Re: GTK+ 2.4.4 released

On Wed, 2004-07-14 at 20:42, Federico Mena Quintero wrote:

> <flame>
> It's pretty ridiculous to advertise a toolkit that supports accesibility
> without having a full-time person dedicated to it.  We get a bunch of
> a11y bugs from Sun (that's a good thing), and sporadic patches (that's
> not so good).  But since we have no a11y experts in the GTK+ team, then
> sometimes people don't know if a patch is really right.

I agree, but why does Sun have to do it all?  It doesn't make sense for
the only fulltime accessibility engineers to all be from the same
company, for a number of reasons.  The patches to GTK+ from Padraig are
of high quality, and not at all sporadic.  It's the patch review process
on the GTK+ end that seems to be the primary bottleneck.  I don't know
what your comment about lacking a11y expertise means in the context of
review of patches from Padraig, as they are already reviewed for
correctness from an a11y perspective.

> Have you considered committing someone from Sun to work on GTK+'s a11y
> close to full-time?

Padraig's work on gtk+ accessibility, including ATK and libgail, is
nearly full time already.  We are already committing the maximum
available resources, but the resulting patches aren't getting reviewed
in a timely enough fashion to get them committed.  We have no commit
privileges for gtk+, so all we can do it revise patches ASAP after they
are reviewed.  We (Sun) don't have an army of people to do this work. 

Very few patches are accepted into gtk+ from the original patch
submitters anyhow, the standard procedure seems to be for the gtk+ core
team to rewrite them when they get time.  I don't think this is so much
a quality issue as a maintainership-style issue.  It does however make
things tough when the accessibility bugs/patches get put on the next
milestone, then fail to get committed before the release (which has
happenned several times in the GTK+ 2.4.X series).  The GtkCombo issue
was, for instance, put on the 2.4.0 milestone, then 2.4.2, then 2.4.4,
but still hasn't been committed.

- Bill

> </flame>
>   Federico

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]