[gentoo-dev] Help needed for maintaining GNOME

[gentoo-dev] Help needed for maintaining GNOME

[Mart Raudsepp]

Hello


The state of GNOME in Gentoo could use a bunch of work and we don't
really have the developers currently to take care of it.
My own grandiose plans to be back and keep things in good order faced
sad reality (a bunch of personal things) and we could really use some
help in at least carefully reviewing and merging a contributors lots of
work, and hopefully more.
While I don't have the time, energy and currently motivation to be
hands-on in the git tree right now, I am online on IRC rather
constantly (lets say while waiting on work rust code compilation ;) and
happy to help on-board an interested developer or developers with any
peculiarities (which hopefully would get documented in the process),
questions, module interactions, whatever is needed, including during PR
reviews, just please do keep in touch, so I can offload things from my
brain for a more sustainable future.

Given a huge stack of available pull requests from a contributor, the
main need right now is someone with a Gentoo developer hat who can help
review these and get good things merged, fixed what's needed, and keep
an eye out on the bugs. Of course contributors are welcome too, but
please keep an eye out for not duplicating work that's already waiting
review.

To my knowledge, one big issue right now is a circular dep between
glib->gobject-introspection->glib, which needs proper solving to move
forward with things. Or at least look at reducing the glib requirement
in gnome-shell some way to at least unleash that core stack for our
users. There are ideas we can talk about.


Mart

Re: [gentoo-dev] Help needed for maintaining GNOME

[Eli Schwartz]

[-- Attachment #1.1: Type: text/plain, Size: 2411 bytes --]

On 9/25/24 1:47 PM, Mart Raudsepp wrote:
> Hello
> 
> 
> The state of GNOME in Gentoo could use a bunch of work and we don't
> really have the developers currently to take care of it.
> My own grandiose plans to be back and keep things in good order faced
> sad reality (a bunch of personal things) and we could really use some
> help in at least carefully reviewing and merging a contributors lots of
> work, and hopefully more.
> While I don't have the time, energy and currently motivation to be
> hands-on in the git tree right now, I am online on IRC rather
> constantly (lets say while waiting on work rust code compilation ;) and
> happy to help on-board an interested developer or developers with any
> peculiarities (which hopefully would get documented in the process),
> questions, module interactions, whatever is needed, including during PR
> reviews, just please do keep in touch, so I can offload things from my
> brain for a more sustainable future.


Thanks for bringing this up. I do not use gnome myself, but I use
cinnamon which was based on it and uses much of its technology stack, so
color me interested.


> Given a huge stack of available pull requests from a contributor, the
> main need right now is someone with a Gentoo developer hat who can help
> review these and get good things merged, fixed what's needed, and keep
> an eye out on the bugs. Of course contributors are welcome too, but
> please keep an eye out for not duplicating work that's already waiting
> review.


And this much I can certainly do. :)


> To my knowledge, one big issue right now is a circular dep between
> glib->gobject-introspection->glib, which needs proper solving to move
> forward with things. Or at least look at reducing the glib requirement
> in gnome-shell some way to at least unleash that core stack for our
> users. There are ideas we can talk about.


This is... tricky. The obvious possibility is to try to build a
bootstrap copy of AAA inside the ebuild for BBB as a bootstrap thing, as
long as that is sufficient to build a proper copy of AAA using the
partially valid installation of BBB.

Chewi was also trying to see if portage can be convinced to stage the
same package twice with different USE flags in the same emerge process
in order to solve cyclical USE flag dependencies, but it's not a
guarantee...


-- 
Eli Schwartz


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 236 bytes --]

Re: [gentoo-dev] Help needed for maintaining GNOME

[Ionen Wolkens]

[-- Attachment #1: Type: text/plain, Size: 480 bytes --]

On Wed, Sep 25, 2024 at 02:05:10PM -0400, Eli Schwartz wrote:
> Chewi was also trying to see if portage can be convinced to stage the
> same package twice with different USE flags in the same emerge process
> in order to solve cyclical USE flag dependencies, but it's not a
> guarantee...

