[PATCH 1/2] cairo: only use show_text_glyphs if the surface supports it and the font has toUnicode

[Adrian Johnson]

When generating pdf output, fonts that do not have toUnicode cause an
excessive number of empty text ActualText entries to be written due to
glyphs that do not have have a mapping.
---
 poppler/CairoOutputDev.cc |   12 ++++++++----
 poppler/CairoOutputDev.h  |    1 +
 poppler/GfxFont.h         |    3 +++
 3 files changed, 12 insertions(+), 4 deletions(-)

diff --git a/poppler/CairoOutputDev.cc b/poppler/CairoOutputDev.cc
index 08c28ec..e894b4e 100644
--- a/poppler/CairoOutputDev.cc
+++ b/poppler/CairoOutputDev.cc
@@ -138,6 +138,7 @@ CairoOutputDev::CairoOutputDev() {
   currentFont = NULL;
   prescaleImages = gTrue;
   printing = gTrue;
+  use_show_text_glyphs = gFalse;
   inType3Char = gFalse;
   t3_glyph_has_bbox = gFalse;
 
@@ -578,6 +579,9 @@ void CairoOutputDev::updateFont(GfxState *state) {
 
   font_face = currentFont->getFontFace();
   cairo_set_font_face (cairo, font_face);
+
+  use_show_text_glyphs = state->getFont()->hasToUnicodeCMap() &&
+    cairo_surface_has_show_text_glyphs (cairo_get_target (cairo));
  
   double fontSize = state->getFontSize();
   double *m = state->getTextMat();
@@ -1001,7 +1005,7 @@ void CairoOutputDev::beginString(GfxState *state, GooString *s)
 
   glyphs = (cairo_glyph_t *) gmallocn (len, sizeof (cairo_glyph_t));
   glyphCount = 0;
-  if (printing) {
+  if (use_show_text_glyphs) {
     clusters = (cairo_text_cluster_t *) gmallocn (len, sizeof (cairo_text_cluster_t));
     clusterCount = 0;
     utf8Max = len*2; // start with twice the number of glyphs. we will realloc if we need more.
@@ -1020,7 +1024,7 @@ void CairoOutputDev::drawChar(GfxState *state, double x, double y,
     glyphs[glyphCount].x = x - originX;
     glyphs[glyphCount].y = y - originY;
     glyphCount++;
-    if (printing) {
+    if (use_show_text_glyphs) {
       if (utf8Max - utf8Count < uLen*6) {
         // utf8 encoded characters can be up to 6 bytes
 	if (utf8Max > uLen*6)
@@ -1070,7 +1074,7 @@ void CairoOutputDev::endString(GfxState *state)
   if (!(render & 1) && !haveCSPattern) {
     LOG (printf ("fill string\n"));
     cairo_set_source (cairo, fill_pattern);
-    if (printing)
+    if (use_show_text_glyphs)
       cairo_show_text_glyphs (cairo, utf8, utf8Count, glyphs, glyphCount, clusters, clusterCount, (cairo_text_cluster_flags_t)0);
     else
         cairo_show_glyphs (cairo, glyphs, glyphCount);
@@ -1118,7 +1122,7 @@ void CairoOutputDev::endString(GfxState *state)
 
   gfree (glyphs);
   glyphs = NULL;
-  if (printing) {
+  if (use_show_text_glyphs) {
     gfree (clusters);
     clusters = NULL;
     gfree (utf8);
diff --git a/poppler/CairoOutputDev.h b/poppler/CairoOutputDev.h
index 67f7810..386e844 100644
--- a/poppler/CairoOutputDev.h
+++ b/poppler/CairoOutputDev.h
@@ -301,6 +301,7 @@ protected:
   cairo_matrix_t orig_matrix;
   GBool needFontUpdate;                // set when the font needs to be updated
   GBool printing;
+  GBool use_show_text_glyphs;
   cairo_surface_t *surface;
   cairo_glyph_t *glyphs;
   int glyphCount;
diff --git a/poppler/GfxFont.h b/poppler/GfxFont.h
index 09af062..9550398 100644
--- a/poppler/GfxFont.h
+++ b/poppler/GfxFont.h
@@ -221,6 +221,9 @@ public:
 			  Unicode **u, int *uLen,
 			  double *dx, double *dy, double *ox, double *oy) = 0;
 
+  // Does this font have a toUnicode map?
+  GBool hasToUnicodeCMap() { return hasToUnicode; }
+
   /* XXX: dfp shouldn't be public, however the font finding code is currently in
    * GlobalParams. Instead it should be inside the GfxFont class. However,
    * getDisplayFont currently uses FCcfg so moving it is not as simple. */
-- 
1.7.4.1


--------------070204050201080606010207
Content-Type: text/x-patch;
 name="0002-update-SEE-ALSO-section-of-man-pages.patch"
Content-Transfer-Encoding: 8bit
Content-Disposition: attachment;
 filename="0002-update-SEE-ALSO-section-of-man-pages.patch"