Created attachment 23056 [details]
Panjabi for Pakistan (pa_pk) orthography

Created attachment 23057 [details]
Panjabi in India (pa_in) orthography

Roozbeh, this change caused issues for Qt which has a hardcoded list of languages.  Gah.

See: https://bugzilla.redhat.com/show_bug.cgi?id=497984

Should we rename pa-in to pa?  And, what's the glibc locale name for it?

To answer my own question, glibc has pa_IN and pa_PK.  Still, makes some sense to keep pa, if one is dominant?

(In reply to comment #0)
> The Punjabi orthography file is Gurumukhi only, while most speakers live in
> Pakistan and write in Arabic.
> 
Right, Most speaker are in Pakistan, but no one is using Shahmukhi script (actually part of Arabic), but Urdu only.

> glibc has a pa_PK locale too, in Arabic.
> 

Smiliar Bug was there in Unicode:
http://www.unicode.org/cldr/bugs/locale-bugs?findid=1300

(In reply to comment #4)
> To answer my own question, glibc has pa_IN and pa_PK.  Still, makes some sense
> to keep pa, if one is dominant?
> 

Yes, I would like this, but also this is mostly used. 
'pa' is default for 'pa-IN' only

I agree to the change, but not because of Qt depending on unpromised backward compatibility. Also for what is more dominant, I can't say the Indian one is more dominant, and the present orthography files don't follow any dominance architecture.

But the reason Mark Davis writes about here is valid (that the correct ISO 639 code for Pakistani Punjabi is actually 'lah': Lahnda):

http://www.unicode.org/cldr/bugs/locale-bugs?findid=1300

(Still, the whole term "Lahnda" to refer to the Pakistani Punjabi dialects is quite convtroversial.)

I will create a patch with all the references and renames in my git tree and will comment here when done.

Updated in my git tree, here: git://fedorapeople.org/~roozbeh/fontconfig.git

Roozbeh, any reason to keep pa_in?  I'd rather just rename pa-in.orth back to pa.orth.

(In reply to comment #9)
> Roozbeh, any reason to keep pa_in?  I'd rather just rename pa-in.orth back to
> pa.orth.

Makes sense. Updated in my git tree.

commit 844db442fb745f23087fe2a6f64c9cfcbe18ae47
Author: Roozbeh Pournader <roozbeh@gmail.com>
Date:   Thu May 7 14:31:11 2009 -0700

    Reorganize Panjabi/Punjabi and Lahnda orthographies (bug #19890)
    
    The correct ISO 639 code for Pakistani/Western Panjabi seems to be 'lah',
    not 'pa'. We are keeping 'pa_pk.orth' for compatiblity with glibc.
    
    Signed-off-by: Behdad Esfahbod <behdad@behdad.org>

I believe I've fixed this in 2.7.0.  Please reopen otherwise.