I see it more as giving us hope of being able to remove whatever
horrible hack we implement eventually, but I think the horrible
hack will have to exist in the interim.
-- 
ionen

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: [gentoo-dev] Help needed for maintaining GNOME

[James Le Cuirot]

[-- Attachment #1: Type: text/plain, Size: 1102 bytes --]

On Wed, 2024-09-25 at 14:41 -0400, Ionen Wolkens wrote:
> On Wed, Sep 25, 2024 at 02:05:10PM -0400, Eli Schwartz wrote:
> > Chewi was also trying to see if portage can be convinced to stage the
> > same package twice with different USE flags in the same emerge process
> > in order to solve cyclical USE flag dependencies, but it's not a
> > guarantee...
> 
> I see it more as giving us hope of being able to remove whatever
> horrible hack we implement eventually, but I think the horrible
> hack will have to exist in the interim.

Yes, the glib/gobject-introspection conflict was my main test case. It was an
interesting one because it also involved some blockers. I pushed my half baked
idea up to GitHub in the hope that Zac or someone could maybe take it and
actually make it work properly. The results so far have shown that it at least
seems feasible. Once complete, we could get it out the door quite quickly.

As for GNOME, I'm afraid I don't use it, but I do recognise the importance
these packages have even on my KDE system, so I am very grateful for Leio's
hard work.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 858 bytes --]

Re: [gentoo-dev] Help needed for maintaining GNOME

[James Le Cuirot]

[-- Attachment #1: Type: text/plain, Size: 1289 bytes --]

On Wed, 2024-09-25 at 23:46 +0100, James Le Cuirot wrote:
> On Wed, 2024-09-25 at 14:41 -0400, Ionen Wolkens wrote:
> > On Wed, Sep 25, 2024 at 02:05:10PM -0400, Eli Schwartz wrote:
> > > Chewi was also trying to see if portage can be convinced to stage the
> > > same package twice with different USE flags in the same emerge process
> > > in order to solve cyclical USE flag dependencies, but it's not a
> > > guarantee...
> > 
> > I see it more as giving us hope of being able to remove whatever
> > horrible hack we implement eventually, but I think the horrible
> > hack will have to exist in the interim.
> 
> Yes, the glib/gobject-introspection conflict was my main test case. It was an
> interesting one because it also involved some blockers. I pushed my half baked
> idea up to GitHub in the hope that Zac or someone could maybe take it and
> actually make it work properly. The results so far have shown that it at least
> seems feasible. Once complete, we could get it out the door quite quickly.
> 
> As for GNOME, I'm afraid I don't use it, but I do recognise the importance
> these packages have even on my KDE system, so I am very grateful for Leio's
> hard work.

Sorry, meant to include a link to that.
https://github.com/gentoo/portage/pull/1385

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 858 bytes --]

Re: [gentoo-dev] Help needed for maintaining GNOME

[Zac Medico]

On Wed, Sep 25, 2024 at 3:46 PM James Le Cuirot <chewi@gentoo.org> wrote:
> Yes, the glib/gobject-introspection conflict was my main test case. It was an
> interesting one because it also involved some blockers. I pushed my half baked
> idea up to GitHub in the hope that Zac or someone could maybe take it and
> actually make it work properly. The results so far have shown that it at least
> seems feasible. Once complete, we could get it out the door quite quickly.

I took a peek at https://github.com/gentoo/portage/pull/1385, and it
looks like you may need the --autounmask-bactrack=y option to, since
any automunmask USE changes currently put a halt backtracking.
-- 
Thanks,
Zac

Re: [gentoo-dev] Help needed for maintaining GNOME

[Sam James]

Eli Schwartz <eschwartz@gentoo.org> writes:

> On 9/25/24 1:47 PM, Mart Raudsepp wrote:
>> Hello
>> 
>>
> [...]
>> To my knowledge, one big issue right now is a circular dep between
>> glib->gobject-introspection->glib, which needs proper solving to move
>> forward with things. Or at least look at reducing the glib requirement
>> in gnome-shell some way to at least unleash that core stack for our
>> users. There are ideas we can talk about.
>
>
> This is... tricky. The obvious possibility is to try to build a
> bootstrap copy of AAA inside the ebuild for BBB as a bootstrap thing, as
> long as that is sufficient to build a proper copy of AAA using the
> partially valid installation of BBB.

