Bug 689516

Summary:	problem reading PDF: Failed to interpret TT instructions (ref 7814)
Product:	Ghostscript	Reporter:	Marcos H. Woehrmann <marcos.woehrmann>
Component:	Text	Assignee:	leonardo <leonardo>
Status:	NOTIFIED FIXED
Severity:	normal	CC:	leonardo
Priority:	P2
Version:	master
Hardware:	PC
OS:	Linux
Customer:	700	Word Size:	---

Description Marcos H. Woehrmann 2007-10-16 08:16:02 UTC

The customer reports and I've verified that the attached generates an error with
gshead (r8296): 

GPL Ghostscript SVN PRE-RELEASE 8.61: Failed to interpret TT instructions in
font 80000006. Continue ignoring instructions of the font.
Error: /invalidfont in --run--

Both Acrobat Reader (v7.0) and evince display this file correctly, evince does
generate several warnings: "some font thing failed" (even more cryptic than our
messages :-).

The command line I'm using for testing:

  bin/gs -sDEVICE=ppmraw -sOutputFile=test.ppm ./A2.32_Floor_Plan_Level-2x.PDF

Comment 1 Marcos H. Woehrmann 2007-10-16 08:16:32 UTC

Created attachment 3475 [details]
A2.32_Floor_Plan_Level-2x.PDF

Comment 2 Ray Johnston 2007-10-16 10:17:23 UTC

This is probably similar to 689471.

Comment 3 Marcos H. Woehrmann 2008-03-11 16:38:16 UTC

Per request I'm retesting this customer's bugs:  this one still occurs with r8596.

Comment 4 Henry Stiles 2008-05-20 22:26:25 UTC

support has requested a status update for this problem.  Please add a new
comment entry or send mail to Marcos.

Comment 5 leonardo 2008-05-24 07:06:20 UTC

The test document embeds a TT font, in which  numGlyphs=5 and loca length = 
28. This may be obtained when running  with -ZY. The failed condition is 
gstype42.c ln 518 : 

 if (glyph_index >= pfont->data.trueNumGlyphs)
 	return_error(gs_error_invalidfont);

So I've got an idea how to work around : when loading a font. check if 
numGlyphs is smaller than locaa length. If so< chech whether all loca entries 
are correct. If so, fix numGlyphs and print warning.

I'm not sure that it will work fine, but chances look pretty good. So at this 
point we can talk to the customer that the document is incorrect, and we're 
trying to develop a workaround.

Comment 6 leonardo 2008-05-24 15:09:24 UTC

The suggested change works and passed the regression test with comparefiles, 
so I'll commit it soon.

Comment 7 leonardo 2008-05-25 00:31:16 UTC

Patch to HEAD :
http://ghostscript.com/pipermail/gs-cvs/2008-May/008354.html