Bug 691116

Summary:	Greek Characters missing in conversion to TIFF
Product:	Ghostscript	Reporter:	Dave <dave.harrington>
Component:	PDF Interpreter	Assignee:	Alex Cherepanov <alex>
Status:	RESOLVED FIXED
Severity:	normal
Priority:	P4
Version:	master
Hardware:	PC
OS:	Windows XP
Customer:		Word Size:	---
Attachments:	Sample pictures of what is missing.doc Sample file used to create TIFF Simplified sample file patch

Description Dave 2010-02-17 15:31:23 UTC

When running GS version 8.14 and converting our PDF to a TIFF we can see some 
Greek characters that we can not see when taking the same PDF and converting it 
to a TIFF using 8.71.

Comment 1 Dave 2010-02-17 15:46:04 UTC

Created attachment 5962 [details]
Sample pictures of what is missing.doc

This is picture examples if I need to give you a PDF to convert please let me
knwo and I will try to get one to you

Comment 2 Hin-Tak Leung 2010-02-17 15:47:58 UTC

pdf would be preferable - there are too many ways of going from a word doc to a pdf.

Comment 3 Dave 2010-02-17 15:50:52 UTC

The word doc is examples of what I'm seeing I will try and get a PDF to you 
tommorrow

Comment 4 Dave 2010-02-18 07:55:06 UTC

Created attachment 5964 [details]
Sample file used to create TIFF

Here is the sample file that we ran on 8.14 and 8.71.

Comment 5 Alex Cherepanov 2010-02-23 20:22:19 UTC

Created attachment 5984 [details]
Simplified sample file

Comment 6 Alex Cherepanov 2010-02-23 20:25:30 UTC

This is some case of mixed-up encoding of in a TrueType font.
I'm taking over.

Comment 7 Alex Cherepanov 2010-02-27 12:31:55 UTC

Created attachment 5995 [details]
patch

Upgrade Adobe Glyph List to v. 2.0. 
AGL is used for character mapping in non-symbolic TrueType fonts in PDF files.

Many more characters, including Greek ones, are mapped correctly now.

The patch has been committed as a rev. 10822.
Regression testing shows no relevant differences. The different memory layout
affects one case of memory corruption.