[cairo-bugs] [Bug 70364] New: Umlaut conversion error for Type1 fonts in WinAscii Encoding

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=70364

          Priority: medium
            Bug ID: 70364
          Assignee: ajohnson at redneon.com
           Summary: Umlaut conversion error for Type1 fonts in WinAscii
                    Encoding
        QA Contact: cairo-bugs at cairographics.org
          Severity: normal
    Classification: Unclassified
                OS: Linux (All)
          Reporter: winter at bfw-online.de
          Hardware: x86 (IA32)
            Status: NEW
           Version: unspecified
         Component: postscript backend
           Product: cairo

Created attachment 87422
  --> https://bugs.freedesktop.org/attachment.cgi?id=87422&action=edit
Input document which is printed incorrectly

Hi,

when printing (or running the print preview of) the attached PDF with evince
(utilizing the PostScript backend of Cairo of course), the umlauts (like ö) are
lost. The fonts used in the PDF are not embedded, the PDF seems to be encoded
in a Windows charset and the resulting PostScript seems to include the combined
characters for umlauts like 'odieresis', however it cannot render them hinting
the possibility it is a combined glyph made up of 'o' and 'dieresis' and one of
those might be missing. However my tries to manually include 'dieresis' in the
resulting type1 font subset dict were futile.
Also probably noteworthy but not relevant to this bug is the fact that xpdf
crashes when one tries to open the PDF, which suggests it might by a "strange"
PDF.

My approach of including 'dieresis' in the dictionary was (patched against
cairo-1.12.16 of Debian, the git log of the file seemed to suggest there were
no later patches to the file after the Debian release):

diff --git a/src/cairo-type1-subset.c b/src/cairo-type1-subset.c
index 4b64403..dfcd1a1 100644
--- a/src/cairo-type1-subset.c
+++ b/src/cairo-type1-subset.c
@@ -1407,6 +1407,15 @@ skip_subrs:
        cairo_type1_font_subset_use_glyph (font, index);
     }

+    for (i = 0; i < font->base.num_glyphs; i++) if (font->glyph_names[i])
+    { fprintf (stderr, "glyph: %s\n", font->glyph_names[i]);
+      if (!strcmp (font->glyph_names[i], "dieresis")
+       || !strcmp (font->glyph_names[i], "odieresis") || !strcmp
(font->glyph_names[i], "o"))
+      { cairo_type1_font_subset_use_glyph (font, i);
+        fprintf (stderr, "-----> added\n");
+      }
+    }
+
     /* Go through the charstring of each glyph in use, get the glyph
      * width and figure out which extra glyphs may be required by the
      * seac operator (which may cause font->num_glyphs to increase

Best regards,
Leon

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cairographics.org/archives/cairo-bugs/attachments/20131011/f18a4583/attachment.html>