Issue 102603

Summary: Hebrew vowel points (e.g., holam) misplaced on PDF export
Product: gsl Reporter: kaminsky <kaminsky>
Component: codeAssignee: h.ilter
Status: CLOSED FIXED QA Contact: issues@gsl <issues>
Severity: Trivial    
Priority: P3 CC: hdu, issues, philipp.lohmann
Version: OOO310m9Keywords: regression
Target Milestone: OOo 3.1.1   
Hardware: PC   
OS: Unix, all   
Issue Type: DEFECT Latest Confirmation in: ---
Developer Difficulty: ---
Issue Depends on:    
Issue Blocks: 108760, 101565    
Attachments:
Description Flags
sample Writer document with various holam tests (on-screen looks okay)
none
PDF export of sample holam test
none
screenshot of same Writer document
none
uncompressed version of the problematic PDF none

Description kaminsky 2009-06-09 06:10:19 UTC
[Actually, my OO reports version OOO311m11, but that wasn't an option above.]

When exporting a Hebrew document with vowel points, the spacing is off for the
holam point (a single dot above/between letters).  Unicode seems to have two
different versions of the holam--one that goes above a letter and the other that
is supposed to replace a vav.  Both seem to be mis-rendered on PDF export even
though they appear okay on-screen (sort of).

I'll attach a sample odt document and the corresponding pdf showing the problems
with both types of holam points.  Note that depending on what the next letter
is, the problem manifests it self differently.  Aleph is particularly
problematic, but even that depends on whether the aleph itself has a vowel under
it.  Note that even on-screen in Writer, the dot shifts position between lines 1
and 2.  These documents use the SBL Hebrew font.
Comment 1 kaminsky 2009-06-09 06:11:51 UTC
Created attachment 62873 [details]
sample Writer document with various holam tests (on-screen looks okay)
Comment 2 kaminsky 2009-06-09 06:12:31 UTC
Created attachment 62874 [details]
PDF export of sample holam test
Comment 3 kaminsky 2009-06-09 06:14:01 UTC
Created attachment 62875 [details]
screenshot of same Writer document
Comment 4 kaminsky 2009-06-09 06:16:57 UTC
Sorry, I had a typo on the version, it's actually: OOO310m11 (build 9399).
Comment 5 hdu@apache.org 2009-06-09 11:08:09 UTC
Confirming and reassigning.
kaminsky: great bugdocs, thanks!
Comment 6 hdu@apache.org 2009-06-09 15:47:48 UTC
Created attachment 62887 [details]
uncompressed version of the problematic PDF
Comment 7 hdu@apache.org 2009-06-09 15:51:20 UTC
The uncompressed PDF shows that the exporter code does some extra effort to push the glyphs for 
U+05B9 and U+05BA to the wrong position.
Comment 8 hdu@apache.org 2009-06-09 16:26:34 UTC
This problem was introduced with the workaround for issue 100522. Disabling the fabs() solves it. Also 
the workaround in issue 100522 that suggested first would have solved it.
@pl: ping
Comment 9 hdu@apache.org 2009-06-09 16:27:38 UTC
*** Issue 101671 has been marked as a duplicate of this issue. ***
Comment 10 hdu@apache.org 2009-06-10 14:21:23 UTC
Added the regression keyword. This deserves at least a 3.2 target.
Comment 11 Olaf Felka 2009-06-10 14:53:07 UTC
Setting target 3.1.1.
Comment 12 hdu@apache.org 2009-06-10 15:07:00 UTC
Fixed in CWS ooo311gsl04.
Comment 13 hdu@apache.org 2009-06-15 16:03:13 UTC
.
Comment 14 hdu@apache.org 2009-06-23 14:41:52 UTC
@hi: please verify in CWS ooo311gsl04
Comment 15 h.ilter 2009-06-25 15:26:07 UTC
Verified with cws OOO311gsl04 = ok
Comment 16 hdu@apache.org 2009-07-08 08:16:38 UTC
*** Issue 103400 has been marked as a duplicate of this issue. ***
Comment 17 hdu@apache.org 2009-07-08 11:50:02 UTC
*** Issue 30153 has been marked as a duplicate of this issue. ***
Comment 18 thorsten.ziehm 2010-06-07 11:49:21 UTC
This issue is closed automatically. It is in state 'verified/fixed' since 2
releases (OOo 3.1.1 and OOo 3.2). The policy [1] indicates that such older
issues should be closed.

If this issue still occur in a current build (OOo 3.2.1 or >DEV300m80) please
reopen the issue and set the target accordingly.

[1] : http://wiki.services.openoffice.org/wiki/Handle_fixed_verified_issues