currChar != '"') ) return eTriNo; - ths->fQuoteChar = currChar; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - -} // MatchOpenQuote - - -// ================================================================================================= -// MatchCloseQuote -// =============== - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::MatchCloseQuote ( PacketMachine * ths, const char * /* unused */ ) -{ - - return MatchChar ( ths, &ths->fQuoteChar ); - -} // MatchCloseQuote - - -// ================================================================================================= -// CaptureAttrName -// =============== - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::CaptureAttrName ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - char currChar; - - if ( ths->fPosition == 0 ) { // Get the first character in the name. - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - currChar = *ths->fBufferPtr; - if ( ths->fAttrName.size() == 0 ) { - if ( ! ( ( ('a' <= currChar) && (currChar <= 'z') ) || - ( ('A' <= currChar) && (currChar <= 'Z') ) || - (currChar == '_') || (currChar == ':') ) ) { - return eTriNo; - } - } - - ths->fAttrName.erase ( ths->fAttrName.begin(), ths->fAttrName.end() ); - #if UseStringPushBack - ths->fAttrName.push_back ( currChar ); - #else - ths->fAttrName.insert ( ths->fAttrName.end(), currChar ); - #endif - ths->fBufferPtr += bytesPerChar; - - } - - while ( ths->fBufferPtr < ths->fBufferLimit ) { // Get the remainder of the name. - - currChar = *ths->fBufferPtr; - if ( ! ( ( ('a' <= currChar) && (currChar <= 'z') ) || - ( ('A' <= currChar) && (currChar <= 'Z') ) || - ( ('0' <= currChar) && (currChar <= '9') ) || - (currChar == '-') || (currChar == '.') || (currChar == '_') || (currChar == ':') ) ) { - break; - } - - #if UseStringPushBack - ths->fAttrName.push_back ( currChar ); - #else - ths->fAttrName.insert ( ths->fAttrName.end(), currChar ); - #endif - ths->fBufferPtr += bytesPerChar; - - } - - if ( ths->fBufferPtr < ths->fBufferLimit ) return eTriYes; - ths->fPosition = ths->fAttrName.size(); // The name might span into the next buffer. - return eTriMaybe; - -} // CaptureAttrName - - -// ================================================================================================= -// CaptureAttrValue -// ================ -// -// Recognize the equal sign and the quoted string value, capture the value along the way. - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::CaptureAttrValue ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - char currChar = 0; - TriState result = eTriMaybe; - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - switch ( ths->fPosition ) { - - case 0 : // The name should haved ended at the '=', nulls already skipped. - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - if ( *ths->fBufferPtr != '=' ) return eTriNo; - ths->fBufferPtr += bytesPerChar; - ths->fPosition = 1; - // fall through OK because MatchOpenQuote will check the buffer limit and nulls ... - - case 1 : // Look for the open quote. - - result = MatchOpenQuote ( ths, NULL ); - if ( result != eTriYes ) return result; - ths->fPosition = 2; - // fall through OK because the buffer limit and nulls are checked below ... - - default : // Look for the close quote, capturing the value along the way. - - assert ( ths->fPosition == 2 ); - - const char quoteChar = ths->fQuoteChar; - - while ( ths->fBufferPtr < ths->fBufferLimit ) { - currChar = *ths->fBufferPtr; - if ( currChar == quoteChar ) break; - #if UseStringPushBack - ths->fAttrValue.push_back ( currChar ); - #else - ths->fAttrValue.insert ( ths->fAttrValue.end(), currChar ); - #endif - ths->fBufferPtr += bytesPerChar; - } - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - assert ( currChar == quoteChar ); - ths->fBufferPtr += bytesPerChar; // Advance past the closing quote. - return eTriYes; - - } - -} // CaptureAttrValue - - -// ================================================================================================= -// RecordStart -// =========== -// -// Note that this routine looks at bytes, not logical characters. It has to figure out how many -// bytes per character there are so that the other recognizers can skip intervening nulls. - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::RecordStart ( PacketMachine * ths, const char * /* unused */ ) -{ - - while ( true ) { - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currByte = *ths->fBufferPtr; - - switch ( ths->fPosition ) { - - case 0 : // Record the length. - assert ( ths->fCharForm == eChar8Bit ); - assert ( ths->fBytesPerChar == 1 ); - ths->fPacketStart = ths->fBufferOffset + ((ths->fBufferPtr - 1) - ths->fBufferOrigin); - ths->fPacketLength = 0; - ths->fPosition = 1; - // ! OK to fall through here, we didn't consume a byte in this step. - - case 1 : // Look for the first null byte. - if ( currByte != 0 ) return eTriYes; // No nulls found. - ths->fCharForm = eChar16BitBig; // Assume 16 bit big endian for now. - ths->fBytesPerChar = 2; - ths->fBufferPtr++; - ths->fPosition = 2; - break; // ! Don't fall through, have to check for the end of the buffer between each byte. - - case 2 : // One null was found, look for a second. - if ( currByte != 0 ) return eTriYes; // Just one null found. - ths->fBufferPtr++; - ths->fPosition = 3; - break; - - case 3 : // Two nulls were found, look for a third. - if ( currByte != 0 ) return eTriNo; // Just two nulls is not valid. - ths->fCharForm = eChar32BitBig; // Assume 32 bit big endian for now. - ths->fBytesPerChar = 4; - ths->fBufferPtr++; - return eTriYes; - break; - - } - - } - -} // RecordStart - - -// ================================================================================================= -// RecognizeBOM -// ============ -// -// Recognizing the byte order marker is a surprisingly messy thing to do. It can't be done by the -// normal string matcher, there are no intervening nulls. There are 4 transitions after the opening -// quote, the closing quote or one of the three encodings. For the actual BOM there are then 1 or 2 -// following bytes that depend on which of the encodings we're in. Not to mention that the buffer -// might end at any point. -// -// The intervening null count done earlier determined 8, 16, or 32 bits per character, but not the -// big or little endian nature for the 16/32 bit cases. The BOM must be present for the 16 and 32 -// bit cases in order to determine the endian mode. There are six possible byte sequences for the -// quoted BOM string, ignoring the differences for quoting with ''' versus '"'. -// -// Keep in mind that for the 16 and 32 bit cases there will be nulls for the quote. In the table -// below the symbol <quote> means just the one byte containing the ''' or '"'. The nulls for the -// quote character are explicitly shown. -// -// <quote> <quote> - 1: No BOM, this must be an 8 bit case. -// <quote> \xEF \xBB \xBF <quote> - 1.12-13: The 8 bit form. -// -// <quote> \xFE \xFF \x00 <quote> - 1.22-23: The 16 bit, big endian form -// <quote> \x00 \xFF \xFE <quote> - 1.32-33: The 16 bit, little endian form. -// -// <quote> \x00 \x00 \xFE \xFF \x00 \x00 \x00 <quote> - 1.32.43-45.56-57: The 32 bit, big endian form. -// <quote> \x00 \x00 \x00 \xFF \xFE \x00 \x00 <quote> - 1.32.43.54-57: The 32 bit, little endian form. - -enum { - eBOM_8_1 = 0xEF, - eBOM_8_2 = 0xBB, - eBOM_8_3 = 0xBF, - eBOM_Big_1 = 0xFE, - eBOM_Big_2 = 0xFF, - eBOM_Little_1 = eBOM_Big_2, - eBOM_Little_2 = eBOM_Big_1 -}; - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::RecognizeBOM ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - while ( true ) { // Handle one character at a time, the micro-state (fPosition) changes for each. - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const unsigned char currChar = *ths->fBufferPtr; // ! The BOM bytes look like integers bigger than 127. - - switch ( ths->fPosition ) { - - case 0 : // Look for the opening quote. - if ( (currChar != '\'') && (currChar != '"') ) return eTriNo; - ths->fQuoteChar = currChar; - ths->fBufferPtr++; - ths->fPosition = 1; - break; // ! Don't fall through, have to check for the end of the buffer between each byte. - - case 1 : // Look at the byte immediately following the opening quote. - if ( currChar == ths->fQuoteChar ) { // Closing quote, no BOM character, must be 8 bit. - if ( ths->fCharForm != eChar8Bit ) return eTriNo; - ths->fBufferPtr += bytesPerChar; // Skip the nulls after the closing quote. - return eTriYes; - } else if ( currChar == eBOM_8_1 ) { // Start of the 8 bit form. - if ( ths->fCharForm != eChar8Bit ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 12; - } else if ( currChar == eBOM_Big_1 ) { // Start of the 16 bit big endian form. - if ( ths->fCharForm != eChar16BitBig ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 22; - } else if ( currChar == 0 ) { // Start of the 16 bit little endian or either 32 bit form. - if ( ths->fCharForm == eChar8Bit ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 32; - } else { - return eTriNo; - } - break; - - case 12 : // Look for the second byte of the 8 bit form. - if ( currChar != eBOM_8_2 ) return eTriNo; - ths->fPosition = 13; - ths->fBufferPtr++; - break; - - case 13 : // Look for the third byte of the 8 bit form. - if ( currChar != eBOM_8_3 ) return eTriNo; - ths->fPosition = 99; - ths->fBufferPtr++; - break; - - case 22 : // Look for the second byte of the 16 bit big endian form. - if ( currChar != eBOM_Big_2 ) return eTriNo; - ths->fPosition = 23; - ths->fBufferPtr++; - break; - - case 23 : // Look for the null before the closing quote of the 16 bit big endian form. - if ( currChar != 0 ) return eTriNo; - ths->fBufferPtr++; - ths->fPosition = 99; - break; - - case 32 : // Look at the second byte of the 16 bit little endian or either 32 bit form. - if ( currChar == eBOM_Little_1 ) { - ths->fPosition = 33; - } else if ( currChar == 0 ) { - ths->fPosition = 43; - } else { - return eTriNo; - } - ths->fBufferPtr++; - break; - - case 33 : // Look for the third byte of the 16 bit little endian form. - if ( ths->fCharForm != eChar16BitBig ) return eTriNo; // Null count before assumed big endian. - if ( currChar != eBOM_Little_2 ) return eTriNo; - ths->fCharForm = eChar16BitLittle; - ths->fPosition = 99; - ths->fBufferPtr++; - break; - - case 43 : // Look at the third byte of either 32 bit form. - if ( ths->fCharForm != eChar32BitBig ) return eTriNo; // Null count before assumed big endian. - if ( currChar == eBOM_Big_1 ) { - ths->fPosition = 44; - } else if ( currChar == 0 ) { - ths->fPosition = 54; - } else { - return eTriNo; - } - ths->fBufferPtr++; - break; - - case 44 : // Look for the fourth byte of the 32 bit big endian form. - if ( currChar != eBOM_Big_2 ) return eTriNo; - ths->fPosition = 45; - ths->fBufferPtr++; - break; - - case 45 : // Look for the first null before the closing quote of the 32 bit big endian form. - if ( currChar != 0 ) return eTriNo; - ths->fPosition = 56; - ths->fBufferPtr++; - break; - - case 54 : // Look for the fourth byte of