Enable pdfimages to print filenames to stdout.

Add a command line option -print-filenames so that after writing each
imaage to file, pdfimages will print the filename to stdout. This is in
line with the --batch-print option of scanimage and the stream_filelist
parameter of tesseract. It allows easy parallelization of different
stages of an image processing pipeline.

4 files changed, 15 insertions, 0 deletions

diff --git a/utils/ImageOutputDev.cc b/utils/ImageOutputDev.cc
index be12f053..74a9a608 100644
--- a/utils/ImageOutputDev.cc
+++ b/utils/ImageOutputDev.cc
@@ -66,6 +66,7 @@ ImageOutputDev::ImageOutputDev(char *fileRootA, bool pageNamesA, bool listImages
     dumpJBIG2 = false;
     dumpCCITT = false;
     pageNames = pageNamesA;
+    printFilenames = false;
     imgNum = 0;
     pageNum = 0;
     errorCode = 0;
@@ -690,6 +691,10 @@ void ImageOutputDev::writeImage(GfxState *state, Object *ref, Stream *str, int w
     if (inlineImg) {
         embedStr->restore();
     }
+
+    if (printFilenames) {
+        printf("%s\n", fileName);
+    }
 }
 
 bool ImageOutputDev::tilingPatternFill(GfxState *state, Gfx *gfx, Catalog *cat, GfxTilingPattern *tPat, const double *mat, int x0, int y0, int x1, int y1, double xStep, double yStep)
diff --git a/utils/ImageOutputDev.h b/utils/ImageOutputDev.h
index 40bc7713..d23c1638 100644
--- a/utils/ImageOutputDev.h
+++ b/utils/ImageOutputDev.h
@@ -90,6 +90,9 @@ public:
     // Use CCITT format for CCITT files
     void enableCCITT(bool ccitt) { dumpCCITT = ccitt; }
 
+    // Print filenames to stdout after writing
+    void enablePrintFilenames(bool filenames) { printFilenames = filenames; }
+
     // Get the error code
     // 0 = No error, 1 = Error opening a PDF file, 2 = Error opening an output file, 3 = Error related to PDF permissions, 99 = Other error.
     int getErrorCode() const { return errorCode; }
@@ -150,6 +153,7 @@ private:
     bool outputPNG; // set to output in PNG format
     bool outputTiff; // set to output in TIFF format
     bool pageNames; // set to include page number in file names
+    bool printFilenames; // set to print image filenames to stdout after writing
     int pageNum; // current page number
     int imgNum; // current image number
     int errorCode; // code for any error creating the output files
diff --git a/utils/pdfimages.1 b/utils/pdfimages.1
index 8afa9b79..1a5386a0 100644
--- a/utils/pdfimages.1
+++ b/utils/pdfimages.1
@@ -219,6 +219,9 @@ Specify the user password for the PDF file.
 .B \-p
 Include page numbers in output file names.
 .TP
+.B \-print\-filenames
+Print image filenames to stdout.
+.TP
 .B \-q
 Don't print any messages or errors.
 .TP
diff --git a/utils/pdfimages.cc b/utils/pdfimages.cc
index 031a4418..937dafc6 100644
--- a/utils/pdfimages.cc
+++ b/utils/pdfimages.cc
@@ -64,6 +64,7 @@ static bool dumpJBIG2 = false;
 static bool dumpCCITT = false;
 static bool allFormats = false;
 static bool pageNames = false;
+static bool printFilenames = false;
 static char ownerPassword[33] = "\001";
 static char userPassword[33] = "\001";
 static bool quiet = false;
@@ -87,6 +88,7 @@ static const ArgDesc argDesc[] = { { "-f", argInt, &firstPage, 0, "first page to
                                    { "-opw", argString, ownerPassword, sizeof(ownerPassword), "owner password (for encrypted files)" },
                                    { "-upw", argString, userPassword, sizeof(userPassword), "user password (for encrypted files)" },
                                    { "-p", argFlag, &pageNames, 0, "include page numbers in output file names" },
+                                   { "-print-filenames", argFlag, &printFilenames, 0, "print image filenames to stdout" },
                                    { "-q", argFlag, &quiet, 0, "don't print any messages or errors" },
                                    { "-v", argFlag, &printVersion, 0, "print copyright and version info" },
                                    { "-h", argFlag, &printHelp, 0, "print usage information" },
@@ -188,6 +190,7 @@ int main(int argc, char *argv[])
             imgOut->enableJBig2(dumpJBIG2);
             imgOut->enableCCITT(dumpCCITT);
         }
+        imgOut->enablePrintFilenames(printFilenames);
         doc->displayPages(imgOut, firstPage, lastPage, 72, 72, 0, true, false, false);
     }
     const int exitCode = imgOut->isOk() ? 0 : imgOut->getErrorCode();