On Thu, 2008-01-10 at 11:15 -0600, Aaron Griffin wrote:
> On Jan 10, 2008 10:48 AM, Jan de Groot <email@example.com> wrote:
> > Hi all,
> > It's quite common knowledge that pygtk/pygobject programs are bad for
> > power usage. Newer versions of pygtk and pygobject contain code that
> > requires API from python trunk to solve this issue. However, our python
> > 2.5 doesn't have that code.
> > http://blogs.gnome.org/johan/2008/01/04/enough-wakeups-in-python-programs/
> > describes the issue, it's from the pygobject/pygtk developer that
> > integrated the new code.
> > http://cvs.fedora.redhat.com/viewcvs/devel/python/python2.6-set_wakeup_fd4.patch?rev=1.1
> > is the backport to python 2.5 done by redhat.
> > I would like to apply the python wakeup patch, put this new python in
> > testing and build the new pygobject/pygtk releases against it.
> > Though the added python API is development code, it's an API addition
> > that doesn't touch existing API, so our python will stay compatible with
> > old python scripts/libraries/etc. The downside of the patch is that this
> > new API could be not the final API, which requires us to update the
> > patch whenever it changes. As this patch has been developed by python
> > and pygtk developers together, I don't see problems here.
> > So the question is: is it ok to apply the patch to our python package?
> It sounds good to me assuming we have decent testing done for it.
The new python is in testing for both architectures. Please test
carefully. Take note that python is also in testing for the tcl/tk
upgrade, so this new patch can't get moved to extra before tcl/tk is in