[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ft-devel] Regarding the 2.1.10 release
From: |
Owen Taylor |
Subject: |
Re: [ft-devel] Regarding the 2.1.10 release |
Date: |
Sat, 26 Mar 2005 09:09:15 -0500 |
I've always been quite willing to accept that if FreeType changes
break Pango, I'll fix Pango. But I'm a little worried about timing
here ... removing the internal headers completely is going
to require pretty major changes to the opentype code ... so I'd
like to have some time to try those changes in Pango HEAD before
I need to do a Pango-1.8.x release with the changes.
Do you have a planned timescale for when the final 2.1.10 is
planned? If we can do the pango/opentype changes to make it
use the raw tables in the next few weeks, and still have a few
more weeks to test that in HEAD, then that probably will work
fine.
Another possiblity would be to up the major version of FreeType to
allow people to keep using Pango compiled against the old version.
I don't think this is a good idea - you can get bad problems if an
application is linked against libfreetype.so.6 but a library it depends
upon (fontconfig, Pango, whatever) is linked against libfreetype.so.7.
So, basically, changing the major version of FreeType requires
immediately fixing uses of the internal headers and rebuilding all
apps on the system. The main difference from not changing the
major number is that innocent applications that don't use
FreeType internals also have to be rebuilt.
In terms of the opentype code ... I'd certainly like to see it
being maintained someplace shared rather than copied around. My
thoughts last summer are outlined in:
http://lists.gnu.org/archive/html/freetype-devel/2004-08/msg00036.html
Basically to redo the otlayout/ module in freetype CVS starting
from the current Pango code. But I never really got everybody
"signed off on" that plan, and other development priorities arose.
But I think it's still a good idea. Behdad Esfahbod (cc'ed) is
starting to look at doing some major work on the opentype code
in Pango, so maybe he'd be able to help out moving to an
independent code base.
(I think rewriting to use Werner's validator should be done
separately.)
Regards,
Owen
[ Note, I'm away from my email until 3/31 ]
signature.asc
Description: This is a digitally signed message part