the 32 bit little endian form. - ths->fCharForm = eChar32BitLittle; - if ( currChar != eBOM_Little_1 ) return eTriNo; - ths->fPosition = 55; - ths->fBufferPtr++; - break; - - case 55 : // Look for the fifth byte of the 32 bit little endian form. - if ( currChar != eBOM_Little_2 ) return eTriNo; - ths->fPosition = 56; - ths->fBufferPtr++; - break; - - case 56 : // Look for the next to last null before the closing quote of the 32 bit forms. - if ( currChar != 0 ) return eTriNo; - ths->fPosition = 57; - ths->fBufferPtr++; - break; - - case 57 : // Look for the last null before the closing quote of the 32 bit forms. - if ( currChar != 0 ) return eTriNo; - ths->fPosition = 99; - ths->fBufferPtr++; - break; - - default : // Look for the closing quote. - assert ( ths->fPosition == 99 ); - if ( currChar != ths->fQuoteChar ) return eTriNo; - ths->fBufferPtr += bytesPerChar; // Skip the nulls after the closing quote. - return eTriYes; - break; - - } - - } - -} // RecognizeBOM - - -// ================================================================================================= -// RecordHeadAttr -// ============== - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::RecordHeadAttr ( PacketMachine * ths, const char * /* unused */ ) -{ - - if ( ths->fAttrName == "encoding" ) { - - assert ( ths->fEncodingAttr.empty() ); - ths->fEncodingAttr = ths->fAttrValue; - - } else if ( ths->fAttrName == "bytes" ) { - - long value = 0; - size_t count = ths->fAttrValue.size(); - size_t i; - - assert ( ths->fBytesAttr == -1 ); - - if ( count > 0 ) { // Allow bytes='' to be the same as no bytes attribute. - - for ( i = 0; i < count; i++ ) { - const char currChar = ths->fAttrValue[i]; - if ( ('0' <= currChar) && (currChar <= '9') ) { - value = (value * 10) + (currChar - '0'); - } else { - ths->fBogusPacket = true; - value = -1; - break; - } - } - ths->fBytesAttr = value; - - if ( CharFormIs16Bit ( ths->fCharForm ) ) { - if ( (ths->fBytesAttr & 1) != 0 ) ths->fBogusPacket = true; - } else if ( CharFormIs32Bit ( ths->fCharForm ) ) { - if ( (ths->fBytesAttr & 3) != 0 ) ths->fBogusPacket = true; - } - - } - - } - - ths->fAttrName.erase ( ths->fAttrName.begin(), ths->fAttrName.end() ); - ths->fAttrValue.erase ( ths->fAttrValue.begin(), ths->fAttrValue.end() ); - - return eTriYes; - -} // RecordHeadAttr - - -// ================================================================================================= -// CaptureAccess -// ============= - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::CaptureAccess ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - while ( true ) { - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - - switch ( ths->fPosition ) { - - case 0 : // Look for the opening quote. - if ( (currChar != '\'') && (currChar != '"') ) return eTriNo; - ths->fQuoteChar = currChar; - ths->fBufferPtr += bytesPerChar; - ths->fPosition = 1; - break; // ! Don't fall through, have to check for the end of the buffer between each byte. - - case 1 : // Look for the 'r' or 'w'. - if ( (currChar != 'r') && (currChar != 'w') ) return eTriNo; - ths->fAccess = currChar; - ths->fBufferPtr += bytesPerChar; - ths->fPosition = 2; - break; - - default : // Look for the closing quote. - assert ( ths->fPosition == 2 ); - if ( currChar != ths->fQuoteChar ) return eTriNo; - ths->fBufferPtr += bytesPerChar; - return eTriYes; - break; - - } - - } - -} // CaptureAccess - - -// ================================================================================================= -// RecordTailAttr -// ============== - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::RecordTailAttr ( PacketMachine * ths, const char * /* unused */ ) -{ - - // There are no known "general" attributes for the packet trailer. - - ths->fAttrName.erase ( ths->fAttrName.begin(), ths->fAttrName.end() ); - ths->fAttrValue.erase ( ths->fAttrValue.begin(), ths->fAttrValue.end() ); - - return eTriYes; - - -} // RecordTailAttr - - -// ================================================================================================= -// CheckPacketEnd -// ============== -// -// Check for trailing padding and record the packet length. We have trailing padding if the bytes -// attribute is present and has a value greater than the current length. - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::CheckPacketEnd ( PacketMachine * ths, const char * /* unused */ ) -{ - const int bytesPerChar = ths->fBytesPerChar; - - if ( ths->fPosition == 0 ) { // First call, decide if there is trailing padding. - - const SInt64 currLength = (ths->fBufferOffset + (ths->fBufferPtr - ths->fBufferOrigin)) - ths->fPacketStart; - - if ( (ths->fBytesAttr != -1) && (ths->fBytesAttr != currLength) ) { - if ( ths->fBytesAttr < currLength ) { - ths->fBogusPacket = true; // The bytes attribute value is too small. - } else { - ths->fPosition = (signed long)(ths->fBytesAttr - currLength); - if ( (ths->fPosition % ths->fBytesPerChar) != 0 ) { - ths->fBogusPacket = true; // The padding is not a multiple of the character size. - ths->fPosition = (ths->fPosition / ths->fBytesPerChar) * ths->fBytesPerChar; - } - } - } - - } - - while ( ths->fPosition > 0 ) { - - if ( ths->fBufferPtr >= ths->fBufferLimit ) return eTriMaybe; - - const char currChar = *ths->fBufferPtr; - - if ( (currChar != ' ') && (currChar != '\t') && (currChar != '\n') && (currChar != '\r') ) { - ths->fBogusPacket = true; // The padding is not whitespace. - break; // Stop the packet here. - } - - ths->fPosition -= bytesPerChar; - ths->fBufferPtr += bytesPerChar; - - } - - ths->fPacketLength = (ths->fBufferOffset + (ths->fBufferPtr - ths->fBufferOrigin)) - ths->fPacketStart; - return eTriYes; - -} // CheckPacketEnd - - -// ================================================================================================= -// CheckFinalNulls -// =============== -// -// Do some special case processing for little endian characters. We have to make sure the presumed -// nulls after the last character actually exist, i.e. that the stream does not end too soon. Note -// that the prior character scanning has moved the buffer pointer to the address following the last -// byte of the last character. I.e. we're already past the presumed nulls, so we can't check their -// content. All we can do is verify that the stream does not end too soon. -// -// Doing this check is simple yet subtle. If we're still in the current buffer then the trailing -// bytes obviously exist. If we're exactly at the end of the buffer then the bytes also exist. -// The only question is when we're actually past this buffer, partly into the next buffer. This is -// when "ths->fBufferPtr > ths->fBufferLimit" on entry. For that case we have to wait until we've -// actually seen enough extra bytes of input. -// -// Since the normal buffer processing is already adjusting for this partial character overrun, all -// that needs to be done here is wait until "ths->fBufferPtr <= ths->fBufferLimit" on entry. In -// other words, if we're presently too far, ths->fBufferPtr will be adjusted by the amount of the -// overflow the next time QEScanner::Scan is called. This might still be too far, so just keep -// waiting for enough data to pass by. -// -// Note that there is a corresponding special case for big endian characters, we must decrement the -// starting offset by the number of leading nulls. But we don't do that here, we leave it to the -// outer code. This is because the leading nulls might have been at the exact end of a previous -// buffer, in which case we have to also decrement the length of that raw data snip. - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::CheckFinalNulls ( PacketMachine * ths, const char * /* unused */ ) -{ - - if ( (ths->fCharForm != eChar8Bit) && CharFormIsLittleEndian ( ths->fCharForm ) ) { - if ( ths->fBufferPtr > ths->fBufferLimit ) return eTriMaybe; - } - - return eTriYes; - -} // CheckFinalNulls - - -// ================================================================================================= -// SetNextRecognizer -// ================= - -void -QEScanner::PacketMachine::SetNextRecognizer ( RecognizerKind nextRecognizer ) -{ - - fRecognizer = nextRecognizer; - fPosition = 0; - -} // SetNextRecognizer - - -// ================================================================================================= -// FindNextPacket -// ============== - -// *** When we start validating intervening nulls for 2 and 4 bytes characters, throw an exception -// *** for errors. Don't return eTriNo, that might skip at an optional point. - -QEScanner::PacketMachine::TriState -QEScanner::PacketMachine::FindNextPacket () -{ - - TriState status; - - #define kPacketHead "?xpacket begin=" - #define kPacketID "W5M0MpCehiHzreSzNTczkc9d" - #define kPacketTail "?xpacket end=" - - static const RecognizerInfo recognizerTable [eRecognizerCount] = { // ! Would be safer to assign these explicitly. - - // proc successNext failureNext literal - - { NULL, eFailureRecognizer, eFailureRecognizer, NULL}, // eFailureRecognizer - { NULL, eSuccessRecognizer, eSuccessRecognizer, NULL}, // eSuccessRecognizer - - { FindLessThan, eHeadStartRecorder, eFailureRecognizer, "H" }, // eLeadInRecognizer - { RecordStart, eHeadStartRecognizer, eLeadInRecognizer, NULL }, // eHeadStartRecorder - { MatchString, eBOMRecognizer, eLeadInRecognizer, kPacketHead }, // eHeadStartRecognizer - - { RecognizeBOM, eIDTagRecognizer, eLeadInRecognizer, NULL }, // eBOMRecognizer - - { MatchString, eIDOpenRecognizer, eLeadInRecognizer, " id=" }, // eIDTagRecognizer - { MatchOpenQuote, eIDValueRecognizer, eLeadInRecognizer, NULL }, // eIDOpenRecognizer - { MatchString, eIDCloseRecognizer, eLeadInRecognizer, kPacketID }, // eIDValueRecognizer - { MatchCloseQuote, eAttrSpaceRecognizer_1, eLeadInRecognizer, NULL }, // eIDCloseRecognizer - - { MatchChar, eAttrNameRecognizer_1, eHeadEndRecognizer, " " }, // eAttrSpaceRecognizer_1 - { CaptureAttrName, eAttrValueRecognizer_1, eLeadInRecognizer, NULL }, // eAttrNameRecognizer_1 - { CaptureAttrValue, eAttrValueRecorder_1, eLeadInRecognizer, NULL }, // eAttrValueRecognizer_1 - { RecordHeadAttr, eAttrSpaceRecognizer_1, eLeadInRecognizer, NULL }, // eAttrValueRecorder_1 - - { MatchString, eBodyRecognizer, eLeadInRecognizer, "?>" }, // eHeadEndRecognizer - - { FindLessThan, eTailStartRecognizer, eBodyRecognizer, "T"}, // eBodyRecognizer - - { MatchString, eAccessValueRecognizer, eBodyRecognizer, kPacketTail }, // eTailStartRecognizer - { CaptureAccess, eAttrSpaceRecognizer_2, eBodyRecognizer, NULL }, // eAccessValueRecognizer - - { MatchChar, eAttrNameRecognizer_2, eTailEndRecognizer, " " }, // eAttrSpaceRecognizer_2 - { CaptureAttrName, eAttrValueRecognizer_2, eBodyRecognizer, NULL }, // eAttrNameRecognizer_2 - { CaptureAttrValue, eAttrValueRecorder_2, eBodyRecognizer, NULL }, // eAttrValueRecognizer_2 - { RecordTailAttr, eAttrSpaceRecognizer_2, eBodyRecognizer, NULL }, // eAttrValueRecorder_2 - - { MatchString, ePacketEndRecognizer, eBodyRecognizer, "?