[Bug 540390] [all_langs except ml_IN] Boxes appearing on Font GUI instead of latin characters

bugzilla at redhat.com bugzilla at redhat.com
Tue Dec 1 16:26:57 UTC 2009

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


K. Sethu <skhome at gmail.com> changed:

           What    |Removed                     |Added
                 CC|                            |skhome at gmail.com

--- Comment #3 from K. Sethu <skhome at gmail.com>  2009-12-01 11:26:56 EDT ---
When wanting to check for the presence or the absence of this bug in a
different language locale, rather than going through the procedure of a log-out
and the log back in with another locale, it is more convenient to just launch
from a terminal the gnome-appearance-properties applet together with the export
of LANG environment variable set to the relevant language 

For e.g., for Tamil,regardless of with which language locale the current
session is being run, from a terminal launch the applet with the following

export LANG=ta_IN && gnome-appearance-properties

Similarly the applet can be launched for any number of other languages from the
different required number of terminals. They can be viewed simultaneously for
making comparisons.  

I made the following screen-shot in Fedora 11 wherein the 3 shown simultaneous
instances of the gnome-appearance-properties are for US English (LANG=
en_US.UTF-8), Tamil LANG=ta_IN) and Kannada (LANG=kn_IN) : 


The additional image in the same screen-shot is a gnome-terminal in which I had
checked which fonts Fontconfig had matched for the generic font "serif" in case
of each of the 3 languages. It is the "serif" (as I have empirically found and
proved to myself) that is used for those sample glyphs cages in the "Appearance
Preferences" dialog. 

The fact that whenever the Fontconfig framework has to match a font lacking the
Latin range glyphs to "serif" this bug being caused to occur, points to two

First and more likely cause is that the applet could be faulty in that while
making a font match it is not passing to Fontconfig the correct information
that the sample glyph data are only fixed ascii range codes meant to be Latin
characters only and not as per the language locale in use. The other
possibility is the sample data is up for l10n translations which have not been
carried out and the Latin data are repeated falsely for each language. 

I haven't checked the source codes and my conjecture are speculative.

Well, hopefully upstream fixing would solve this bug once and for all. However 
I like to suggest that it is better to also include Latin range glyphs (at
least English) in all Lohit fonts to make them more competitive in the field.


Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

More information about the fonts-bugs mailing list