bug#27179: gtk+ in dependency graph for qt

bug#27179: gtk+ in dependency graph for qt

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 490 bytes --]

gtk+ is currently in the dependency tree for qt, but is not an input.
Either we should add gtk+ as an input so the styling can be uniform
between the two toolkits, or we should find which input pulls it in and
see what other options we have for removing it from the graph.

-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

bug#27179: gtk+ in dependency graph for qt

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 691 bytes --]

On Thu, Jun 01, 2017 at 09:14:40AM +0300, Efraim Flashner wrote:
> gtk+ is currently in the dependency tree for qt, but is not an input.
> Either we should add gtk+ as an input so the styling can be uniform
> between the two toolkits, or we should find which input pulls it in and
> see what other options we have for removing it from the graph.
> 

It appears that qtbase depends on libinput which depends on gtk+@3, and
also qtbase -> libinput -> libwacom -> gtk+@3.

-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

bug#27179: gtk+ in dependency graph for qt

[Ricardo Wurmus]

Efraim Flashner <efraim@flashner.co.il> writes:

> On Thu, Jun 01, 2017 at 09:14:40AM +0300, Efraim Flashner wrote:
>> gtk+ is currently in the dependency tree for qt, but is not an input.
>> Either we should add gtk+ as an input so the styling can be uniform
>> between the two toolkits, or we should find which input pulls it in and
>> see what other options we have for removing it from the graph.
>>
>
> It appears that qtbase depends on libinput which depends on gtk+@3, and
> also qtbase -> libinput -> libwacom -> gtk+@3.

Can we split libinput into a qt part and a separate gtk part?

--
Ricardo

GPG: BCA6 89B6 3655 3801 C3C6  2150 197A 5888 235F ACAC
https://elephly.net

bug#27179: gtk+ in dependency graph for qt

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 1150 bytes --]

On Sun, Jul 02, 2017 at 09:23:44PM +0200, Ricardo Wurmus wrote:
> 
> Efraim Flashner <efraim@flashner.co.il> writes:
> 
> > On Thu, Jun 01, 2017 at 09:14:40AM +0300, Efraim Flashner wrote:
> >> gtk+ is currently in the dependency tree for qt, but is not an input.
> >> Either we should add gtk+ as an input so the styling can be uniform
> >> between the two toolkits, or we should find which input pulls it in and
> >> see what other options we have for removing it from the graph.
> >>
> >
> > It appears that qtbase depends on libinput which depends on gtk+@3, and
> > also qtbase -> libinput -> libwacom -> gtk+@3.
> 
> Can we split libinput into a qt part and a separate gtk part?
> 

libwacom is built with glib-or-gtk-build-system, so if we switch
libinput to libinput-minimal then we can drop the GUI for libinput and
libwacom support and remove gtk+ as an input. Or we can try to build
libwacom without gtk.

-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

bug#27179: gtk+ in dependency graph for qt

[Efraim Flashner]

[-- Attachment #1: Type: text/plain, Size: 1434 bytes --]

On Sun, Jul 02, 2017 at 11:16:32PM +0300, Efraim Flashner wrote:
> On Sun, Jul 02, 2017 at 09:23:44PM +0200, Ricardo Wurmus wrote:
> > 
> > Efraim Flashner <efraim@flashner.co.il> writes:
> > 
> > > On Thu, Jun 01, 2017 at 09:14:40AM +0300, Efraim Flashner wrote:
> > >> gtk+ is currently in the dependency tree for qt, but is not an input.
> > >> Either we should add gtk+ as an input so the styling can be uniform
> > >> between the two toolkits, or we should find which input pulls it in and
> > >> see what other options we have for removing it from the graph.
> > >>
> > >
> > > It appears that qtbase depends on libinput which depends on gtk+@3, and
> > > also qtbase -> libinput -> libwacom -> gtk+@3.
> > 
> > Can we split libinput into a qt part and a separate gtk part?
> > 
> 
> libwacom is built with glib-or-gtk-build-system, so if we switch
> libinput to libinput-minimal then we can drop the GUI for libinput and
> libwacom support and remove gtk+ as an input. Or we can try to build
> libwacom without gtk.
> 

I just noticed that efl in elightenment.scm also has libinput as an
input. I think for now we should replace both qtbase's and efl's
libinput with libinput-minimal.


