Kenichi Handa writes:
> In article <86d3z4eskv.fsf@chateau.d.if>, wahjava.ml@gmail.com (Ashish SHUKLA) writes:

>> Following are the characters (in sequence) mentioned in the above image:

>> ज़ - 095Bh
>> ा - 093Eh
>> क़ - 0958h
>> ि - 093Fh
>> र - 0930h

> Ah, I see.  I have forgotten to add consonants
> U+0958..U+095F and etc. in devanagari-composable-pattern.
> I've just installed the folloing patch in emacs-23 branch.

> ---
> Kenichi Handa
> handa@m17n.org

> === modified file 'lisp/language/indian.el'
> --- lisp/language/indian.el	2010-01-13 08:35:10 +0000
> +++ lisp/language/indian.el	2010-03-16 06:58:18 +0000
> @@ -141,7 +141,7 @@
>  (defconst devanagari-composable-pattern
>    (let ((table
>  	 '(("V" . "[\u0904-\u0914\u0960-\u0961\u0972]") ; independent vowel
> -	   ("C" . "[\u0915-\u0939]")		 ; consonant
> +	   ("C" . "[\u0915-\u0939\u0958-\u095F\u097B-\u097C\u097E-\u097F]") ; consonant
>  	   ("R" . "\u0930")			 ; RA
>  	   ("n" . "\u093C")			 ; NUKTA
>  	   ("H" . "\u094D")			 ; HALANT

Your patch works[1] for me.

References:
[1]  http://www.flickr.com/photos/wahjava/4437998520/

Thanks
-- 
Ashish SHUKLA