>" }, // eTailEndRecognizer - { CheckPacketEnd, eCloseOutRecognizer, eBodyRecognizer, "" }, // ePacketEndRecognizer - { CheckFinalNulls, eSuccessRecognizer, eBodyRecognizer, "" } // eCloseOutRecognizer - - }; - - while ( true ) { - - switch ( fRecognizer ) { - - case eFailureRecognizer : - return eTriNo; - - case eSuccessRecognizer : - return eTriYes; - - default : - - // ------------------------------------------------------------------- - // For everything else, the normal cases, use the state machine table. - - const RecognizerInfo * thisState = &recognizerTable [fRecognizer]; - - status = thisState->proc ( this, thisState->literal ); - - switch ( status ) { - - case eTriNo : - SetNextRecognizer ( thisState->failureNext ); - continue; - - case eTriYes : - SetNextRecognizer ( thisState->successNext ); - continue; - - case eTriMaybe : - fBufferOverrun = (unsigned char)(fBufferPtr - fBufferLimit); - return eTriMaybe; // Keep this recognizer intact, to be resumed later. - - } - - } // switch ( fRecognizer ) { ... - - } // while ( true ) { ... - -} // FindNextPacket - - -// ================================================================================================= -// ================================================================================================= -// class InternalSnip -// ================== - - -// ================================================================================================= -// InternalSnip -// ============ - -QEScanner::InternalSnip::InternalSnip ( SInt64 offset, SInt64 length ) -{ - - fInfo.fOffset = offset; - fInfo.fLength = length; - -} // InternalSnip - - -// ================================================================================================= -// InternalSnip -// ============ - -QEScanner::InternalSnip::InternalSnip ( const InternalSnip & rhs ) : - fInfo ( rhs.fInfo ), - fMachine ( NULL ) -{ - - assert ( rhs.fMachine.get() == NULL ); // Don't copy a snip with a machine. - assert ( (rhs.fInfo.fEncodingAttr == 0) || (*rhs.fInfo.fEncodingAttr == 0) ); // Don't copy a snip with an encoding. - -} // InternalSnip - - -// ================================================================================================= -// ~InternalSnip -// ============= - -QEScanner::InternalSnip::~InternalSnip () -{ -} // ~InternalSnip - - - -// ================================================================================================= -// ================================================================================================= -// class QEScanner -// ================ - - -// ================================================================================================= -// DumpSnipList -// ============ - -static const char * snipStateName [6] = { "not-seen", "pending", "raw-data", "good-packet", "partial", "bad-packet" }; - -void -QEScanner::DumpSnipList ( const char * title ) -{ -#if 1 - InternalSnipIterator currPos = fInternalSnips.begin(); - InternalSnipIterator endPos = fInternalSnips.end(); - - cout << endl << title << " snip list: " << fInternalSnips.size() << endl; - - for ( ; currPos != endPos; ++currPos ) { - SnipInfo * currSnip = &currPos->fInfo; - cout << '\t' << currSnip << ' ' << snipStateName[currSnip->fState] << ' ' - << currSnip->fOffset << ".." << (currSnip->fOffset + currSnip->fLength - 1) - << ' ' << currSnip->fLength << ' ' << endl; - } -#endif -} // DumpSnipList - - -// ================================================================================================= -// PrevSnip and NextSnip -// ===================== - -QEScanner::InternalSnipIterator -QEScanner::PrevSnip ( InternalSnipIterator snipPos ) -{ - - InternalSnipIterator prev = snipPos; - return --prev; - -} // PrevSnip - -QEScanner::InternalSnipIterator -QEScanner::NextSnip ( InternalSnipIterator snipPos ) -{ - - InternalSnipIterator next = snipPos; - return ++next; - -} // NextSnip - - -// ================================================================================================= -// QEScanner -// ========== -// -// Initialize the scanner object with one "not seen" snip covering the whole stream. - -QEScanner::QEScanner ( SInt64 streamLength ) : - - fStreamLength ( streamLength ) - -{ - InternalSnip rootSnip ( 0, streamLength ); - - if ( streamLength > 0 ) fInternalSnips.push_front ( rootSnip ); // Be nice for empty files. - // DumpSnipList ( "New QEScanner" ); - -} // QEScanner - - -// ================================================================================================= -// ~QEScanner -// =========== - -QEScanner::~QEScanner() -{ - -} // ~QEScanner - - -// ================================================================================================= -// GetSnipCount -// ============ - -size_t -QEScanner::GetSnipCount () -{ - - return fInternalSnips.size(); - -} // GetSnipCount - - -// ================================================================================================= -// StreamAllScanned -// ================ - -bool -QEScanner::StreamAllScanned () -{ - InternalSnipIterator currPos = fInternalSnips.begin(); - InternalSnipIterator endPos = fInternalSnips.end(); - - for ( ; currPos != endPos; ++currPos ) { - if ( currPos->fInfo.fState == eNotSeenSnip ) return false; - } - return true; - -} // StreamAllScanned - - -// ================================================================================================= -// SplitInternalSnip -// ================= -// -// Split the given snip into up to 3 pieces. The new pieces are inserted before and after this one -// in the snip list. The relOffset is the first byte to be kept, it is relative to this snip. If -// the preceeding or following snips have the same state as this one, just shift the boundaries. -// I.e. move the contents from one snip to the other, don't create a new snip. - -// *** To be thread safe we ought to lock the entire list during manipulation. Let data scanning -// *** happen in parallel, serialize all mucking with the list. - -void -QEScanner::SplitInternalSnip ( InternalSnipIterator snipPos, SInt64 relOffset, SInt64 newLength ) -{ - - assert ( (relOffset + newLength) > relOffset ); // Check for overflow. - assert ( (relOffset + newLength) <= snipPos->fInfo.fLength ); - - // ----------------------------------- - // First deal with the low offset end. - - if ( relOffset > 0 ) { - - InternalSnipIterator prevPos; - if ( snipPos != fInternalSnips.begin() ) prevPos = PrevSnip ( snipPos ); - - if ( (snipPos != fInternalSnips.begin()) && (snipPos->fInfo.fState == prevPos->fInfo.fState) ) { - prevPos->fInfo.fLength += relOffset; // Adjust the preceeding snip. - } else { - InternalSnip headExcess ( snipPos->fInfo.fOffset, relOffset ); - headExcess.fInfo.fState = snipPos->fInfo.fState; - headExcess.fInfo.fOutOfOrder = snipPos->fInfo.fOutOfOrder; - fInternalSnips.insert ( snipPos, headExcess ); // Insert the head piece before the middle piece. - } - - snipPos->fInfo.fOffset += relOffset; // Adjust the remainder of this snip. - snipPos->fInfo.fLength -= relOffset; - - } - - // ---------------------------------- - // Now deal with the high offset end. - - if ( newLength < snipPos->fInfo.fLength ) { - - InternalSnipIterator nextPos = NextSnip ( snipPos ); - const SInt64 tailLength = snipPos->fInfo.fLength - newLength; - - if ( (nextPos != fInternalSnips.end()) && (snipPos->fInfo.fState == nextPos->fInfo.fState) ) { - nextPos->fInfo.fOffset -= tailLength; // Adjust the following snip. - nextPos->fInfo.fLength += tailLength; - } else { - InternalSnip tailExcess ( (snipPos->fInfo.fOffset + newLength), tailLength ); - tailExcess.fInfo.fState = snipPos->fInfo.fState; - tailExcess.fInfo.fOutOfOrder = snipPos->fInfo.fOutOfOrder; - fInternalSnips.insert ( nextPos, tailExcess ); // Insert the tail piece after the middle piece. - } - - snipPos->fInfo.fLength = newLength; - - } - -} // SplitInternalSnip - - -// ================================================================================================= -// MergeInternalSnips -// ================== - -QEScanner::InternalSnipIterator -QEScanner::MergeInternalSnips ( InternalSnipIterator firstPos, InternalSnipIterator secondPos ) -{ - - firstPos->fInfo.fLength += secondPos->fInfo.fLength; - fInternalSnips.erase ( secondPos ); - return firstPos; - -} // MergeInternalSnips - - -// ================================================================================================= -// Scan -// ==== - -void -QEScanner::Scan ( const void * bufferOrigin, SInt64 bufferOffset, SInt64 bufferLength ) -{ - SInt64 relOffset; - - #if 0 - cout << "Scan: @ " << bufferOrigin << ", " << bufferOffset << ", " << bufferLength << endl; - #endif - - if ( bufferLength == 0 ) return; - - // ---------------------------------------------------------------- - // These comparisons are carefully done to avoid overflow problems. - - if ( (bufferOffset >= fStreamLength) || - (bufferLength > (fStreamLength - bufferOffset)) || - (bufferOrigin == 0) ) { - throw ScanError ( "Bad origin, offset, or length" ); - } - - // ---------------------------------------------------------------------------------------------- - // This buffer must be within a not-seen snip. Find it and split it. The first snip whose whose - // end is beyond the buffer must be the enclosing one. - - // *** It would be friendly for rescans for out of order problems to accept any buffer postion. - - const SInt64 endOffset = bufferOffset + bufferLength - 1; - InternalSnipIterator snipPos = fInternalSnips.begin(); - - while ( endOffset > (snipPos->fInfo.fOffset + snipPos->fInfo.fLength - 1) ) ++ snipPos; - if ( snipPos->fInfo.fState != eNotSeenSnip ) throw ScanError ( "Already seen" ); - - relOffset = bufferOffset - snipPos->fInfo.fOffset; - if ( (relOffset + bufferLength) > snipPos->fInfo.fLength ) throw ScanError ( "Not within existing snip" ); - - SplitInternalSnip ( snipPos, relOffset, bufferLength ); // *** If sequential & prev is partial, just tack on, - - // -------------------------------------------------------- - // Merge this snip with the preceeding snip if appropriate. - - // *** When out of order I/O is supported we have to do something about buffers who's predecessor is not seen. - - if ( snipPos->fInfo.fOffset > 0 ) { - InternalSnipIterator prevPos = PrevSnip ( snipPos ); - if ( prevPos->fInfo.fState == ePartialPacketSnip ) snipPos = MergeInternalSnips ( prevPos, snipPos ); - } - - // ---------------------------------- - // Look for packets within this snip. - - snipPos->fInfo.fState = ePendingSnip; - PacketMachine* thisMachine = snipPos->fMachine.get(); - // DumpSnipList ( "Before scan" ); - - if ( thisMachine != 0 ) { - thisMachine->AssociateBuffer ( bufferOffset, bufferOrigin, bufferLength ); - } else { - // *** snipPos->fMachine.reset ( new PacketMachine ( bufferOffset, bufferOrigin, bufferLength ) ); VC++ lacks reset - #if 0 - snipPos->fMachine = auto_ptr<PacketMachine> ( new PacketMachine ( bufferOffset, bufferOrigin, bufferLength ) ); - #else - { - // Some versions of gcc complain about the assignment operator above. This avoids the