From 35d248cab1f0d4800f72abb5cb6afb56f40d9083 Mon Sep 17 00:00:00 2001
From: Michael Stahl <Michael.Stahl@cib.de>
Date: Mon, 28 Oct 2019 14:31:23 +0100
Subject: svl: HTMLParser: stop inserting control character garbage into Writer

E.g. rhbz433940-1.html contains literal ^G characters that are inserted
as-is into SwTextNodes.

This now triggers assert about CH_TXT_ATR_FIELDSTART in
SwSubFont::GetTextSize_() that was added in
19a559b0ec9b806519c405651d6d2b2e14712b4a.

Change-Id: I6aa7de41a04069e15b40865fd57894dae0fc10db
Reviewed-on: https://gerrit.libreoffice.org/81606
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
Tested-by: Michael Stahl <michael.stahl@cib.de>
---
 svtools/source/svhtml/parhtml.cxx | 19 +++++++++++++++----
 1 file changed, 15 insertions(+), 4 deletions(-)

(limited to 'svtools')

diff --git a/svtools/source/svhtml/parhtml.cxx b/svtools/source/svhtml/parhtml.cxx
index a048ab92d923..6ce236566e61 100644
--- a/svtools/source/svhtml/parhtml.cxx
+++ b/svtools/source/svhtml/parhtml.cxx
@@ -30,6 +30,7 @@
 #include <tools/datetime.hxx>
 #include <unotools/datetime.hxx>
 #include <svl/inettype.hxx>
+#include <svl/lngmisc.hxx>
 #include <com/sun/star/beans/PropertyAttribute.hpp>
 #include <com/sun/star/document/XDocumentProperties.hpp>
 
@@ -454,8 +455,12 @@ HtmlTokenId HTMLParser::ScanText( const sal_Unicode cBreak )
                     else
                         nNextCh = 0U;
 
-                    if ( ! rtl::isUnicodeCodePoint( cChar ) )
+                    if (!rtl::isUnicodeCodePoint(cChar)
+                        || (linguistic::IsControlChar(cChar)
+                            && cChar != '\r' && cChar != '\n' && cChar != '\t'))
+                    {
                         cChar = '?';
+                    }
                 }
                 else if( rtl::isAsciiAlpha( nNextCh ) )
                 {
@@ -751,8 +756,11 @@ HtmlTokenId HTMLParser::ScanText( const sal_Unicode cBreak )
             else
             {
                 do {
+                    if (!linguistic::IsControlChar(nNextCh))
+                    {
                     // All remaining characters make their way into the text.
-                    sTmpBuffer.appendUtf32( nNextCh );
+                        sTmpBuffer.appendUtf32( nNextCh );
+                    }
                     if( MAX_LEN == sTmpBuffer.getLength() )
                     {
                         aToken += sTmpBuffer;
@@ -987,8 +995,11 @@ HtmlTokenId HTMLParser::GetNextRawToken()
             }
             [[fallthrough]];
         default:
-            // all remaining characters are appended to the buffer
-            sTmpBuffer.appendUtf32( nNextCh );
+            if (!linguistic::IsControlChar(nNextCh) || nNextCh == '\t')
+            {
+                // all remaining characters are appended to the buffer
+                sTmpBuffer.appendUtf32( nNextCh );
+            }
             break;
         }
 
-- 
cgit v1.2.3