Re: new dependencies in gvfs

2009/5/31 Emmanuele Bassi <ebassi gmail com>:
> On Sun, 2009-05-31 at 10:37 +0200, Luca Ferretti wrote:
>> I know this is temporary, that sooner or later distro will include
>> needed stuff, but in next 6 months the only _stable_ distro
>> "certified" to test all GNOME 2.27 features will be Fedora 11.
> I fail to see the point: if another distro really wants to be "bleeding
> edge" enough to allow development on it then it should ship the bleeding
> edge dependencies. it's always been as simple as that.
> we're not talking about GNOME components here: we're talking about
> system level libraries; it happened with HAL, it happened with D-Bus, it
> happened with every low-level system component; JHBuild cannot always
> build them because JHBuild is not a distribution tool or a deployment
> tool[1].
> so you should start packaging DeviceKit for Ubuntu, if you want your
> distribution to be a development environment.

Here is the point, as I see it. g-d-u is proposed for inclusion in
2.28. It has a hard dependency on DeviceKit-disks and a lot of smaller
utility libraries. All those libraries must be used by a person
running the next version of gnome if you want to use the g-d-u

jhbuild is a tool that people use to QA the next release of
gnome. Since it is hard to build g-d-u (and it's not just me, there
are 3-4 dupes of #582096) and the libraries below it, that component
won't receive a lot of testing. I'd love to try g-d-u and find some
bugs, but I can't because someone has decided that building system
libraries is to dangerous for me.

Which means a high risk of a lot of serious bugs not being discovered
before release. I think that is problematic because g-d-u and friends
represent a huge amount of code, especially since those bugs may lead
to data losses.

It happened with HAL and it happened with D-Bus, those libs were tough
to jhbuild and introduced regressions in the desktop. And it will
happen with g-d-u too.

The solution is to make g-d-u play nice with jhbuild, which
technically isn't very hard.

mvh Björn

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