In my tree:
Created commit 8a5f273: Add Sindhi .orth file. (#17140)
 1 files changed, 28 insertions(+), 0 deletions(-)

Behdad, would you please consider checking in this file under sd_in.orth instead? It seems that the vast majority of Sindhi speakers live in Pakistan, and use the Arabic script.

(In reply to comment #2)
> Behdad, would you please consider checking in this file under sd_in.orth
> instead? It seems that the vast majority of Sindhi speakers live in Pakistan,
> and use the Arabic script.

Humm.  The question is, what's in the "sd" glibc locale?  And it should be "sd-in.orth", not "sd-in.orth".

(In reply to comment #3)
> The question is, what's in the "sd" glibc locale?

glibc has an sd_IN and an sd_IN@devanagari. Interestingly, sd_IN is in Arabic, while sd_IN@devenagari is of course in Devenagari. Wikipedia says Indian Sindhi speakers write in both Arabic and Devanagari.

glibc doesn't have sd_PK (if it had, it would be Arabic, of course).

> And it should be "sd-in.orth", not "sd-in.orth".

All current file names in the fc-lang directory seem to use underscores (they get converted to "-" internally somehow).

OK, I did more research, and found that even in India, most Sindhi speakers write in the Arabic script, not Devanagari. For example, see this report, created by the government of India:

http://tdil.mit.gov.in/sindhidesignguideoct02.pdf

Quote: "The present script predominantly used in Sindh as
well as in many states in India and elsewhere where
migrant Sindhis have settled, is Arabic in Naskh
styles having fifty two alphabets. However, in some
circles in India, Devanagari is used for writing
Sindhi."

So even renaming this to sd_in would be wrong, as sd_in is also Arabic. I will await a resolution on bug 19869 and bug 17208.

Behdad, I think the present sd.orth file in your tree should either be removed or renamed to something like sd_Deva, sd_in_x_devanagari, or sd_in@devanagari.

I will try to create a correct sd.orth using the Arabic script.

Created attachment 22452 [details]
corrected Sindhi (sd) orthography using Arabic script

In my tree:

commit e26f934469b537350732d0c5cdaf66e02590745c
Author: Roozbeh Pournader <roozbeh@gmail.com>
Date:   Sun Feb 1 18:32:21 2009 -0800

    Correct Sindhi orthography to use Arabic script (bug #17140)
    
    The previous version used the Devanagari script. But in both Pakistan and
    India, Sindhi is generally written in Arabic. The Devanagari data could
    prove to be useful, if we decide on how we should name such files (see
    bug #17208 and bug #19869).

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