Summary: | Regression: substituting a TrueType font for a CID font is broken | ||
---|---|---|---|
Product: | Ghostscript | Reporter: | Marcos H. Woehrmann <marcos.woehrmann> |
Component: | Text | Assignee: | Marcos H. Woehrmann <marcos.woehrmann> |
Status: | NOTIFIED FIXED | ||
Severity: | normal | ||
Priority: | P1 | ||
Version: | master | ||
Hardware: | PC | ||
OS: | Linux | ||
Customer: | Word Size: | --- |
Description
Marcos H. Woehrmann
2008-07-09 08:23:37 UTC
Created attachment 4197 [details]
64_384_6_180_5a_682822_361223.pdf
This bug is related to bug 689623 (and the test file is the same). Current version can load TrueType font using cidfmap without PostScript errors but the text is shown is incorrect. Rev. 8775 shows the same incorrect text. So there's no regression now. PDF interpreter can already load TrueType fonts as CIDFont resources. This hack shows correct text in the sample file on my box. So we need to scan FONTPATH and load a matching TrueType font the same way as the embedded font stream. --- E:\gs_svn\gs\lib\pdf_font.ps 2008-07-04 16:32:09.000000000 -0400 +++ pdf_font.ps 2008-07-12 00:29:38.000000000 -0400 @@ -1569,6 +1569,14 @@ } if } if } if + + + dup //null eq { + pop + dup /FontFile2 << /F (c:/winnt/fonts/arial.ttf) >> put + dup /FontFile2 get + } if + % /key res res desc stream dup //null ne { PDFfile fileposition Comment #3 is not correct; Ghostscript head (r8853) cannot load a TrueType font via cidfmap. Ghostscript 8.62 and earlier can. If my comment #0 cidfmap file is incorrect please post a copy of one that works with gshead. Returning to the author. Please check whether rev 8881 closes this one. If not, please attach the font file and assign to Leonardo. Confirmed fixed in r8881. |