Sugar and GTK updates
pbrobinson at gmail.com
Tue Aug 16 16:49:29 EDT 2011
Thank you Daniel,
This is what I've been saying for some time!
On Tue, Aug 16, 2011 at 8:28 PM, Daniel Drake <dsd at laptop.org> wrote:
> On Tue, Aug 16, 2011 at 6:49 PM, Benjamin M. Schwartz
> <bmschwar at fas.harvard.edu> wrote:
>> 2. The system requirements (especially disk space) are affected more by
>> changes in Fedora than changes in Sugar. A large amount of disk space is
>> taken up by files whose presence is unnecessary. Customizing the build to
>> exclude these files takes significant human effort to execute and test
>> (see http://dev.laptop.org/ticket/4281), especially if the tradeoff is
>> different for XO-1 vs. XO-1.5.
> And just to build on this point a bit more...
> The changes referred to in the mail that opened this thread - GTK3 and
> PyGI - are not expected to be items of bloat. Rather the opposite.
> Moving to PyGI is expected to reduce startup time and reduce memory
> usage. No need for me to repeat Tomeu's arguments;
> This move will also allow us to drop a load of bindings packages, as
> the way PyGI works is that it allows python to call into C libraries
> through only a small generic introspection layer, plus a few small
> introspection data files. That will save some disk space.
> Moving from GTK2 to GTK3 is not a significant change. The version
> number bump was mostly for some internal restructuring and
> backwards-incompatible API changes, the actual end result is not that
> different from GTK2. Some things are faster and smaller, and GTK3 .so
> size has dropped 0.2mb to 4.1mb size on my system. A few things might
> be slightly slower, but the benchmarks I saw indicating this largely
> blamed the GTK3 theme for the small performance losses (which Sugar
> won't be using), and you won't notice it.
> Right now we ship both Mozilla and webkit in our builds. Both steps
> combined will allow us to drop mozilla, saving 30-50mb of disk space.
> I'll let others jump in with the "and webkit is faster/better"
> Sugar is growing features, but nothing that I feel that challenges the
> XO-1 at this point. The XO-1 images are big, and Fedora creep from
> 9-11 and 11-14 has hurt us a little, but we expect XO-1 deployments to
> ship without GNOME (which several do, at least), so the issue isn't as
> big as it appears.
Part of the growth in F-14 was new features like webM and other things
that we so actually want and get benefit from. Some of the growth was
due to changes for gnome3 that didn't end up making it to Fedora 14
that will be reduced in upcoming releases as things are deprecated and
cleaned up post gnome3 (believe me, I've found out how much of a mess
this was from building the entire distro from scratch for ARM)!
I've been working on keeping Fedora's deps in check for quite some
time (from around Fedora 8 or 9 if my memory serves me correctly). Its
unfortunately one of those pretty thankless and never ending jobs.
There's always improvements that can be made and there's always things
that people miss. If people pick up on things where we can split
certain features out into sub-packages so we don't pull in big deps do
let me know, I will get the changes upstream, even if I do them
More information about the Devel