-- 
Efraim Flashner   <efraim@flashner.co.il>   אפרים פלשנר
GPG key = A28B F40C 3E55 1372 662D  14F7 41AA E7DC CA3D 8351
Confidentiality cannot be guaranteed on emails sent or received unencrypted

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

bug#27179: gtk+ in dependency graph for qt

[Ludovic Courtès]

Efraim Flashner <efraim@flashner.co.il> skribis:

> On Sun, Jul 02, 2017 at 11:16:32PM +0300, Efraim Flashner wrote:
>> On Sun, Jul 02, 2017 at 09:23:44PM +0200, Ricardo Wurmus wrote:
>> > 
>> > Efraim Flashner <efraim@flashner.co.il> writes:
>> > 
>> > > On Thu, Jun 01, 2017 at 09:14:40AM +0300, Efraim Flashner wrote:
>> > >> gtk+ is currently in the dependency tree for qt, but is not an input.
>> > >> Either we should add gtk+ as an input so the styling can be uniform
>> > >> between the two toolkits, or we should find which input pulls it in and
>> > >> see what other options we have for removing it from the graph.
>> > >>
>> > >
>> > > It appears that qtbase depends on libinput which depends on gtk+@3, and
>> > > also qtbase -> libinput -> libwacom -> gtk+@3.
>> > 
>> > Can we split libinput into a qt part and a separate gtk part?
>> > 
>> 
>> libwacom is built with glib-or-gtk-build-system, so if we switch
>> libinput to libinput-minimal then we can drop the GUI for libinput and
>> libwacom support and remove gtk+ as an input. Or we can try to build
>> libwacom without gtk.
>> 
>
> I just noticed that efl in elightenment.scm also has libinput as an
> input. I think for now we should replace both qtbase's and efl's
> libinput with libinput-minimal.

I guess you can reasonably start with EFL since it has few dependents.
I suppose we’ll also need a “libwacom-minimal” package no?

Or perhaps we should simply add a “lib” output since only the libwacom-*
commands depend on GTK+:

--8<---------------cut here---------------start------------->8---
$ grep -r 8f5bm9zqakgl7mc4lr5ck8l91vhlzv85 /gnu/store/4p4vw5shvj1ywjy91m5q3mpjyr1wp1hx-libwacom-0.25
/gnu/store/4p4vw5shvj1ywjy91m5q3mpjyr1wp1hx-libwacom-0.25/bin/libwacom-list-local-devices:export XDG_DATA_DIRS="/gnu/store/s0mc7alnzxpvvfay3pcpw8n3ipcq0a0r-shared-mime-info-1.8/share:/gnu/store/1k9zfb8wqd3pm97jsiy8n5szy56hgn8k-glib-2.52.2/share:/gnu/store/8f5bm9zqakgl7mc4lr5ck8l91vhlzv85-gtk+-3.22.15/share${XDG_DATA_DIRS:+:}$XDG_DATA_DIRS"
/gnu/store/4p4vw5shvj1ywjy91m5q3mpjyr1wp1hx-libwacom-0.25/bin/libwacom-list-local-devices:export GTK_PATH="/gnu/store/8f5bm9zqakgl7mc4lr5ck8l91vhlzv85-gtk+-3.22.15/lib/gtk-3.0${GTK_PATH:+:}$GTK_PATH"
$ ldd /gnu/store/4p4vw5shvj1ywjy91m5q3mpjyr1wp1hx-libwacom-0.25/bin/.libwacom-list-local-devices-real |grep
gtk
--8<---------------cut here---------------end--------------->8---

It’s not even clear that GTK+ is needed, given that it doesn’t show up
in ldd.

The same thing probably applies to libinput.

Thoughts?

Ludo’.

bug#27179: gtk+ in dependency graph for qt

[Maxim Cournoyer]

Hello,

Efraim Flashner <efraim@flashner.co.il> writes:

> gtk+ is currently in the dependency tree for qt, but is not an input.
> Either we should add gtk+ as an input so the styling can be uniform
> between the two toolkits, or we should find which input pulls it in and
> see what other options we have for removing it from the graph.

Apologies for not seeing this earlier, but I fixed it with commit
206193fec9cf527dec73841c16b6ae6ae52aaf37 last June.

Now, we should ensure gtk@4 doesn't depend on qtbase though... (see
#51994).

Closing,

Thanks!

Maxim