Apache OpenOffice (AOO) Bugzilla – Issue 11061
Bengali rendering related bugs
Last modified: 2008-05-17 22:58:47 UTC
Bengali rendering related bugs in OpenOffice.org developers' build version 643C =============================================================================== NOTE: I am selecting the language as Hindi - Bengali is not yet there - but the OTF rules are more or less the same. 1. Split vowels signs (to be more specific, Bengali vowel sign 'O' - 09CB and Bengali Vowel sign 'AU' - 09CC are not being rendered properly. The spacing and the alignment is getting all goofed up. 2. The conjuncts are getting displayed properly only when the screen is refreshed. In this specific case, I am minimising and maximising the window - and that seems to work fine. 3. The program is not being able to handle two of the Bengali Open Type Fonts - Akaash and Likhan from www.nongnu.org/freebangfont/. As soon any input is made while the font(s) is selected, the entire thing is autosaving and crashing. (it may also be a bug with the fonts - I am trying to find that out.) 4. Open Type rendering is not yet there in Impress and Draw.
DL->HDU: Would you please takeover?
The problem with the vowel signs seems to be wrong printer metrics. When selecting the menu item "View -> Online Layout" the rendering looks ok to me in 6.1beta. I'll have a closer look at this. The problem with the Akaash and Likhan fonts seem to be fixed in the OOo 1.1beta.
Just a hint: when you encounter a refresh problem in OOo, you may want to use the keyboard shortcut Ctrl+Shift+R to manually force a refresh.
Changing target as the window to Beta2 is closing.
This task cannot be fixed until OOo 1.1. I set this task to target OOo 2.0. Perhaps we can fix it in a Product Update for OOo 1.1.
In View->Online View mode and on the Win32 platform it looks ok. HDU->HDU: In Layout Mode on the Unx platform the problem happens when adjusting the glyph array to the reference device when the unicode to glyph mapping resulted in not continuous glyph clusters. There is a bug in this adjustation code.
.
According to the OpenOffice.org roadmap (http://tools.openoffice.org/releases) this issue was retargeted to 'OOo Later'.
Finally got access to a windows machine and did some testing in it. Apparently the rendering there is done by the Uniscribe Engine (usp10.dll) - not by the thingy that is doing it in Linux.
Yes. On UNX we use ICU + VCL-ICU glue code. I think the VCL-ICU glue code is at fault. Unfortunately the related sources are quite often called and we don't want to risk instability there until at least 2.0.
*The bug in rendering Malayalam Language (15883), Tamil (15013) and this bug are the same. So it will be effecting all Indian languages. Need immedate consideration
After some poking around, it looks as if the vowels are not being split at all. For example, a normal rendering engine should split U+09CB BENGALI VOWEL SIGN O into U+09C7 BENGALI VOWEL SIGN E and U+09BE BENGALI VOWEL SIGN AA. It seems as if the OOo rendering system is just placing the glyph for U+09CB BENGALI VOWEL SIGN O , without doing the splitting. Can anyone confirm?
This seems to be an issue with ICU, specifically, the ICU Layout Engine (libicule). The relevant ICU bug is at http://www.jtcsv.com/cgibin/icu-bugs/layout?id=4026;page=4;user=guest, and Ithink it has been resolved in the ICU 3x series. The ICU version (2.6) used by the snapshot builds of OOo 2.0 do not have this fix incorporated. It would be great if the relevant patches (http://oss.software.ibm.com/cvs/icu-jinfo?id=4026&ignore_attic=on) can be backported to ICU 2.6.
Thanks a lot for the pointer to the ICU bug and the diff! Another reason to upgrade to the latest ICU version.
The split vowel sign issue is solved by the patch at http://cvs.fedora.redhat.com/viewcvs/rpms/openoffice.org/devel/patches-vclplug-i18n-indic-bug-fixes.diff?rev=1.1&view=log However, the version (2.2) of ICU included in OOo 1.1.2 (the only version for which I have the source) has a quite a few bugs when it comes to Bengali rendering. This bugs were also present in Pango, so I am atm porting the fixes made to Pango to ICU 2.2. I'll try to release the patches for ICU (and a build incorporating these fixes) within the weekend. A screenshot of OOo rendering Bengali properly is at http://sayamindu.randomink.org/gallery/Screenshots/ooo_bn
So when the fedora indic patches fixes most of the problem then please check the latest version of SRC680 OOo again after the CWS vcl34 has been integrated. CWS vcl34 contains the fix for issue 39982, which is a modified version of the indic patch.
Created attachment 21506 [details] A simple test case for the problem mentioned in the first comment.
The test case above looks good to me in CWS vcl34.
Following characters do not render propery with other character in OOo 2.0 Beta, 09C1 BENGALI VOWEL SIGN U 09C2 BENGALI VOWEL SIGN UU 09C3 BENGALI VOWEL SIGN VOCALIC R 09C4 BENGALI VOWEL SIGN VOCALIC RR Screenshot of the bug is, http://www.bengalinux.org/screenshots/others/OOo_bug_sample.png Here is the shot of correct rendering, http://www.bengalinux.org/screenshots/others/OOo_sample.png
Created attachment 24158 [details] A simple test case for the problem mentioned above.
Sorry, I guess the problem I mentioned is font issue. I have filed another bug (#46144) for it. Thanks!
The Issue you raised has been marked as 'Resolved' and not updated within the last 1 year+. I am therefore setting this issue to 'Verified' as the first step towards Closing it. If you feel this is incorrect, please re-open the issue and add any comments. Many thanks, Andrew Cleaning-up and Closing old Issues ~ The Grand Bug Squash, pre v3 ~ http://marketing.openoffice.org/3.0/announcementbeta.html
As per previous posting: Verified -> Closed. A Closed Issue is a Happy Issue (TM). Regards, Andrew