gcc bug. - PacketMachine * pm = new PacketMachine ( bufferOffset, bufferOrigin, bufferLength ); - auto_ptr<PacketMachine> ap ( pm ); - snipPos->fMachine = ap; - } - #endif - thisMachine = snipPos->fMachine.get(); - } - - bool bufferDone = false; - while ( ! bufferDone ) { - - PacketMachine::TriState foundPacket = thisMachine->FindNextPacket(); - - if ( foundPacket == PacketMachine::eTriNo ) { - - // ----------------------------------------------------------------------- - // No packet, mark the snip as raw data and get rid of the packet machine. - // We're done with this buffer. - - snipPos->fInfo.fState = eRawInputSnip; - #if 0 - snipPos->fMachine = auto_ptr<PacketMachine>(); // *** snipPos->fMachine.reset(); VC++ lacks reset - #else - { - // Some versions of gcc complain about the assignment operator above. This avoids the gcc bug. - auto_ptr<PacketMachine> ap ( 0 ); - snipPos->fMachine = ap; - } - #endif - bufferDone = true; - - } else { - - // --------------------------------------------------------------------------------------------- - // Either a full or partial packet. First trim any excess off of the front as a raw input snip. - // If this is a partial packet mark the snip and keep the packet machine to be resumed later. - // We're done with this buffer, the partial packet by definition extends to the end. If this is - // a complete packet first extract the additional information from the packet machine. If there - // is leftover data split the snip and transfer the packet machine to the new trailing snip. - - if ( thisMachine->fPacketStart > snipPos->fInfo.fOffset ) { - - // There is data at the front of the current snip that must be trimmed. - SnipState savedState = snipPos->fInfo.fState; - snipPos->fInfo.fState = eRawInputSnip; // ! So it gets propagated to the trimmed front part. - relOffset = thisMachine->fPacketStart - snipPos->fInfo.fOffset; - SplitInternalSnip ( snipPos, relOffset, (snipPos->fInfo.fLength - relOffset) ); - snipPos->fInfo.fState = savedState; - - } - - if ( foundPacket == PacketMachine::eTriMaybe ) { - - // We have only found a partial packet. - snipPos->fInfo.fState = ePartialPacketSnip; - bufferDone = true; - - } else { - - // We have found a complete packet. Extract all the info for it and split any trailing data. - - InternalSnipIterator packetSnip = snipPos; - SnipState packetState = eValidPacketSnip; - - if ( thisMachine->fBogusPacket ) packetState = eBadPacketSnip; - - packetSnip->fInfo.fAccess = thisMachine->fAccess; - packetSnip->fInfo.fCharForm = thisMachine->fCharForm; - packetSnip->fInfo.fBytesAttr = thisMachine->fBytesAttr; - packetSnip->fInfo.fEncodingAttr = thisMachine->fEncodingAttr.c_str(); - thisMachine->fEncodingAttr.erase ( thisMachine->fEncodingAttr.begin(), thisMachine->fEncodingAttr.end() ); - - if ( (thisMachine->fCharForm != eChar8Bit) && CharFormIsBigEndian ( thisMachine->fCharForm ) ) { - - // ------------------------------------------------------------------------------ - // Handle a special case for big endian characters. The packet machine works as - // though things were little endian. The packet starting offset points to the - // byte containing the opening '<', and the length includes presumed nulls that - // follow the last "real" byte. If the characters are big endian we now have to - // decrement the starting offset of the packet, and also decrement the length of - // the previous snip. - // - // Note that we can't do this before the head trimming above in general. The - // nulls might have been exactly at the end of a buffer and already in the - // previous snip. We are doing this before trimming the tail from the raw snip - // containing the packet. We adjust the raw snip's size because it ends with - // the input buffer. We don't adjust the packet's size, it is already correct. - // - // The raw snip (the one before the packet) might entirely disappear. A simple - // example of this is when the packet is at the start of the file. - - assert ( packetSnip != fInternalSnips.begin() ); // Leading nulls were trimmed! - - if ( packetSnip != fInternalSnips.begin() ) { // ... but let's program defensibly. - - InternalSnipIterator prevSnip = PrevSnip ( packetSnip ); - const unsigned int nullsToAdd = ( CharFormIs16Bit ( thisMachine->fCharForm ) ? 1 : 3 ); - - assert ( nullsToAdd <= prevSnip->fInfo.fLength ); - prevSnip->fInfo.fLength -= nullsToAdd; - if ( prevSnip->fInfo.fLength == 0 ) (void) fInternalSnips.erase ( prevSnip ); - - packetSnip->fInfo.fOffset -= nullsToAdd; - packetSnip->fInfo.fLength += nullsToAdd; - thisMachine->fPacketStart -= nullsToAdd; - - } - - } - - if ( thisMachine->fPacketLength == snipPos->fInfo.fLength ) { - - // This packet ends exactly at the end of the current snip. - #if 0 - snipPos->fMachine = auto_ptr<PacketMachine>(); // *** snipPos->fMachine.reset(); VC++ lacks reset - #else - { - // Some versions of gcc complain about the assignment operator above. This avoids the gcc bug. - auto_ptr<PacketMachine> ap ( 0 ); - snipPos->fMachine = ap; - } - #endif - bufferDone = true; - - } else { - - // There is trailing data to split from the just found packet. - SplitInternalSnip ( snipPos, 0, thisMachine->fPacketLength ); - - InternalSnipIterator tailPos = NextSnip ( snipPos ); - - tailPos->fMachine = snipPos->fMachine; // auto_ptr assignment - taking ownership - thisMachine->ResetMachine (); - - snipPos = tailPos; - - } - - packetSnip->fInfo.fState = packetState; // Do this last to avoid messing up the tail split. - // DumpSnipList ( "Found a packet" ); - - - } - - } - - } - - // -------------------------------------------------------- - // Merge this snip with the preceeding snip if appropriate. - - // *** When out of order I/O is supported we have to check the following snip too. - - if ( (snipPos->fInfo.fOffset > 0) && (snipPos->fInfo.fState == eRawInputSnip) ) { - InternalSnipIterator prevPos = PrevSnip ( snipPos ); - if ( prevPos->fInfo.fState == eRawInputSnip ) snipPos = MergeInternalSnips ( prevPos, snipPos ); - } - - // DumpSnipList ( "After scan" ); - -} // Scan - - -// ================================================================================================= -// Report -// ====== - -void -QEScanner::Report ( SnipInfoVector& snips ) -{ - const size_t count = fInternalSnips.size(); - InternalSnipIterator snipPos = fInternalSnips.begin(); - - size_t s; - - // DumpSnipList ( "Report" ); - - snips.erase ( snips.begin(), snips.end() ); // ! Should use snips.clear, but VC++ doesn't have it. - snips.reserve ( count ); - - for ( s = 0; s < count; s += 1 ) { - snips.push_back ( SnipInfo ( snipPos->fInfo.fState, snipPos->fInfo.fOffset, snipPos->fInfo.fLength ) ); - snips[s] = snipPos->fInfo; // Pick up all of the fields. - ++ snipPos; - } - -} // Report diff --git a/samples/source/common/QEScanner.hpp b/samples/source/common/QEScanner.hpp deleted file mode 100644 index 0c24dce..0000000 --- a/samples/source/common/QEScanner.hpp +++ /dev/null @@ -1,331 +0,0 @@ -#ifndef __QEScanner_hpp__ -#define __QEScanner_hpp__ - -// ================================================================================================= -// Copyright 2002-2005 Adobe Systems Incorporated -// All Rights Reserved. -// -// NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms -// of the Adobe license agreement accompanying it. -// -// Adobe patent application tracking #P435, entitled 'Unique markers to simplify embedding data of -// one format in a file with a different format', inventors: Sean Parent, Greg Gilley. -// ================================================================================================= - -#include <list> -#include <vector> -#include <string> -#include <memory> -#include <stdexcept> - -// ================================================================================================= -// The QEScanner class is used to scan a stream of input for XMP packets. A scanner object is -// constructed then fed the input through a series of calls to Scan. Report may be called at any -// time to get the current knowledge of the input. -// -// A packet starts when a valid header is found and ends when a valid trailer is found. If the -// header contains a "bytes" attribute, additional whitespace must follow. -// -// *** RESTRICTIONS: The current implementation of the scanner has the the following restrictions: -// - The input must be presented in order. -// - Not fully thread safe, don't make concurrent calls to the same QEScanner object. -// ================================================================================================= - -class QEScanner { -public: - - // ============================================================================================= - // The entire input stream is represented as a series of snips. Each snip defines one portion - // of the input stream that either has not been seen, has been seen and contains no packets, is - // exactly one packet, or contains the start of an unfinished packet. Adjacent snips with the - // same state are merged, so the number of snips is always minimal. - // - // A newly constructed QEScanner object has one snip covering the whole input with a state - // of "not seen". A block of input that contains a full XMP packet is split into 3 parts: a - // (possibly empty) raw input snip, the packet, and another (possibly empty) raw input snip. A - // block of input that contains the start of an XMP packet is split into two snips, a (possibly - // empty) raw input snip and the packet start; the following snip must be a "not seen" snip. - // - // It is possible to have ill-formed packets. These have a syntactically valid header and - // trailer, but some semantic error. For example, if the "bytes" attribute length does not span - // to the end of the trailer, or if the following packet begins within trailing padding. - - typedef unsigned char UInt8; - typedef unsigned long UInt32; - typedef long long SInt64; - - enum { - eNotSeenSnip, // This snip has not been seen yet. - ePendingSnip, // This snip is an input buffer being processed. - eRawInputSnip, // This snip is raw input, it doesn't contain any part of an XMP packet. - eValidPacketSnip, // This snip is a complete, valid XMP packet. - ePartialPacketSnip, // This snip contains the start of a possible XMP packet. - eBadPacketSnip // This snip contains a complete, but semantically incorrect XMP packet. - }; - typedef UInt8 SnipState; - - enum { // The values allow easy testing for 16/32 bit and big/little endian. - eChar8Bit = 0, - eChar16BitBig = 2, - eChar16BitLittle = 3, - eChar32BitBig = 4, - eChar32BitLittle = 5 - }; - typedef UInt8 CharacterForm; - - enum { - eChar16BitMask = 2, // These constant shouldn't be used directly, they are mainly - eChar32BitMask = 4, // for the CharFormIsXyz macros below. - eCharLittleEndianMask = 1 - }; - - #define CharFormIs16Bit(f) ( ((int)(f) & QEScanner::eChar16BitMask) != 0 ) - #define CharFormIs32Bit(f) ( ((int)(f) & QEScanner::eChar32BitMask) != 0 ) - - #define