[gentoo-dev] add global useflag: webkit

[gentoo-dev] add global useflag: webkit

[hasufell]

# grep :webkit use.local.desc | wc -l
33

I would vote to make this a global useflag:

webkit - Adds support for the webkit library/module

Re: [gentoo-dev] add global useflag: webkit

[Ben]

On 6 May 2012 08:34, hasufell <hasufell@gentoo.org> wrote:
> # grep :webkit use.local.desc | wc -l
> 33
>
> I would vote to make this a global useflag:
>
> webkit - Adds support for the webkit library/module
>

IIRC this was already voted on ~2 years ago (before I retired). It's
just waiting for someone to implement it.
See also bug #285743.

Cheers,
Ben | yngwin

Re: [gentoo-dev] add global useflag: webkit

[Markos Chandras]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

On 05/06/2012 02:55 AM, Ben wrote:
> On 6 May 2012 08:34, hasufell <hasufell@gentoo.org> wrote:
>> # grep :webkit use.local.desc | wc -l 33
>> 
>> I would vote to make this a global useflag:
>> 
>> webkit - Adds support for the webkit library/module
>> 
> 
> IIRC this was already voted on ~2 years ago (before I retired).
> It's just waiting for someone to implement it. See also bug
> #285743.
> 
> Cheers, Ben | yngwin
> 
Any logs about that? I don't remember but to be honest there is
nothing to discuss here. 33 packages are more than enough to justify a
new global use flag. The description also looks good to me

- -- 
Regards,
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
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=RJdP
-----END PGP SIGNATURE-----

Re: [gentoo-dev] add global useflag: webkit

[Davide Pesavento]

On Sun, May 6, 2012 at 2:34 AM, hasufell <hasufell@gentoo.org> wrote:
> # grep :webkit use.local.desc | wc -l
> 33
>
> I would vote to make this a global useflag:
>
> webkit - Adds support for the webkit library/module
>

I suggest the following description:

"Add support for the WebKit HTML rendering/layout engine."

Otherwise +1 from me.

Thanks,
Davide

Re: [gentoo-dev] add global useflag: webkit

[hasufell]

On 05/06/2012 12:52 PM, Davide Pesavento wrote:
> On Sun, May 6, 2012 at 2:34 AM, hasufell <hasufell@gentoo.org> wrote:
>> # grep :webkit use.local.desc | wc -l
>> 33
>>
>> I would vote to make this a global useflag:
>>
>> webkit - Adds support for the webkit library/module
>>
> 
> I suggest the following description:
> 
> "Add support for the WebKit HTML rendering/layout engine."
> 
> Otherwise +1 from me.
> 
> Thanks,
> Davide
> 

I took the liberty to apply the change.
https://bugs.gentoo.org/show_bug.cgi?id=285743

I have not changed all metadata.xml files, because quite a few of them
have detailed descriptions on what the flag really does.
I don't see a problem with that, so the maintainers should decide.

There is still "net-irc/kvirc" which uses "qt-webkit". I would suggest
to change that to "webkit" like in "net-irc/quassel" (with pretty much
the same description).
Any reservations against that?

Re: [gentoo-dev] add global useflag: webkit

[Arfrever Frehtes Taifersar Arahesis]

[-- Attachment #1: Type: Text/Plain, Size: 324 bytes --]

2012-05-06 02:34:26 hasufell napisał(a):
> # grep :webkit use.local.desc | wc -l
> 33
> 
> I would vote to make this a global useflag:
> 
> webkit - Adds support for the webkit library/module

I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk USE flags.

-- 
Arfrever Frehtes Taifersar Arahesis

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[hasufell]

On 05/07/2012 02:47 AM, Arfrever Frehtes Taifersar Arahesis wrote:
> 2012-05-06 02:34:26 hasufell napisał(a):
>> # grep :webkit use.local.desc | wc -l 33
>> 
>> I would vote to make this a global useflag:
>> 
>> webkit - Adds support for the webkit library/module
> 
> I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk
> USE flags.
> 

You mean that for example KDE users who set +webkit in make.conf would
possibly get some weird gtk-deps too?

Re: [gentoo-dev] add global useflag: webkit

[Arfrever Frehtes Taifersar Arahesis]

[-- Attachment #1: Type: Text/Plain, Size: 1382 bytes --]

2012-05-07 03:00:31 hasufell napisał(a):
> On 05/07/2012 02:47 AM, Arfrever Frehtes Taifersar Arahesis wrote:
> > 2012-05-06 02:34:26 hasufell napisał(a):
> >> # grep :webkit use.local.desc | wc -l 33
> >> 
> >> I would vote to make this a global useflag:
> >> 
> >> webkit - Adds support for the webkit library/module
> > 
> > I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk
> > USE flags.
> > 
> 
> You mean that for example KDE users who set +webkit in make.conf would
> possibly get some weird gtk-deps too?