This is bug 937616 for those following along. Blake has been poking at
this suggestion a bit over there and in
https://github.com/gentoo/gentoo/pull/38618.

>
> Chewi was also trying to see if portage can be convinced to stage the
> same package twice with different USE flags in the same emerge process
> in order to solve cyclical USE flag dependencies, but it's not a
> guarantee...

Right, this is very much worth exploring but isn't an answer to this
given it'll take a while for the feature to be mature anyway.

Re: [gentoo-dev] Help needed for maintaining GNOME

[Yuan Liao (Leo3418)]

[-- Attachment #1: Type: text/plain, Size: 3323 bytes --]

On Wed, Sep 25, 2024 at 08:47:30PM +0300, Mart Raudsepp wrote:
> Hello
> 
> 
> The state of GNOME in Gentoo could use a bunch of work and we don't
> really have the developers currently to take care of it.
> My own grandiose plans to be back and keep things in good order faced
> sad reality (a bunch of personal things) and we could really use some
> help in at least carefully reviewing and merging a contributors lots of
> work, and hopefully more.
> While I don't have the time, energy and currently motivation to be
> hands-on in the git tree right now, I am online on IRC rather
> constantly (lets say while waiting on work rust code compilation ;) and
> happy to help on-board an interested developer or developers with any
> peculiarities (which hopefully would get documented in the process),
> questions, module interactions, whatever is needed, including during PR
> reviews, just please do keep in touch, so I can offload things from my
> brain for a more sustainable future.
> 
> Given a huge stack of available pull requests from a contributor, the
> main need right now is someone with a Gentoo developer hat who can help
> review these and get good things merged, fixed what's needed, and keep
> an eye out on the bugs. Of course contributors are welcome too, but
> please keep an eye out for not duplicating work that's already waiting
> review.

As a Gentoo user on GNOME, let me first use this opportunity to thank
leio, Matt (who previously maintained GNOME), and flint2 on GitHub (who
has been submitting many PRs for GNOME) for your work on this desktop
environment.  GNOME was something I would like to work on if we didn't
have flint2's many pending PRs because, of course, we wouldn't like
duplicate work.  At this moment, I don't have too much time, but as a
stakeholder of this GNOME issue on Gentoo, I'm happy to do my best to
help with it.

I just can't wrap my head around a deadlock I'm seeing.  We surely want
and need a Gentoo developer's involvement in GNOME, at least in
reviewing PRs; but if no Gentoo developer could review flint2's PRs,
then flint2 couldn't even become a Gentoo developer merely by showing a
stable stream of _merged_ contributions to GNOME on Gentoo -- because
those PRs wouldn't be reviewed and merged.  I was quite close to
becoming a Gentoo developer but had to pause the recruiting process due
to a personal life change; after I sorted that change out, I am no
longer interested in the area of Gentoo I contributed to previously
(namely Java, because (1) a lot of things about Java have been changed
and I have been struggling to keep up, and (2) I don't even use Java a
lot personally).  I would have definitely loved to help with this if I
were already a Gentoo developer.

The key point I wanted to make here is that we probably should get
flint2 on board as a Gentoo developer.

> To my knowledge, one big issue right now is a circular dep between
> glib->gobject-introspection->glib, which needs proper solving to move
> forward with things. Or at least look at reducing the glib requirement
> in gnome-shell some way to at least unleash that core stack for our
> users. There are ideas we can talk about.

Is this covered by one of flint2's PRs already?

> 
> Mart
> 

Best regards,
Leo3418

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 963 bytes --]