CharFormIsBigEndian(f) ( ((int)(f) & QEScanner::eCharLittleEndianMask) == 0 ) - #define CharFormIsLittleEndian(f) ( ((int)(f) & QEScanner::eCharLittleEndianMask) != 0 ) - - struct SnipInfo { - - SInt64 fOffset; // The byte offset of this snip within the input stream. - SInt64 fLength; // The length in bytes of this snip. - SnipState fState; // The state of this snip. - bool fOutOfOrder; // If true, this snip was seen before the one in front of it. - char fAccess; // The read-only/read-write access from the end attribute. - CharacterForm fCharForm; // How the packet is divided into characters. - const char * fEncodingAttr; // The value of the encoding attribute, if any, with nulls removed. - SInt64 fBytesAttr; // The value of the bytes attribute, -1 if not present. - - SnipInfo() : - fOffset ( 0 ), - fLength ( 0 ), - fState ( eNotSeenSnip ), - fOutOfOrder ( false ), - fAccess ( ' ' ), - fCharForm ( eChar8Bit ), - fEncodingAttr ( "" ), - fBytesAttr( -1 ) - { } - - SnipInfo ( SnipState state, SInt64 offset, SInt64 length ) : - fOffset ( offset ), - fLength ( length ), - fState ( state ), - fOutOfOrder ( false ), - fAccess ( ' ' ), - fCharForm ( eChar8Bit ), - fEncodingAttr ( "" ), - fBytesAttr( -1 ) - { } - - }; - - typedef std::vector<SnipInfo> SnipInfoVector; - - QEScanner ( SInt64 streamLength ); - // Constructs a new QEScanner object for a stream with the given length. - - ~QEScanner(); - - size_t GetSnipCount(); - // Returns the number of snips that the stream has been divided into. - - bool StreamAllScanned(); - // Returns true if all of the stream has been seen. - - void Scan ( const void * bufferOrigin, SInt64 bufferOffset, SInt64 bufferLength ); - // Scans the given part of the input, incorporating it in to the known snips. - // The bufferOffset is the offset of this block of input relative to the entire stream. - // The bufferLength is the length in bytes of this block of input. - - void Report ( SnipInfoVector & snips ); - // Produces a report of what is known about the input stream. - - class ScanError : public std::logic_error { - public: - ScanError() throw() : std::logic_error ( "" ) {} - explicit ScanError ( const char * message ) throw() : std::logic_error ( message ) {} - virtual ~ScanError() throw() {} - }; - -private: // QEScanner - - class PacketMachine; - - class InternalSnip { - public: - - SnipInfo fInfo; // The public info about this snip. - std::auto_ptr<PacketMachine> fMachine; // The state machine for "active" snips. - - InternalSnip() {}; // Let everything default. - InternalSnip ( SInt64 offset, SInt64 length ); - InternalSnip ( const InternalSnip & ); - ~InternalSnip(); - - }; // InternalSnip - - typedef std::list<InternalSnip> InternalSnipList; - typedef InternalSnipList::iterator InternalSnipIterator; - - class PacketMachine { - public: - - SInt64 fPacketStart; // Byte offset relative to the entire stream. - SInt64 fPacketLength; // Length in bytes to the end of the trailer processing instruction. - SInt64 fBytesAttr; // The value of the bytes attribute, -1 if not present. - std::string fEncodingAttr; // The value of the encoding attribute, if any, with nulls removed. - CharacterForm fCharForm; // How the packet is divided into characters. - char fAccess; // The read-only/read-write access from the end attribute. - bool fBogusPacket; // True if the packet has an error such as a bad "bytes" attribute value. - - void ResetMachine(); - - enum TriState { - eTriNo, - eTriMaybe, - eTriYes - }; - - TriState FindNextPacket(); - - void AssociateBuffer ( SInt64 bufferOffset, const void * bufferOrigin, SInt64 bufferLength ); - - PacketMachine ( SInt64 bufferOffset, const void * bufferOrigin, SInt64 bufferLength ); - ~PacketMachine(); - - private: // PacketMachine - - PacketMachine() {}; // ! Hide the default constructor. - - enum RecognizerKind { - - eFailureRecognizer, // Not really recognizers, special states to end one buffer's processing. - eSuccessRecognizer, - - eLeadInRecognizer, // Anything up to the next '<'. - eHeadStartRecorder, // Save the starting offset, count intervening nulls. - eHeadStartRecognizer, // The literal string "?xpacket begin=". - - eBOMRecognizer, // Recognize and record the quoted byte order marker. - - eIDTagRecognizer, // The literal string " id=". - eIDOpenRecognizer, // The opening quote for the ID. - eIDValueRecognizer, // The literal string "W5M0MpCehiHzreSzNTczkc9d". - eIDCloseRecognizer, // The closing quote for the ID. - - eAttrSpaceRecognizer_1, // The space before an attribute. - eAttrNameRecognizer_1, // The name of an attribute. - eAttrValueRecognizer_1, // The equal sign and quoted string value for an attribute. - eAttrValueRecorder_1, // Record the value of an attribute. - - eHeadEndRecognizer, // The string literal "?>". - - eBodyRecognizer, // The packet body, anything up to the next '<'. - - eTailStartRecognizer, // The string literal "?xpacket end=". - eAccessValueRecognizer, // Recognize and record the quoted r/w access mode. - - eAttrSpaceRecognizer_2, // The space before an attribute. - eAttrNameRecognizer_2, // The name of an attribute. - eAttrValueRecognizer_2, // The equal sign and quoted string value for an attribute. - eAttrValueRecorder_2, // Record the value of an attribute. - - eTailEndRecognizer, // The string literal "?>". - ePacketEndRecognizer, // Look for trailing padding, check and record the packet size. - eCloseOutRecognizer, // Look for final nulls for little endian multibyte characters. - - eRecognizerCount - - }; - - SInt64 fBufferOffset; // The offset of the data buffer within the input stream. - const char * fBufferOrigin; // The starting address of the data buffer for this snip. - const char * fBufferPtr; // The current postion in the data buffer. - const char * fBufferLimit; // The address one past the last byte in the data buffer. - - RecognizerKind fRecognizer; // Which recognizer is currently active. - /*signed long*/ size_t fPosition; // The internal position within a string literal, etc. - unsigned char fBytesPerChar; // The number of bytes per logical character, 1, 2, or 4. - unsigned char fBufferOverrun; // Non-zero if suspended while skipping intervening nulls. - char fQuoteChar; // The kind of quote seen at the start of a quoted value. - std::string fAttrName; // The name for an arbitrary attribute (other than "begin" and "id"). - std::string fAttrValue; // The value for an arbitrary attribute (other than "begin" and "id"). - - void SetNextRecognizer ( RecognizerKind nextRecognizer ); - - typedef TriState (* RecognizerProc) ( PacketMachine *, const char * ); - - static TriState - FindLessThan ( PacketMachine * ths, const char * which ); - - static TriState - MatchString ( PacketMachine * ths, const char * literal ); - - static TriState - MatchChar ( PacketMachine * ths, const char * literal ); - - static TriState - MatchOpenQuote ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - MatchCloseQuote ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CaptureAttrName ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CaptureAttrValue ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecordStart ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecognizeBOM ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecordHeadAttr ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CaptureAccess ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - RecordTailAttr ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CheckPacketEnd ( PacketMachine * ths, const char * /* unused */ ); - - static TriState - CheckFinalNulls ( PacketMachine * ths, const char * /* unused */ ); - - struct RecognizerInfo { - RecognizerProc proc; - RecognizerKind successNext; - RecognizerKind failureNext; - const char * literal; - }; - - }; // PacketMachine - - SInt64 fStreamLength; - InternalSnipList fInternalSnips; - - void - SplitInternalSnip ( InternalSnipIterator snipPos, SInt64 relOffset, SInt64 newLength ); - - InternalSnipIterator - MergeInternalSnips ( InternalSnipIterator firstPos, InternalSnipIterator secondPos ); - - void - DumpSnipList ( const char * title ); - - InternalSnipIterator - PrevSnip ( InternalSnipIterator snipPos ); - - InternalSnipIterator - NextSnip ( InternalSnipIterator snipPos ); - -}; // QEScanner - - -#endif // __QEScanner_hpp__ diff --git a/samples/source/common/TagTree.cpp b/samples/source/common/TagTree.cpp index 2870057..9a5ac23 100644 --- a/samples/source/common/TagTree.cpp +++ b/samples/source/common/TagTree.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -23,6 +23,8 @@ #include "TagTree.h" #include <stdarg.h> +// silent by default +bool TagTree::verbose = false; #if WIN_ENV //should preferably be within a #if XMP_WinBuild, not doable here... @@ -58,24 +60,40 @@ TagTree::TagTree() rootNode.key="rootkey"; //all never seen rootNode.value="rootvalue"; rootNode.comment="rootcomment"; - tagMap.clear(); // making extra sure it's clear, avoid OS X std:: issues - nodeStack.clear(); - nodeStack.push_back(&rootNode); - lastNode=NULL; + reset(); } TagTree::~TagTree() { + +} + +void TagTree::reset() +{ tagMap.clear(); nodeStack.clear(); + rootNode.children.clear(); nodeStack.push_back(&rootNode); - //any further cleanup needed? + lastNode = NULL; } -void TagTree::setKeyValue(const std::string key,const std::string value, const std::string comment) + +// verbosity control: +void TagTree::setMute() +{ + TagTree::verbose = false; +} + +void TagTree::setVerbose() +{ + TagTree::verbose = true; +} + + +void TagTree::setKeyValue(const std::string key,const std::string value, const std::string _comment) { Node* pCurNode=*nodeStack.rbegin(); //current Node - pCurNode->children.push_back(Node(key,value,comment)); + pCurNode->children.push_back(Node(key,value, _comment)); if(key.size()==0) { // standalone comment? if (value.size()!