GNOME Bugzilla – Bug 481223
[ta_IN] - Codepoint 0B83 showing wrong shape in Gtk applications
Last modified: 2010-04-08 23:05:42 UTC
Please describe the problem: https://bugzilla.redhat.com/show_bug.cgi?id=288901 Opened by A S Alam (aalam@redhat.com) on 2007-09-13 03:33 EST [reply] Private Description of problem: Tamil Codepoint - Various Sign Visarga - Unicode Value U+0B83 is showing wrong shape and appearing with dotted circle. Version-Release number of selected component (if applicable): pango-1.18.1-1.fc8 How reproducible: Everytime during use Steps to Reproduce: 1. Login into machine with tamil language support 2. Open gedit 3. Select RAWCODE from the SCIM table (or copy paste [ஃ] ) 4. Type 0b83 5. Observe the char appearing with dotted circle. Actual results: Char is appearing with dotted circle Expected results: Char should appear without dotted circle as its place independently everywhere. Additional info: font has not this issue. qt render it correctly. Comment #1 From A S Alam (aalam@redhat.com) on 2007-09-13 03:33 EST [reply] Private Created an attachment (id=194341) [edit] Screenshot showing Image of Character Steps to reproduce: Actual results: Expected results: Does this happen every time? Other information:
Created attachment 97464 [details] [review] my patch I wrote a patch for fixing this bug. It works fine.
Thanks to A.S. Alam and LingNing Zhang who had opened this nearly 20 months back. The patch of LingNing Zhang works alright in my Ubuntu desktops after I correct as per that patch and rebuild libpango1.0-x.. and install In the IndicClassTables.cpp file under icu which renders OpenOffice apps in GNU/Linux the U+0B83 has been set to _xx probably two years ago (just like what LingNing's patch prescribes for the indic-ot-class-tables.c) and so OpenOffice has no such problem. It can be clearly seen from the Unicode chart for Tamil : http://www.unicode.org/charts/PDF/Unicode-5.1/U51-0B80.pdf that Aytham (that's the Tamil name for the so-called Tamil Sign Visarga) does not have a dotted circle in front. Unicode standards older than 3.0 apparently had that mistake but was corrected from 4 onwards. As it stands now, Pango remains to be the only one still to change. Hope someone will triage this soon. K. Sethu
Hello. Can this very simple and elementary bug be fixed? The patch is just changing two simple bytes but this is affecting a lot of downstream applications. It should be a trivial thing.
Thanks all for reporting and commenting here. I have included this bug fix in bug 594101.
Behdad, Do you still see description of bug is not enough for accepting patch submitted in comment#1?
Further to description by LingNing Zhang and my comment#2, I wish to point to the Errata section of Unicode Standard Annex #28 (Unicode 3.2) in http://www.unicode.org/reports/tr28/tr28-3.html#errata wherein in the row 3 of the Table (entry dated 2001 September 25) the correction made to Unicode classifications to 0B83 with effect from Unicode 3.1 are shown. The removal of dotted circle is shown with figures. Hope this is useful information to speed up fixing this bug K. Sethu
K. Sethu, Thanks for your quick reply.
I now see that in my Ubuntu 10.04 (Lucid Lynx) Beta-1 up-to-date (4th April 2010), the latest version pango1.0-1.28.0 is bug fixed as per patch in comment #1. I thank all concerned upstream for triaging and fixing this bug. Another thing I observed which i need to get clarified. The Changelog file in the source folder pango1.0-1.28.0/ as well as the other Changelog in pango1.0-1.28.0/debian do not have any reference to fixing this bug or this bug report id. Is it act of omission or is it that such minor changes are usually not mentioned in Changelogs? Thanks in advance for reply K. Sethu
The bug was fixed as part of a larger Indic fix, hence the omission of the bug number for this one. Closing as per comment 8.