Issue 99241 - vocal sign on arabic letter hamza out of place
Summary: vocal sign on arabic letter hamza out of place
Status: CLOSED FIXED
Alias: None
Product: gsl
Classification: Code
Component: code (show other issues)
Version: OOO310m1
Hardware: PC Unix, all
: P3 Trivial (vote)
Target Milestone: OOo 3.1
Assignee: stefan.baltzer
QA Contact: issues@gsl
URL:
Keywords: regression
Depends on:
Blocks:
 
Reported: 2009-02-15 17:34 UTC by cemu
Modified: 2009-10-28 15:46 UTC (History)
4 users (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: ---
Developer Difficulty: ---


Attachments
example odt file (16.67 KB, application/vnd.oasis.opendocument.text)
2009-02-15 17:35 UTC, cemu
no flags Details
correct placement of signs below and above hamza (206.12 KB, image/jpeg)
2009-02-15 17:36 UTC, cemu
no flags Details
wrong placement of signs below and above hamza (208.24 KB, image/jpeg)
2009-02-15 17:37 UTC, cemu
no flags Details
screenshot after reverting trac6249-change (42.88 KB, image/png)
2009-02-17 08:58 UTC, hdu@apache.org
no flags Details

Note You need to log in before you can comment on or make changes to this issue.
Description cemu 2009-02-15 17:34:43 UTC
This bug is new in OOdev 3.1m1, in OO 3.0.1 it was correct.
On the arabic letter hamza the vocal sign is out of place, please look to the example files. In the
second attachment there is the correct placement in OO 3.0.1 (look inside the green circles), in
the third attachment there is the wrong placement in OOdev 3.1m1 (look inside the red circles).
There is too much space between the letter hamza and the vocal sign (or other signs) above and
below. Attachment one is the example odt file.
Comment 1 cemu 2009-02-15 17:35:26 UTC
Created attachment 60175 [details]
example odt file
Comment 2 cemu 2009-02-15 17:36:31 UTC
Created attachment 60176 [details]
correct placement of signs below and above hamza
Comment 3 cemu 2009-02-15 17:37:16 UTC
Created attachment 60177 [details]
wrong placement of signs below and above hamza
Comment 4 stefan.baltzer 2009-02-16 11:49:02 UTC
SBA: Set Prio to P3. For setting Prios, please see
http://qa.openoffice.org/scdocs/ddIssues_EnterModify.html#priority
Thank you for your comprehension.
Put hdu on CC.
Comment 5 hdu@apache.org 2009-02-16 12:15:57 UTC
The problem probably got introduced by the change from ICU 3.6+patches to plain ICU 4.0.
Comment 6 stefan.baltzer 2009-02-16 13:56:54 UTC
SBA: When verifying issue 86483 ("Upgrade icu to 4.0") (CWS i18n42, integrated
into DEV300_m35), it seems that Karl and me did not cover all possible
"surprises"...
Set keyword "regression".
Comment 7 stefan.baltzer 2009-02-16 15:19:47 UTC
CCing myself.
Comment 8 hdu@apache.org 2009-02-17 08:39:58 UTC
The fix for ICU-ticket 6249 (http://bugs.icu-project.org/trac/ticket/6249) has introduced this problem. 
Now I'm not sure yet whether the fix for that ICU-problem introduced a bug or if the font (Scheherezade) 
is the one to blame. At a first glance it is the font, because it seems to set a Non_Joining attribute for the 
problematic glyphs.

For the OOo31 release I'm about to revert the change http://bugs.icu-project.org/trac/changeset/24081  
which we got from upstream-ICU.
Comment 9 hdu@apache.org 2009-02-17 08:58:03 UTC
Created attachment 60229 [details]
screenshot after reverting trac6249-change
Comment 10 hdu@apache.org 2009-02-17 08:59:33 UTC
@cemu: please check the screenshot above
Comment 11 cemu 2009-02-17 19:58:20 UTC
Hamza handling looks correct now. I don't know if this is a Scheherazade (from scripts.sil.org)
font problem or not, but under MS Office, ArabTex or XeTeX (Scheherazade is standard font in
XeTeX for arabic) there is no problem.
I chose that font because it's seem to more "professional" than other fonts. There are many
other fonts that are not usable under OpenOffice together with arabic vocal signs (e.g. Arial).
Maybe there are more informations at scripts.sil.org (They have many unicode fonts and
information). Arabic font page is under http://scripts.sil.org/ArabicFonts (There is also some little
information about OpenOffice).
Comment 12 hdu@apache.org 2009-02-19 07:42:36 UTC
In CWS ooo31gsl2 the upstream ICU change "trac6249" is reverted -> FIXED.
The regression is reported upstream but as of now there is no feedback from the ICU team yet.
Comment 13 hdu@apache.org 2009-02-19 07:43:52 UTC
@sba: please verify in CWS ooo31gsl2
Comment 14 stefan.baltzer 2009-02-20 11:11:59 UTC
Verified in CWS ooo31gsl2.
Comment 15 ooo 2009-02-23 12:30:35 UTC
@hdu: please add to
http://wiki.services.openoffice.org/wiki/ICU/bugs_and_patches to keep track of
it, thanks.
Comment 16 hdu@apache.org 2009-02-23 13:46:02 UTC
Change the OOo&ICU wiki page:
http://wiki.services.openoffice.org/w/index.php?
title=ICU%2Fbugs_and_patches&diff=112812&oldid=99838
Comment 17 stefan.baltzer 2009-10-28 15:46:35 UTC
OK in OOO320_m2. Closed.