16 packages have "webkit" USE flag, which enables dependency on net-libs/webkit-gtk:
app-misc/gramps
app-office/gnucash
app-pda/gtkpod
app-text/xiphos
dev-java/swt
dev-util/geany-plugins
dev-util/mono-tools
gnome-extra/avant-window-navigator-extras
gnome-extra/zenity
mail-client/balsa
media-gfx/gimp
media-sound/gmusicbrowser
media-sound/gpodder
media-sound/rhythmbox
net-im/empathy
x11-misc/google-gadgets

15 packages have "webkit" USE flag, which enables dependency on x11-libs/qt-webkit:
dev-python/PyQt4
dev-python/pyside
kde-base/kget
kde-base/perlqt
kde-base/qtruby
kde-base/qyoto
kde-base/smokeqt
net-im/psi
net-im/qutim
net-irc/quassel
sci-chemistry/ball
sci-geosciences/merkaartor
x11-libs/qt-assistant
x11-libs/qt-declarative
x11-libs/qt-demo

-- 
Arfrever Frehtes Taifersar Arahesis

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Samuli Suominen]

On 05/07/2012 04:00 AM, hasufell wrote:
> On 05/07/2012 02:47 AM, Arfrever Frehtes Taifersar Arahesis wrote:
>> 2012-05-06 02:34:26 hasufell napisał(a):
>>> # grep :webkit use.local.desc | wc -l 33
>>>
>>> I would vote to make this a global useflag:
>>>
>>> webkit - Adds support for the webkit library/module
>>
>> I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk
>> USE flags.
>>
>
> You mean that for example KDE users who set +webkit in make.conf would
> possibly get some weird gtk-deps too?
>

KDE shouldn't be trying to avoid GTK+ in any case, as USE="gtk" is 
enabled in the default desktop profile and is a graphical toolkit which 
may be the only graphical toolkit available for an package (and likely 
is).  It's not like it's "straightup" GNOME.

-1 for separating them

Re: [gentoo-dev] add global useflag: webkit

[Ben]

On 7 May 2012 08:47, Arfrever Frehtes Taifersar Arahesis
<arfrever.fta@gmail.com> wrote:
> I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk USE flags.

I don't think that is necessary.

Ben | yngwin

Re: [gentoo-dev] add global useflag: webkit

[Jesús J. Guerrero Botella]

[-- Attachment #1: Type: text/plain, Size: 529 bytes --]

So we have 31 packages that can optionally support webkit, and none of them
lets you choose between the qt and gtk branches at compile time. I fail to
see the benefit of splitting the flag.

---
Jesús Guerrero Botella
El 07/05/2012 05:16, "Ben" <yngwin@gmail.com> escribió:

> On 7 May 2012 08:47, Arfrever Frehtes Taifersar Arahesis
> <arfrever.fta@gmail.com> wrote:
> > I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk USE
> flags.
>
> I don't think that is necessary.
>
> Ben | yngwin
>
>

[-- Attachment #2: Type: text/html, Size: 837 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

On 05/06/2012 05:47 PM, Arfrever Frehtes Taifersar Arahesis wrote:
> 2012-05-06 02:34:26 hasufell napisał(a):
>> # grep :webkit use.local.desc | wc -l
>> 33
>>
>> I would vote to make this a global useflag:
>>
>> webkit - Adds support for the webkit library/module
> 
> I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk USE flags.

Another possible way to model this kind of relationship would be to us
REQUIRED_USE to enforce relationships with the qt and gtk flags:

REQUIRED_USE="webkit? ( qt ) !webkit? ( !qt ) qt? ( webkit ) !qt? (
!webkit )"

versus

REQUIRED_USE="webkit? ( gtk ) !webkit? ( !gtk ) gtk? ( webkit ) !gtk? (
!webkit )"

It's pretty awkward with the existing operators, but we could extend the
REQUIRED_USE syntax to support an equivalent operator in a future EAPI.
-- 
Thanks,
Zac

Re: [gentoo-dev] add global useflag: webkit

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 755 bytes --]

On Mon, 07 May 2012 11:11:04 -0700
Zac Medico <zmedico@gentoo.org> wrote:
> REQUIRED_USE="webkit? ( qt ) !webkit? ( !qt ) qt? ( webkit ) !qt? (
> !webkit )"

Why do you need to write it both ways?

> It's pretty awkward with the existing operators, but we could extend
> the REQUIRED_USE syntax to support an equivalent operator in a future
> EAPI.

If we're doing this, can we get it in EAPI 5 please, and not use
workarounds in the tree until EAPI 5 is done? Getting the package
mangler to come up with good error messages for REQUIRED_USE failures
is a huge pain, and it gets worse the more clever tricks people come up
with to get around its inexpressivity.