=0) // must have no value @@ -95,6 +113,9 @@ void TagTree::setKeyValue(const std::string key,const std::string value, const s //add to Map ----------------------------------- lastNode=&*(pCurNode->children.rbegin()); tagMap[extkey]=lastNode; + + if ( verbose ) + Log::info( " setKeyValue( %s |-> %s) [%s]", key.c_str(), value.c_str(), _comment.c_str() ); } void TagTree::digest(LFA_FileRef file,const std::string key /*=NULL*/, @@ -115,7 +136,8 @@ void TagTree::digest(LFA_FileRef file,const std::string key /*=NULL*/, else value=new char[numOfBytes+1]; - if (numOfBytes != LFA_Read ( file, value, numOfBytes, true)) // saying 1,4 guarantes read as ordered (4,1 would not) + // require all == false => leave the throwing to this routine + if (numOfBytes != LFA_Read ( file, value, numOfBytes, false)) // saying 1,4 guarantes read as ordered (4,1 would not) Log::error("could not read %d number of files (End of File reached?)",numOfBytes); char* out=new char[2 + numOfBytes*3 + 5]; //'0x12 34 45 78 ' length formula: 2 ("0x") + numOfBytes x 3 + 5 (padding) @@ -125,7 +147,7 @@ void TagTree::digest(LFA_FileRef file,const std::string key /*=NULL*/, XMP_Int64 i; // *) for (i=0; i < numOfBytes; i++) snprintf(&out[2+i*3],4,"%.2X ",value[i]); //always must allow that extra 0-byte on mac (overwritten again and again) - snprintf(&out[2+i*3],1,"\0"); // *) using i one more time (needed while bug 1613297 regarding snprintf not fixed) + snprintf(&out[2+i*3],1,"%c",'\0'); // *) using i one more time (needed while bug 1613297 regarding snprintf not fixed) setKeyValue(key,out); } @@ -139,14 +161,14 @@ void TagTree::digest(LFA_FileRef file,const std::string key /*=NULL*/, XMP_Int64 TagTree::digest64s(LFA_FileRef file,const std::string key /* ="" */ , bool BigEndian /*=false*/ ) { XMP_Int64 r; - if (8 != LFA_Read ( file, &r, 8, true)) + if (8 != LFA_Read ( file, &r, 8, false)) // require all == false => leave the throwing to this routine Log::error("could not read 8-byte value from file (end of file?)"); if ( ((kBigEndianHost==1) && !BigEndian ) || ((kBigEndianHost==0) && BigEndian )) // "XOR" Flip8(&r); if (!key.empty()) { char out[25]; //longest is "18446744073709551615", 21 chars ==> 25 - snprintf(out,24,"%d",r); //signed, mind the trailing \0 on Mac btw + snprintf(out,24,"%lld",r); //signed, mind the trailing \0 on Mac btw setKeyValue(key,out); } return r; @@ -155,7 +177,7 @@ XMP_Int64 TagTree::digest64s(LFA_FileRef file,const std::string key /* ="" */ , XMP_Uns64 TagTree::digest64u(LFA_FileRef file,const std::string key /* ="" */, bool BigEndian /*=false*/,bool hexDisplay /*=false*/ ) { XMP_Uns64 r; - if (8 != LFA_Read ( file, &r, 8, true)) + if (8 != LFA_Read ( file, &r, 8, false)) // require all == false => leave the throwing to this routine Log::error("could not read 8-byte value from file (end of file?)"); if ( ((kBigEndianHost==1) && !BigEndian ) || ((kBigEndianHost==0) && BigEndian )) // "XOR" Flip8(&r); @@ -168,16 +190,16 @@ XMP_Uns64 TagTree::digest64u(LFA_FileRef file,const std::string key /* ="" */, b snprintf(out , 24 , "%I64u" , r); #else // MAC, UNIX - snprintf(out , 24 , "%0xllu" , r); + snprintf(out , 24 , "%llu" , r); #endif } else { //not working, upper 32 bit empty: #if WIN_ENV - snprintf( out , 24 , "0x%I64.16X" , r ); + snprintf( out , 24 , "0x%.16I64X" , r ); #else - snprintf( out , 24 , "0x%ll.16X" , r ); + snprintf( out , 24 , "0x%.16llX" , r ); #endif } setKeyValue(key,out); @@ -189,7 +211,7 @@ XMP_Uns64 TagTree::digest64u(LFA_FileRef file,const std::string key /* ="" */, b XMP_Int32 TagTree::digest32s(LFA_FileRef file,const std::string key /* ="" */ , bool BigEndian /*=false*/ ) { XMP_Int32 r; - if (4 != LFA_Read ( file, &r, 4, true)) + if (4 != LFA_Read ( file, &r, 4, false)) // require all == false => leave the throwing to this routine Log::error("could not read 4-byte value from file (end of file?)"); if ( ((kBigEndianHost==1) && !BigEndian ) || ((kBigEndianHost==0) && BigEndian )) // "XOR" Flip4(&r); @@ -204,7 +226,7 @@ XMP_Int32 TagTree::digest32s(LFA_FileRef file,const std::string key /* ="" */ , XMP_Uns32 TagTree::digest32u(LFA_FileRef file,const std::string key /* ="" */, bool BigEndian /*=false*/,bool hexDisplay /*=false*/ ) { XMP_Uns32 r; - if (4 != LFA_Read ( file, &r, 4, true)) + if (4 != LFA_Read ( file, &r, 4, false)) // require all == false => leave the throwing to this routine Log::error("could not read 4-byte value from file (end of file?)"); if ( ((kBigEndianHost==1) && !BigEndian ) || ((kBigEndianHost==0) && BigEndian )) // "XOR" Flip4(&r); @@ -222,7 +244,7 @@ XMP_Uns32 TagTree::digest32u(LFA_FileRef file,const std::string key /* ="" */, b XMP_Int16 TagTree::digest16s(LFA_FileRef file,const std::string key /* ="" */ , bool BigEndian /*=false*/ ) { XMP_Int16 r; - if (2 != LFA_Read ( file, &r, 2, true)) + if (2 != LFA_Read ( file, &r, 2, false)) // require all == false => leave the throwing to this routine Log::error("could not read 2-byte value from file (end of file?)"); if ( ((kBigEndianHost==1) && !BigEndian ) || ((kBigEndianHost==0) && BigEndian )) // "XOR" Flip2(&r); @@ -237,7 +259,7 @@ XMP_Int16 TagTree::digest16s(LFA_FileRef file,const std::string key /* ="" */ , XMP_Uns16 TagTree::digest16u(LFA_FileRef file,const std::string key /* ="" */, bool BigEndian /*=false*/,bool hexDisplay /*=false*/ ) { XMP_Uns16 r; - if (2 != LFA_Read ( file, &r, 2, true)) + if (2 != LFA_Read ( file, &r, 2, false)) // require all == false => leave the throwing to this routine Log::error("could not read 2-byte value from file (end of file?)"); if ( ((kBigEndianHost==1) && !BigEndian ) || ((kBigEndianHost==0) && BigEndian )) // "XOR" Flip2(&r); @@ -340,7 +362,7 @@ void TagTree::digest16u(XMP_Uns16* returnValue, LFA_FileRef file,const std::stri ////////////////////////////////////////////////////////////////////////////////////////// -std::string TagTree::digestString(LFA_FileRef file,const std::string key /*=""*/, size_t length /* =0 */, bool verifyZeroTerm /* =false */) +std::string TagTree::digestString(LFA_FileRef file,const std::string key /*=""*/, size_t length /* =0 */, bool verifyZeroTerm /* =false */, bool allowEarlyZeroTerm /* =false */ ) { std::string r(256,'\0'); //give some room in advance (performance) r.clear(); // safety measure (may be needed on mac) @@ -350,6 +372,15 @@ std::string TagTree::digestString(LFA_FileRef file,const std::string key /*=""*/ for ( XMP_Uns32 i = 0; ( i<length ) || (length==0) ; i++ ) { XMP_Uns8 ch = (XMP_Uns8)LFA_GetChar(file); + + // allow early zero termination (useful for fixed length field that may or may not end prematurely) + if ( allowEarlyZeroTerm && ( ch == 0 ) && ( length != 0 ) ) + { + i++; + LFA_Seek( file, length - i, SEEK_CUR ); // compensate for skipped bytes + break; + } + if ( (0x20 <= ch) && (ch <= 0x7E) ) { //outside-case if ( outside ) @@ -357,13 +388,13 @@ std::string TagTree::digestString(LFA_FileRef file,const std::string key /*=""*/ r.push_back(ch); outside = false; } else { + if ( (length==0) && (ch == '\0' ) ) + break; // lenght zero => watch for zero termination... if ( !outside ) r.push_back('<'); //first inside else if (!((length==0) && (ch =='\0'))) r.push_back(' '); //further inside (except very last) outside = true; - if ( (length==0) && (ch == '\0' ) ) - break; // lenght zero => watch for zero termination... char tmp[4]; sprintf(tmp, "%.2X", ch ); r+=tmp; @@ -386,9 +417,9 @@ std::string TagTree::digestString(LFA_FileRef file,const std::string key /*=""*/ return r; } -void TagTree::comment(const std::string comment) +void TagTree::comment(const std::string _comment) { - setKeyValue("","",comment); + setKeyValue("","", _comment); } void TagTree::comment(const char* format, ...) @@ -411,6 +442,9 @@ void TagTree::pushNode(const std::string key) //_and_ push reference to that one on stack Node* pCurNode=*nodeStack.rbegin(); nodeStack.push_back( &*pCurNode->children.rbegin() ); + + if ( verbose ) + Log::info( "pushing %d: %s",nodeStack.size(), key.c_str() ); } //formatstring wrapper @@ -439,6 +473,8 @@ void TagTree::addOffset(LFA_FileRef file) // - addTag()'s will now go to the prior Node void TagTree::popNode() { + // FOR DEBUGGING Log::info( "pop %d",nodeStack.size() ); + if (nodeStack.size() <= 1) Log::error("nodeStack underflow: %d",nodeStack.size()); nodeStack.pop_back(); @@ -473,13 +509,16 @@ void TagTree::changeValue(const char* format, ...) } -void TagTree::addComment(const std::string comment) +void TagTree::addComment(const std::string _comment) { if (!lastNode) Log::error("lastnode NULL"); lastNode->comment=lastNode->comment + ( (lastNode->comment.size())?",":"") //only add comma, if there already is... - +comment; + + _comment; + + if ( verbose ) + Log::info( " addComment: %s", _comment.c_str() ); } void TagTree::addComment(const char* format, ...) @@ -516,13 +555,16 @@ void TagTree::dumpTree(bool commentsFlag /*true*/,Node* pNode /*null*/ ,unsigned std::string indent(depth*2,' '); //read: tab 4 if (commentsFlag) { - Log::info( "%s%s%s%s%s%s%s%s%s", //fancy formatting foo='bar' [re,do] + Log::info( "%s%s%s%s%s%s%s%s%s%s", //fancy formatting foo='bar' [re,do] indent.c_str(), pNode->key.c_str(), pNode->value.size()?" = '":"", pNode->value.c_str(), pNode->value.size()?"'":"", ( pNode->key.size() + pNode->value.size() > 0 ) ? " ":"", + // standalong comments don't need extra indentation: + pNode->comment.size() && ( pNode->key.size() || pNode->value.size() )? + ((std::string("\n ")+indent).c_str()) : "", pNode->comment.size() && ( pNode->key.size() || pNode->value.size() ) ?"[":"", //standalone comments don't need brackets pNode->comment.c_str(), @@ -606,6 +648,47 @@ std::string TagTree::getComment(const std::string key) return tagMap[key]->comment; } +unsigned int TagTree::getSubNodePos( const std::string nodeKey, const std::string parentKey, int skip ) +{ + Node* parent = NULL; + + if( parentKey.empty() ) + { + parent = &rootNode.children.front(); + } + else + { + if ( ! hasNode( parentKey ) ) + Log::error( "parent key %s does not exist", parentKey.c_str() ); + + parent = tagMap[parentKey]; + } + + unsigned int pos = 1; + NodeListIter iter; + for( iter = parent->children.begin(); + iter != parent->children.end() && ( !( (iter->key == (parentKey.empty() ? nodeKey : parentKey + nodeKey)) && skip--<=0 )) ; + iter++, pos++ ); + + if( iter == parent->children.end() ) + pos = 0; + + return pos; +} + + +XMP_Int64 TagTree::getNodeSize( const std::string nodeKey ) +{ + string tmp = tagMap[nodeKey]->comment; + size_t startpos = tmp.find( "size" ) + 5; + if( startpos == string::npos ) + return 0; + tmp = tmp.substr( startpos, tmp.find_first_of( ",", startpos ) ); + + return strtol( tmp.c_str(), NULL, 0 ); +} + + bool TagTree::hasNode(const std::string key) { if ( tagMap.count(key)==0 ) diff --git a/samples/source/common/TagTree.h b/samples/source/common/TagTree.h index 7ec490c..0891711 100644 --- a/samples/source/common/TagTree.h +++ b/samples/source/common/TagTree.h @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated +// Copyright 2008 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -38,6 +38,7 @@ class TagTree { private: struct Node; //forward-looking declaration typedef std::list<Node> NodeList; + typedef std::list<Node>::iterator NodeListIter; struct Node { std::string key; // node structure name resp. tag name @@ -54,10 +55,10 @@ private: } //the one to use - Node(std::string key, std::string value, std::string comment) { - this->key=key; - this->value=value; - this->comment=comment; + Node(std::string _key, std::string _value, std::string _comment) { + this->key = _key; + this->value = _value; + this->comment = _comment; children.clear(); } }; @@ -77,11 +78,20 @@ private: NodeStack nodeStack; Node rootNode; //TODO: ("root",""); - + + // control verbosity to ease debugging: + static bool verbose; + public: TagTree(); ~TagTree(); + void reset(); + + // verbosity control (mute by default ) =================================== + void setMute(); + void setVerbose(); + //input functions ========================================================= void pushNode(const std::string key); @@ -147,11 +157,11 @@ public: //(length is counted w/o trailing zero termination, i.e. length("hans")==4 ) // TODO: length default = 0 not yet implemented // verifyZeroTerm - // - has an effect only if a length!