On 08/27/2012 06:13 PM, Adam Jackson wrote:
On 8/27/12 9:10 AM, Adam Jackson wrote:
pangoxft should be a functional replacement, as far as I can tell, and
the APIs seem to be quite similar.
Actually I'm told pangocairo is the preferred thing instead of pangoxft.
repoquery agrees:
$ repoquery --whatrequires 'libpangocairo-1.0.so.0()(64bit)' | wc -l
1435
$ repoquery --whatrequires 'libpangoxft-1.0.so.0()(64bit)' | wc -l
11
But again, quite similar APIs.
Many of the broken packages appear to acquire their pangox dependency
through their use of gtkglext, which has gone an irritatingly long time
without an upstream release.
It has not seen an upstream update, because
- gtkglext's development has gone lost in the churn gtk3 and
Gnome3 have caused
+ All gtkglext maintainers have gone "silent".
+ Some people started working on a gtkglext3 port.
- There wasn't any need for another gtkglext release.
gtkglext's pangox use was removed with
this commit:
http://git.gnome.org/browse/gtkglext/commit/gdk/x11/Makefile.am?id=df7a7b35b80b395d7ba411c7f727970a46fb0588
But we can't just backport that, because removing those functions is an
ABI break, and those functions are indeed used in some of the affected
packages.
Correct ... This change is an ABI/API breakage, and therefore is not
applicable to Fedora nor to gtkglext.
Still, looking into it.
My view: The pango maintainers can not abandon pangox just because
_they_ do not use some API other works rely upon.
Alternatively, as last resorts,
* gtkglext either could adopt those pieces of code pango has dropped,
* Fedora could ship a compat-pangox package.
Ralf
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
