Bug 200206
Summary: | [bn-IN] Incorrect spellings in localedef-components | ||
---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Runa Bhattacharjee <runab> |
Component: | glibc | Assignee: | Jakub Jelinek <jakub> |
Status: | CLOSED RAWHIDE | QA Contact: | Brian Brock <bbrock> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | rawhide | CC: | ankit, drepper, eng-i18n-bugs, fweimer, majain |
Target Milestone: | --- | Keywords: | i18n |
Target Release: | --- | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Fixed In Version: | 2.4.90-23 | Doc Type: | Bug Fix |
Doc Text: | Story Points: | --- | |
Clone Of: | Environment: | ||
Last Closed: | 2006-08-22 15:35:41 UTC | Type: | --- |
Regression: | --- | Mount Type: | --- |
Documentation: | --- | CRM: | |
Verified Versions: | Category: | --- | |
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
Cloudforms Team: | --- | Target Upstream Version: | |
Embargoed: | |||
Bug Depends On: | |||
Bug Blocks: | 200230 | ||
Attachments: |
Description of problem: Some contents in the Date section of the localedef had incorrect spellings Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1.From the gdm, choose language as Bengali India 2. Click on Date on the mail panel 3. Actual results: Tuesday shows as "মà¦à¦à¦²à¦¬à¦¾à¦°" with short form as "মà¦à§à¦à¦²: Expected results: Should be spelt as "মà¦à§à¦à¦²à¦¬à¦¾à¦°" with short form as "মà¦à§à¦à¦²" Additional info: Corrected localedef file for bn-IN locale is attached Why do you change so much in this file? All possible copy statements should be retained. This applies at least to LC_MEASUREMENT, LC_TELEPHONE. I haven't checked the rest. Please remove these. Also, in LC_MESSAGES, the strings differ from the bn_BD content. Why is there a difference? If the new text in your bn_IN file is correct, shouldn't then the text in bn_BD be fixed? I doubt that the spelling is different just because of the country. Is it possible to unify the LC_TIME section of bn_HI and bn_BD? If yes, please move the real definition to bn_BD. Finally, you shouldn't send entire files. Use the diff command on the command line and send the output. (In reply to comment #2) > Why do you change so much in this file? All possible copy statements should be > retained. This applies at least to LC_MEASUREMENT, LC_TELEPHONE. I haven't > checked the rest. Please remove these. These are copied from the hi_IN localedef file. In case a user decides to install only the bn_IN locale besides the en_US (and hi_IN is not installed) will the bn_IN get the hi_IN file to replicate values from? iirc, the values were put for both the bn_IN and as_IN files as a failsafe. If they are not required I can remove them. > > Also, in LC_MESSAGES, the strings differ from the bn_BD content. Why is there a > difference? If the new text in your bn_IN file is correct, shouldn't then the > text in bn_BD be fixed? I doubt that the spelling is different just because of > the country. There are spelling differences in the two countries. Secondly there were some spelling errors in the original bn_IN file which was a rehashed version of the bn_BD file. bn_BD localedef file is managed by the bn community, hence I personally do not think it prudent to modify it. > > Is it possible to unify the LC_TIME section of bn_HI and bn_BD? If yes, please > move the real definition to bn_BD. Not sure about bn_HI. > > Finally, you shouldn't send entire files. Use the diff command on the command > line and send the output. > Sorry for that. Shall send the diffs from now on. copy statements affect locale sources only, those are small and always the whole set is installed. The compiled locale files created by localedef are self-contained, even if the source contained copy statements. Created attachment 133777 [details]
diff from the bn_IN original localedef file and the modifed bn_IN localedef file
Attaching a diff file.
Modified:
LC_TIME
Entries for Tuesday, January and February had incorrect spellings, which have
been corrected. As mentioned earlier the bn_BD file is maintained by members of
the community and I am informing them. I am part of the bn_IN community.
Added:
LC_PAPER
----------------------------------------
(in respect to earlier attached file):
Reverted:
LC_MEASUREMENT
LC_TELEPHONE
LC_MESSAGES
In LC_MESSAGES (to be copied from bn_CD file), the text for "yesexpr" and
"noexpr" are respectively [<localized yes>yY] and [<localized no>nN]. Is this
the correct format for indic languages? This issue is highlighted as the hi_IN
(hindi provides the basic format for a number of indic languages) file uses the
format [yY] and [nN] *without* the localized 'yes', 'no' content in the yesexpr
and noexpr entities.
--------------------------------------------------------------------------------
Folks can use script in https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=200230#c28 to test the spellings/unicode-points in the locale. Created attachment 134004 [details]
Locale Definition - All Day bn_IN text putout
This is the locale definition output redirected to the text file instead of
terminal. Pls have a look.
Upstream bug - http://sources.redhat.com/bugzilla/show_bug.cgi?id=3034 Please attach the final patch for bn_IN locale to upstream bug too, once its finalized here. Thanks, Mayank Created attachment 134283 [details]
Latest patch, used "diff -uNrp"
Patch generated with "diff -uNrp" command, hence obsoleting previous patch.
Also, checked the resultant bn_IN file with "localedef -i bn_IN -f UTF-8
/usr/lib/locale/bn_IN" command.
Over to you Jakub :) Thanks, Mayank Patch is upstream, will be in next rawhide build. The rawhide build with the changes is 2.4.90-23. Maybe even an earlier one. Please test that code. Thanks guys :) |
Created attachment 133049 [details] Corrected localedef file for Bengali India (bn_IN) locale