Red Hat Bugzilla – Bug 431029
[ml_IN] Wrong combinations used for conjunct 'ന്റ'
Last modified: 2016-07-31 21:30:49 EDT
+++ This bug was initially created as a clone of Bug #402331 +++ Description of problem: In lohit-malayalam font, 0D28 + 0D4D + 200D + 0D31 (key 'V', key 'D', key ']' and key ' shift J' on malayalam inscript keyboard) gives the conjunct ' ന്റ ' . But the correct combination is 0D28 + 0D4D + 0D31 (type key 'V', key 'D' and key ' shift J' on malayalam inscript keyboard) which must give the conjunct ന്റ. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Open gedit/kedit. 2.Type unicode 0D28, 0D4D, 200D , 0D31(or type key 'V', key 'D', key ']' and key ' shift J' respectively using Malayalam Inscript). 3.Result is ന്റ. Actual results: 1.Open gedit/kedit. 2.Type unicode 0D28, 0D4D, 200D , 0D31(or type key 'V', key 'D', key ']' and key ' shift J' respectively using Malayalam Inscript). 3.Result is ന്റ. Expected results: 1.Open gedit/kedit. 2.Type unicodes 0D28 + 0D4D + 0D31 (or type key 'V', key 'D' and key ' shift J' respectively using Malayalam Inscript). 3.Result must be ന്റ. Additional info: -- Additional comment from aalam@redhat.com on 2007-12-14 01:28 EST -- fonts-malayalam-2.1.5-3.fc8 -- Additional comment from apeter@redhat.com on 2008-01-04 01:14 EST -- Appreciate if you could let me know the status of this bug as this has to be fixed immediately before next release. This is a very basic but very critical bug and hence must be fixed urgently. Thanks Ani -- Additional comment from apeter@redhat.com on 2008-01-04 01:14 EST -- Appreciate if you could let me know the status of this bug as this has to be fixed immediately before next release. This is a very basic but very critical bug and hence must be fixed urgently. Thanks Ani -- Additional comment from rbhalera@redhat.com on 2008-01-10 02:17 EST -- Ani,this appears to be working fine in f8. Can you please confirm the problem? -- Additional comment from apeter@redhat.com on 2008-01-10 04:20 EST -- Created an attachment (id=291249) Screen shot for the conjunct combination -- Additional comment from apeter@redhat.com on 2008-01-22 03:20 EST -- Rahul, we had a personal discussion on this bug last week. Is the issue clear for you or do you still have any doubts? Thanks Ani -- Additional comment from rbhalera@redhat.com on 2008-01-22 07:43 EST -- Ani, thanks for the concern. The problem is absolutely clear to me. It also has a relation the other bug about chillaksharam right now. -- Additional comment from rbhalera@redhat.com on 2008-01-22 08:41 EST -- Created an attachment (id=292509) patch for Lohit-Malayalam.sfd This patch fixes the ligature rule in the .sfd file. -- Additional comment from rbhalera@redhat.com on 2008-01-22 08:44 EST -- Created an attachment (id=292510) fixed ttf file Ani, can you please test this font file before I build it for fedora? -- Additional comment from apeter@redhat.com on 2008-01-23 06:10 EST -- Rahul , I still find this is not fixed completely. Now the result is just the key sequence when 0D28 + 0D4D + 0D31 is used. Problem : 0D28 + 0D4D + 200D + 0D31 has to be used for getting the conjunct. Expected result: 0D28 + 0D4D + 0D31 must give the desired conjunct. Please have a look on the screenshot attached on Comment #5 Let me know if any more queries. Thanks Ani -- Additional comment from apeter@redhat.com on 2008-01-23 06:22 EST -- Created an attachment (id=292624) Screenshot of the glyph of the conjunct -- Additional comment from apeter@redhat.com on 2008-01-23 06:23 EST -- Rahul , I still find this is not fixed completely. Now the result is just the key sequence when 0D28 + 0D4D + 0D31 is used. Problem : 0D28 + 0D4D + 200D + 0D31 has to be used for getting the conjunct. Expected result: 0D28 + 0D4D + 0D31 must give the desired conjunct. Please have a look on the screenshot attached on Comment #5 Let me know if any more queries. Thanks Ani -- Additional comment from apeter@redhat.com on 2008-01-23 06:25 EST -- Please ignore the comment #12 as its a repetition of comment #10. Instead please have a look in the follwing: Rahul, One more thing I would like to mention is that, now as the chillu is not working anywhere, currently 0D28 + 0D4D + 200D + 0D31 is also not giving the conjunct. I have already showed you the glyph of the conjunct in the font file. have attached the same as in comment #11 Thanks Ani -- Additional comment from rbhalera@redhat.com on 2008-01-23 08:08 EST -- Ani, from the screenshot in comment #5 attached by you, I assume you do not want the conjunct to be formed for 0D28 + 0D4D + 200D + 0D31? -- Additional comment from rbhalera@redhat.com on 2008-01-23 08:17 EST -- Created an attachment (id=292633) patch for Lohit-Malayalam.sfd This patch fixes the ligature rule in sfd file. -- Additional comment from rbhalera@redhat.com on 2008-01-23 08:17 EST -- Created an attachment (id=292634) fixed ttf file Ani, can you please test this font file now? -- Additional comment from apeter@redhat.com on 2008-01-23 23:39 EST -- Rahul, Yes, its working fine. Good job. Thanks Ani -- Additional comment from rbhalera@redhat.com on 2008-01-25 07:04 EST -- Fixed in lohit-fonts-malayalam-2.1.8-1.fc9.
Hi Caolon, This is a special case where 0D28 + 0D4D + 0D31 gives a special glpyh. Have included this in test case too. Will provide you the test case number once its available on batsman. screenshot attached. Currently, this is fixed in pango, fonts-malayalam and working perfect with gedit. Thanks Ani
Created attachment 293558 [details] Screenshot for 0D28 + 0D4D + 0D31 (special case)
Caolon, I believe this is not an ICU bug. The bug was fixed in font (#402331) and no modifications in pango were required to make it work on pango. So I guess there should not be any modifications in ICU as well. If at all the same font is not working on ICU, I think its because of the special issue of 0x0D31, which is fixed in pango and filed for ICU on #424661. Thanks.
Reverted the hack that we had in place in icu to get the original results for the rest of the something + 0d4d + 0d31 ones that have now been changed. So should be good in icu-3.8.1-4.fc9. The same hack for the something + 0d4d + 0d30 test-cases to get the right results remains, which does sort of me dubious about the hack in icu as pango remains "broken" on those test-cases.
Hi Caolon, Working perfect in icu-3.8.1-4.fc9. Tested and confirmed. Thanks Ani
oops.. my apologies for writing your name wrong Caolan. Best regards Ani
sounds good