=0 is given + // - has an effect only if a length!=0 is given (otherwise things go up to the 0 anyway) // - in this case, asserts that the string has a terminating zero // (_after_ <length> bytes, otherwise that byte is not read which has an impact on the filepointer!) // would throw if any zero (\0) is encountered prior to that - std::string digestString(LFA_FileRef file,const std::string key="", size_t length=0, bool verifyZeroTerm=false ); + std::string digestString(LFA_FileRef file,const std::string key="", size_t length=0, bool verifyZeroTerm=false, bool allowEarlyZeroTerm=false ); // (wrappers) // standalone comment @@ -178,6 +188,8 @@ public: std::string getValue(const std::string key); std::string getComment(const std::string key); + unsigned int getSubNodePos( const std::string nodeKey, const std::string parentKey = "", int skip = 0 ); + XMP_Int64 getNodeSize( const std::string nodeKey ); //returns true if there is such a node, false if not, error if it happens to be key-value pair bool hasNode(const std::string key); diff --git a/samples/source/common/globals.h b/samples/source/common/globals.h index 291ceb3..af9d033 100644 --- a/samples/source/common/globals.h +++ b/samples/source/common/globals.h @@ -1,10 +1,10 @@ // ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated +// Copyright 2007 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms // of the Adobe license agreement accompanying it. -// +// // ================================================================================================= #ifndef XMPQE_GLOBALS_H @@ -18,12 +18,12 @@ #endif #ifdef WIN_ENV - #define XMPQE_LITTLE_ENDIAN + #define XMPQE_LITTLE_ENDIAN 1 #elif (defined(MAC_ENV) || defined(UNIX_ENV)) #if __BIG_ENDIAN__ - #define XMPQE_BIG_ENDIAN + #define XMPQE_BIG_ENDIAN 1 #elif __LITTLE_ENDIAN__ - #define XMPQE_LITTLE_ENDIAN + #define XMPQE_LITTLE_ENDIAN 1 #else #error "Neither __BIG_ENDIAN__ nor __LITTLE_ENDIAN__ is set" #endif @@ -36,8 +36,52 @@ const char OMNI_CSTRING[]={0x41,0xE4,0xB8,0x80,0x42,0xE4,0xBA,0x8C,0x43,0xC3,0x96,0x44,0xF0,0x90,0x81,0x91,0x45,'\0'}; const char BOM_CSTRING[]={0xEF,0xBB,0xBF,'\0'}; // nb: forgetting the '\0' is a very evil mistake. const std::string OMNI_STRING(OMNI_CSTRING); - //if plain utf8 conversion, mac/win local encoding is a different story... + // if plain utf8 conversion, mac/win local encoding is a different story... const std::string OMNI_BUGINESE("A<E4 B8 80>B<E4 BA 8C>C<C3 96>D<F0 90 81 91>E"); + // if utf16LE + const std::string OMNI_BUGINESE_16LE("A<00 4E>B<8C 4E>C<D6 00>D<00 D8 51 DC>E"); + // if utf16 BE + const std::string OMNI_BUGINESE_16BE("A<4E 00>B<4E 8C>C<00 D6>D<D8 00 DC 51>E"); + + // degraded version of omni-strings + // (sometimes useful for asserts after roundtrips, for setting values see OMNI_STRING in TestCase.h) + // ==> filed bug 2302354 on this issue. + #if WIN_ENV + // a *wrongly* degraded omni-string (non-BMP-char to ??) + const std::string DEG_OMNI_BUGINESE("A?B?C<C3 96>D??E"); + // ditto albeit MacRoman encoding + const std::string MAC_OMNI_BUGINESE("A?B?C<85>D??E"); + #else + // a *correctly* degraded omni-string (non-BMP-char to ?) + const std::string DEG_OMNI_BUGINESE("A?B?C<C3 96>D?E"); + // ditto albeit MacRoman encoding + const std::string MAC_OMNI_BUGINESE("A?B?C<85>D?E"); + #endif + // -> #issue# the non-BMP character in OMNI_STRING between D and E gets converted + // into two question marks (wrong) , not into one (correct) + + const char OEAEUE_UTF8_CSTRING[]={0xC3, 0x84, 0xC3, 0x96, 0xC3, 0x9C,'\0'}; + const std::string AEOEUE_UTF8(OEAEUE_UTF8_CSTRING); + const std::string AEOEUE_UTF8_BUGINESE("<C3 84 C3 96 C3 9C>"); + + const std::string AEOEUE_WIN_LOCAL_BUGINESE("<C4 D6 DC>"); + const std::string AEOEUE_MAC_LOCAL_BUGINESE("<80 85 86>"); + + const std::string AEOEUE_WIN_MOJIBAKE_BUGINESE("<E2 82 AC E2 80 A6 E2 80 A0>"); + const std::string AEOEUE_MAC_MOJIBAKE_BUGINESE("<C6 92 C3 B7 E2 80 B9>"); + const std::string AEOEUE_LATIN1_MOJIBAKE_BUGINESE("<C2 80 C2 85 C2 86>"); + + #if MAC_ENV + const std::string AEOEUE_WIN_LOCAL_TO_UTF8 = AEOEUE_MAC_MOJIBAKE_BUGINESE; + const std::string AEOEUE_MAC_LOCAL_TO_UTF8 = AEOEUE_UTF8_BUGINESE; + #elif WIN_ENV + const std::string AEOEUE_WIN_LOCAL_TO_UTF8 = AEOEUE_UTF8_BUGINESE; + const std::string AEOEUE_MAC_LOCAL_TO_UTF8 = AEOEUE_WIN_MOJIBAKE_BUGINESE; + #else + // windows local encoding will work for UNIX (Latin1), but mac will result MOJIBAKE + const std::string AEOEUE_WIN_LOCAL_TO_UTF8 = AEOEUE_UTF8_BUGINESE; + const std::string AEOEUE_MAC_LOCAL_TO_UTF8 = AEOEUE_LATIN1_MOJIBAKE_BUGINESE; + #endif #endif // XMPQE_GLOBALS_H diff --git a/samples/source/dumpfile/main.cpp b/samples/source/dumpfile/main.cpp index d1541a6..327a41e 100644 --- a/samples/source/dumpfile/main.cpp +++ b/samples/source/dumpfile/main.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2004-2008 Adobe Systems Incorporated +// Copyright 2003 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -18,9 +18,9 @@ //sanity check platform/endianess #include "globals.h" -#ifdef WIN_ENV -#pragma warning ( disable : 4267 ) // suppress string conversion warning -// #pragma warning ( disable : 1234 ) // say what you do here +#if XMP_WinBuild + #pragma warning ( disable : 4267 ) // suppress string conversion warning + // #pragma warning ( disable : 1234 ) // say what you do here #endif //only define in one non-public-source, non-header(.cpp) place @@ -46,21 +46,40 @@ const int DUMPFILEVERSION=2; #include "LargeFileAccess.hpp" -void LFA_Throw ( const char* msg, int id ) +//void LFA_Throw ( const char* msg, int id ) +//{ + //switch(id) + //{ + // case kLFAErr_InternalFailure: + // Log::error("LFA User Abort:%s", msg); + // break; + // case kLFAErr_ExternalFailure: + // Log::error("LFA External Failure:%s", msg); + // break; + // case kLFAErr_UserAbort: + // Log::error("LFA User Abort:%s", msg); + // break; + // default: + // Log::error("LFA unknown error:%s", msg); //should not occur + // break; + //} +//} +#define XMP_Throw(msg,id) { throw XMP_Error ( id, msg ); } +void LFA_Throw ( const char* msg, int id ) { switch(id) { case kLFAErr_InternalFailure: - Log::error("LFA User Abort:%s", msg); + XMP_Throw(msg,kXMPErr_InternalFailure); break; case kLFAErr_ExternalFailure: - Log::error("LFA External Failure:%s", msg); + XMP_Throw(msg,kXMPErr_ExternalFailure); break; case kLFAErr_UserAbort: - Log::error("LFA User Abort:%s", msg); + XMP_Throw(msg,kXMPErr_UserAbort); break; default: - Log::error("LFA unknown error:%s", msg); //should not occur + XMP_Throw(msg,kXMPErr_UnknownException); break; } } @@ -105,10 +124,10 @@ void printUsageLong() { Log::info(""); Log::info("-tree Shows the tree structure of the file. (default)"); Log::info(""); - Log::info("-keys Shows only the key-value nodes found in the file, as a list with no");
+ Log::info("-keys Shows only the key-value nodes found in the file, as a list with no"); Log::info(" hierarchical structure, in alphabetical order by key."); Log::info(""); - Log::info("-list Shows only the key-value nodes found in the file, as a list with no");
+ Log::info("-list Shows only the key-value nodes found in the file, as a list with no"); Log::info(" hierarchical structure, in the order of parsing."); Log::info(""); Log::info("-nocomments Supresses the comment portion of key-value nodes."); diff --git a/samples/source/xmpcommand/Actions.cpp b/samples/source/xmpcommand/Actions.cpp index 2324b10..d0e5e32 100644 --- a/samples/source/xmpcommand/Actions.cpp +++ b/samples/source/xmpcommand/Actions.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2004-2008 Adobe Systems Incorporated +// Copyright 2006 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -23,7 +23,6 @@ const char * XMP_EXE_VERSION= "4.4"; #include "globals.h" #include "Log.h" -#include "OutputUtils.h" #include "Actions.h" //must come after XMP.hpp/XMP_Const.h @@ -90,6 +89,146 @@ const char * XMP_EXE_VERSION= "4.4"; return content; } + /* hand over the fileFormat and get a fileFormatName (2-3 Letters ie. PDF) + * and a Description (i.e. "Portable Document Format") back. + * see also public/include/XMP_Const.h + */ + void fileFormatNameToString ( XMP_FileFormat fileFormat, std::string & fileFormatName, std::string & fileFormatDesc ) + { + fileFormatName.erase(); + fileFormatDesc.erase(); + switch(fileFormat) { + case kXMP_PDFFile: + fileFormatName = "PDF ";fileFormatDesc = "Portable Document Format";break; + case kXMP_PostScriptFile: + fileFormatName = "PS ";fileFormatDesc = "Post Script";break; + case kXMP_EPSFile: + fileFormatName = "EPS ";fileFormatDesc = "Encapsulated Post Script";break; + case kXMP_JPEGFile: + fileFormatName = "JPEG";fileFormatDesc = "Joint Photographic Experts Group";break; + case kXMP_JPEG2KFile: + fileFormatName = "JPX ";fileFormatDesc = "JPEG 2000";break; + case kXMP_TIFFFile: + fileFormatName = "TIFF";fileFormatDesc = "Tagged Image File Format";break; + case kXMP_GIFFile: + fileFormatName = "GIF ";fileFormatDesc = "Graphics Interchange Format";break; + case kXMP_PNGFile: + fileFormatName = "PNG ";fileFormatDesc = "Portable Network Graphic";break; + case kXMP_MOVFile: + fileFormatName = "MOV ";fileFormatDesc = "Quicktime";break; + case kXMP_AVIFile: + fileFormatName = "AVI ";fileFormatDesc = "Quicktime";break; + case kXMP_CINFile: + fileFormatName = "CIN ";fileFormatDesc = "Cineon";break; + case kXMP_WAVFile: + fileFormatName = "WAV ";fileFormatDesc = "WAVE Form Audio Format";break; + case kXMP_MP3File: + fileFormatName = "MP3 ";fileFormatDesc = "MPEG-1 Audio Layer 3";break; + case kXMP_SESFile: + fileFormatName = "SES ";fileFormatDesc = "Audition session";break; + case kXMP_CELFile: + fileFormatName = "CEL ";fileFormatDesc = "Audition loop";break; + case kXMP_MPEGFile: + fileFormatName = "MPEG";fileFormatDesc = "Motion Pictures Experts Group";break; + case kXMP_MPEG2File: + fileFormatName = "MP2 ";fileFormatDesc = "MPEG-2";break; + case kXMP_WMAVFile: + fileFormatName = "WMAV";fileFormatDesc = "Windows Media Audio and Video";break; + case kXMP_HTMLFile: + fileFormatName = "HTML";fileFormatDesc = "HyperText Markup Language";break; + case kXMP_XMLFile: + fileFormatName = "XML ";fileFormatDesc = "Extensible Markup Language";break; + case kXMP_TextFile: + fileFormatName = "TXT ";fileFormatDesc = "text";break; + case kXMP_PhotoshopFile: + fileFormatName = "PSD ";fileFormatDesc = "Photoshop Document";break; + case kXMP_IllustratorFile: + fileFormatName = "AI ";fileFormatDesc = "Adobe Illustrator";break; + case kXMP_InDesignFile: + fileFormatName = "INDD";fileFormatDesc = "Adobe InDesign";break; + case kXMP_AEProjectFile: + fileFormatName = "AEP ";fileFormatDesc = "AfterEffects Project";break; + case kXMP_AEFilterPresetFile: + fileFormatName = "FFX ";fileFormatDesc = "AfterEffects Filter Preset";break; + case kXMP_EncoreProjectFile: + fileFormatName = "NCOR";fileFormatDesc = "Encore Project";break; + case kXMP_PremiereProjectFile: + fileFormatName = "PPRJ";fileFormatDesc = "Premier Project";break; + case kXMP_SWFFile: + fileFormatName = "SWF ";fileFormatDesc = "Shockwave Flash";break; + case kXMP_PremiereTitleFile: + fileFormatName = "PRTL";fileFormatDesc = "Premier Title";break; + case kXMP_UnknownFile: + fileFormatName = " ";fileFormatDesc = "Unkown file format";break; + default: + fileFormatName = " ";fileFormatDesc = "no known format constant";break; + } + } //fileFormatNameToString + + /** + * GetFormatInfo-Flags (aka Handler-Flags) + * find this in XMP_Const.h under "Options for GetFormatInfo" + */ + std::string XMPFiles_FormatInfoToString ( const XMP_OptionBits options) { + std::string outString; + + if( options & kXMPFiles_CanInjectXMP ) + outString.append(" CanInjectXMP"); + if( options & kXMPFiles_CanExpand ) + outString.append(" CanExpand"); + if( options & kXMPFiles_CanRewrite ) + outString.append(" CanRewrite"); + if( options & kXMPFiles_PrefersInPlace ) + outString.append(" PrefersInPlace"); + if( options & kXMPFiles_CanReconcile ) + outString.append(" CanReconcile"); + if( options & kXMPFiles_AllowsOnlyXMP ) + outString.append(" AllowsOnlyXMP"); + if( options & kXMPFiles_ReturnsRawPacket ) + outString.append(" ReturnsRawPacket"); + if( options & kXMPFiles_HandlerOwnsFile ) + outString.append(" HandlerOwnsFile"); + if( options & kXMPFiles_AllowsSafeUpdate ) + outString.append(" AllowsSafeUpdate"); + + if (outString.empty()) outString=" (none)"; + return outString; + } + + /** + * openOptions to String, find this in + * XMP_Const.h under "Options for OpenFile" + */ + std::string XMPFiles_OpenOptionsToString ( const XMP_OptionBits options) { + std::string outString; + if( options & kXMPFiles_OpenForRead) + outString.append(" OpenForRead"); + if( options & kXMPFiles_OpenForUpdate) + outString.append(" OpenForUpdate"); + if( options & kXMPFiles_OpenOnlyXMP) + outString.append(" OpenOnlyXMP"); + if( options & kXMPFiles_OpenStrictly) + outString.append(" OpenStrictly"); + if( options & kXMPFiles_OpenUseSmartHandler) + outString.append(" OpenUseSmartHandler"); + if( options & kXMPFiles_OpenUsePacketScanning) + outString.append(" OpenUsePacketScanning"); + if( options & kXMPFiles_OpenLimitedScanning) + outString.append(" OpenLimitedScanning"); + + if (outString.empty()) outString=" (none)"; + return outString; + } + + //just the callback-Function + XMP_Status DumpToString( void * refCon, XMP_StringPtr outStr, XMP_StringLen outLen ) + { + XMP_Status status = 0; + std::string* dumpString = static_cast < std::string* > ( refCon ); + dumpString->append (outStr, outLen); + return status; + } + // ******************************************************************************** void Actions::version() { @@ -100,21 +239,21 @@ void Actions::version() { Log::info("%s", filesVersion.message); Log::info("Executable Version:%s", XMP_EXE_VERSION); -#ifdef WIN_ENV +#if XMP_WinBuild Log::info("compiled on Windows on %s, %s",__DATE__,__TIME__); #endif -#ifdef MAC_ENV +#if XMP_MacBuild Log::info("compiled on Mac on %s, %s",__DATE__,__TIME__); #endif -#ifndef NDEBUG +#if NDEBUG Log::info("Configuration: debug"); #else Log::info("Configuration: release"); #endif Log::info("Edition: Public SDK"); -#if defined(XMPQE_BIG_ENDIAN) +#if XMPQE_BIG_ENDIAN Log::info("Big endian machine"); -#elif defined(XMPQE_LITTLE_ENDIAN) +#elif XMPQE_LITTLE_ENDIAN Log::info("Little endian machine"); #else Log::warn("unknown Endian !!!"); @@ -144,7 +283,7 @@ void Actions::info() { out_filepath.c_str(),this->mediafile.c_str()); //openOptions (also verify that identical to what's been used) - std::string openFlagsString=Utils::XMPFiles_OpenOptionsToString(out_openFlags); + std::string openFlagsString=XMPFiles_OpenOptionsToString(out_openFlags); Log::info("openFlags: %s (%X)", openFlagsString.c_str(), out_openFlags); if ( generalOpenFlags != out_openFlags ) Log::warn("out_openFlags (0x%X) differ from those used for open (0x%X)", @@ -152,12 +291,12 @@ void Actions::info() { //FileFormat (resolves, usually by way of extension unless specified) std::string fileFormatName, fileFormatDescription; - Utils::fileFormatNameToString(out_fileFormat,fileFormatName,fileFormatDescription); + fileFormatNameToString(out_fileFormat,fileFormatName,fileFormatDescription); Log::info("fileFormat: %s (%s,0x%X)", fileFormatDescription.c_str(),fileFormatName.c_str(),out_fileFormat); //FormatInfo aka "HandlerFlags" (show what is possible with this format) - std::string formatInfoString=Utils::XMPFiles_FormatInfoToString(out_handlerFlags); + std::string formatInfoString=XMPFiles_FormatInfoToString(out_handlerFlags); Log::info("formatInfo:%s (0x%X)", formatInfoString.c_str(), out_handlerFlags); xmpFile.CloseFile(generalCloseFlags); //(enabled again now that bug 1352603 is fixed) @@ -209,7 +348,9 @@ void Actions::dump() { if ( !xmpFile.GetXMP(&xmpMeta,0,0)) Log::warn("file contains no XMP. - says xmpFile.GetXMP()"); else { - Utils::dumpXMP(&xmpMeta); + std::string dump; + xmpMeta.DumpObject(DumpToString, &dump); + Log::info( dump ); } xmpFile.CloseFile(generalCloseFlags); //(enabled again now that bug 1352603 is fixed) @@ -217,42 +358,6 @@ void Actions::dump() { /////////////////////////////////////////////////////////////////////////////////////// -void Actions::thumb() { - if(!this->switch_nocheck) verifyFileIsReadable(this->mediafile); //need readability - XMP_FileFormat xmpFileFormat=kXMP_UnknownFile; - - SXMPFiles xmpFile; - XMP_ThumbnailInfo tnailInfo; - if ( !xmpFile.OpenFile(this->mediafile, - xmpFileFormat, - generalOpenFlags | kXMPFiles_OpenCacheTNail) ) - Log::error("error opening file %s",this->mediafile.c_str()); - - if ( !xmpFile.GetThumbnail(&tnailInfo)) { - Log::warn("no thumbnail found. - says xmpFile.GetThumbnail()"); - //TODO: comment-in again once bug 1352603 fixed - //xmpFile.CloseFile(generalCloseFlags); - return; - } - else { - //FileFormat (resolves, usually by way of extension unless specified) - std::string fileFormatName, fileFormatDescription; - Utils::fileFormatNameToString(tnailInfo.fileFormat,fileFormatName,fileFormatDescription); - Log::info("fileFormat: %s (%s,0x%X)", - fileFormatDescription.c_str(),fileFormatName.c_str(),tnailInfo.fileFormat); - - Log::info("full image size:%dx%d",tnailInfo.fullWidth,tnailInfo.fullHeight); - Log::info("full orientation:0x%X",tnailInfo.fullOrientation); - Log::info("thumbnail size:%dx%d (%d bytes)",tnailInfo.tnailWidth,tnailInfo.tnailHeight,tnailInfo.tnailSize); - Log::info("thumbnail orientation:0x%X",tnailInfo.tnailOrientation); - Log::info("thumbnail format 0x%X",tnailInfo.tnailFormat); - } - - xmpFile.CloseFile(generalCloseFlags); //(enabled again now that bug 1352603 is fixed) -} - - -/////////////////////////////////////////////////////////////////////////////////////// void Actions::put() { //need read and write if(!this->switch_nocheck) verifyFileIsWritable(this->mediafile); @@ -293,7 +398,7 @@ void Actions::put() { } else { //handle the error //FileFormat (resolves, usually by way of extension unless specified TODO) std::string fileFormatName, fileFormatDescription; - Utils::fileFormatNameToString(out_fileFormat,fileFormatName,fileFormatDescription); + fileFormatNameToString(out_fileFormat,fileFormatName,fileFormatDescription); Log::error("the File %s of type %s neither contains XMP nor can it be injected (at least kXMPFiles_CanInjectXMP not returned)", this->mediafile.c_str(), fileFormatName.c_str()); }; @@ -315,9 +420,6 @@ void Actions::doAction() { case Actions::INFO: info(); break; - case Actions::THUMB: //GetThumbnail() - thumb(); - break; case Actions::PUT: //PutXMP() put(); break; diff --git a/samples/source/xmpcommand/Actions.h b/samples/source/xmpcommand/Actions.h index 084c260..f43afbc 100644 --- a/samples/source/xmpcommand/Actions.h +++ b/samples/source/xmpcommand/Actions.h @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2004-2008 Adobe Systems Incorporated +// Copyright 2006 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -11,7 +11,7 @@ class Actions { public: - enum ACTION_TYPE { NONE, VERSION, INFO, PUT, GET, DUMP, THUMB}; + enum ACTION_TYPE { NONE, VERSION, INFO, PUT, GET, DUMP }; //the params to be set: bool switch_safe; //asking for closing file with kXMPFiles_UpdateSafely (safe+rename that is) @@ -36,7 +36,6 @@ class Actions { private: void version(void); void info(void); - void thumb(void); void put(); void get(void); void dump(void); diff --git a/samples/source/xmpcommand/PrintUsage.cpp b/samples/source/xmpcommand/PrintUsage.cpp index 3931e18..7560cbb 100644 --- a/samples/source/xmpcommand/PrintUsage.cpp +++ b/samples/source/xmpcommand/PrintUsage.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2004-2008 Adobe Systems Incorporated +// Copyright 2006 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -60,8 +60,6 @@ namespace XMPQE { Log::info(""); Log::info("Actions:"); Log::info(" info <mediafile> Prints basic information about the file."); - Log::info(" thumb <mediafile> Prints information about the native thumbnail "); - Log::info(" if one is included, or informs that none is available."); Log::info(" put <xmpfile> <mediafile> Injects the XMP contained in the specified xmpfile "); Log::info(" into the specified mediafile."); Log::info(" get <mediafile> Retrieves the XMP Packet contained in the specified mediafile."); diff --git a/samples/source/xmpcommand/PrintUsage.h b/samples/source/xmpcommand/PrintUsage.h index a48b24a..982fe41 100644 --- a/samples/source/xmpcommand/PrintUsage.h +++ b/samples/source/xmpcommand/PrintUsage.h @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2004-2008 Adobe Systems Incorporated +// Copyright 2006 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms diff --git a/samples/source/xmpcommand/main.cpp b/samples/source/xmpcommand/XMPCommand.cpp index 2dc4374..a89904c 100644 --- a/samples/source/xmpcommand/main.cpp +++ b/samples/source/xmpcommand/XMPCommand.cpp @@ -1,5 +1,5 @@ // ================================================================================================= -// Copyright 2005-2008 Adobe Systems Incorporated +// Copyright 2005 Adobe Systems Incorporated // All Rights Reserved. // // NOTICE: Adobe permits you to use, modify, and distribute this file in accordance with the terms @@ -34,16 +34,23 @@ //QE related #include "Actions.h" #include "Log.h" +#include "LargeFileAccess.hpp" // -help output #include "PrintUsage.h" + namespace XMPQE { void InitDependencies() { if ( ! SXMPMeta::Initialize() ) Log::error( "XMPMeta::Initialize failed!" ); - if ( ! SXMPFiles::Initialize() ) + XMP_OptionBits options = 0; + #if UNIX_ENV + options |= kXMPFiles_ServerMode; + #endif + + if ( ! SXMPFiles::Initialize ( options ) ) Log::error( "XMPFiles::Initialize failed!" ); } @@ -140,8 +147,6 @@ int main ( int argc, const char * argv[] ) if ( !strcmp("info",argv[i])) { action.actiontype= Actions::INFO; - } else if ( !strcmp("thumb",argv[i])) { - action.actiontype= Actions::THUMB; } else if ( !strcmp("put",argv[i]) || !strcmp("inject",argv[i]) ) { action.actiontype= Actions::PUT; } else if ( !strcmp("get",argv[i]) || !strcmp("extract",argv[i]) ) { |