I propose:

    REQUIRED_USE="== ( qt webkit )"

-- 
Ciaran McCreesh

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Ulrich Mueller]

>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:

> I propose:

>     REQUIRED_USE="== ( qt webkit )"

But this just means that the ebuild has redundant USE flags, so one of
them shouldn't be in IUSE, in the first place.

Ulrich

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
> 
>> I propose:
> 
>>     REQUIRED_USE="== ( qt webkit )"
> 
> But this just means that the ebuild has redundant USE flags, so one of
> them shouldn't be in IUSE, in the first place.

It serves to convey meaning, such that a user who has disabled the qt
USE flag will get a meaningful prompt if that flag is required for
webkit support. This kind of information could be useful to some people,
and it may be preferable to having a separate webkit-qt flag.
-- 
Thanks,
Zac

Re: [gentoo-dev] add global useflag: webkit

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 782 bytes --]

On Mon, 07 May 2012 13:24:31 -0700
Zac Medico <zmedico@gentoo.org> wrote:

> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
> >>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
> > 
> >> I propose:
> > 
> >>     REQUIRED_USE="== ( qt webkit )"
> > 
> > But this just means that the ebuild has redundant USE flags, so one
> > of them shouldn't be in IUSE, in the first place.
> 
> It serves to convey meaning, such that a user who has disabled the qt
> USE flag will get a meaningful prompt if that flag is required for
> webkit support. This kind of information could be useful to some
> people, and it may be preferable to having a separate webkit-qt flag.

If 'qt' flag is required for webkit support, it's 'webkit? ( qt )'.

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 316 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

On 05/07/2012 01:43 PM, Michał Górny wrote:
> On Mon, 07 May 2012 13:24:31 -0700
> Zac Medico <zmedico@gentoo.org> wrote:
> 
>> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
>>>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
>>>
>>>> I propose:
>>>
>>>>     REQUIRED_USE="== ( qt webkit )"
>>>
>>> But this just means that the ebuild has redundant USE flags, so one
>>> of them shouldn't be in IUSE, in the first place.
>>
>> It serves to convey meaning, such that a user who has disabled the qt
>> USE flag will get a meaningful prompt if that flag is required for
>> webkit support. This kind of information could be useful to some
>> people, and it may be preferable to having a separate webkit-qt flag.
> 
> If 'qt' flag is required for webkit support, it's 'webkit? ( qt )'.

What if '!webkit? ( !qt )' also applies though? As an alternative to
listing both constraints separately, you could combine them as '^^ (
webkit !qt )', or add support for '== ( qt webkit )' to make the
expression easier to read.
-- 
Thanks,
Zac

Re: [gentoo-dev] add global useflag: webkit

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 441 bytes --]

On Mon, 07 May 2012 14:41:33 -0700
Zac Medico <zmedico@gentoo.org> wrote:
> What if '!webkit? ( !qt )' also applies though? As an alternative to
> listing both constraints separately, you could combine them as '^^ (
> webkit !qt )', or add support for '== ( qt webkit )' to make the
> expression easier to read.

Forget "easier to read". The important part is being able to produce
error messages for users.

-- 
Ciaran McCreesh

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1293 bytes --]

On Mon, 07 May 2012 14:41:33 -0700
Zac Medico <zmedico@gentoo.org> wrote:

> On 05/07/2012 01:43 PM, Michał Górny wrote:
> > On Mon, 07 May 2012 13:24:31 -0700
> > Zac Medico <zmedico@gentoo.org> wrote:
> > 
> >> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
> >>>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
> >>>
> >>>> I propose:
> >>>
> >>>>     REQUIRED_USE="== ( qt webkit )"
> >>>
> >>> But this just means that the ebuild has redundant USE flags, so
> >>> one of them shouldn't be in IUSE, in the first place.
> >>
> >> It serves to convey meaning, such that a user who has disabled the
> >> qt USE flag will get a meaningful prompt if that flag is required
> >> for webkit support. This kind of information could be useful to
> >> some people, and it may be preferable to having a separate
> >> webkit-qt flag.
> > 
> > If 'qt' flag is required for webkit support, it's 'webkit? ( qt )'.
> 
> What if '!webkit? ( !qt )' also applies though? As an alternative to
> listing both constraints separately, you could combine them as '^^ (
> webkit !qt )', or add support for '== ( qt webkit )' to make the
> expression easier to read.

Then it's pointless to have the 'webkit' flag which doesn't control
anything.

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 316 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 05/07/2012 08:50 PM, Michał Górny wrote:
> On Mon, 07 May 2012 14:41:33 -0700 Zac Medico <zmedico@gentoo.org>
> wrote:
> 
>> On 05/07/2012 01:43 PM, Michał Górny wrote:
>>> On Mon, 07 May 2012 13:24:31 -0700 Zac Medico
>>> <zmedico@gentoo.org> wrote:
>>> 
>>>> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
>>>>>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
>>>>> 
>>>>>> I propose:
>>>>> 
>>>>>> REQUIRED_USE="== ( qt webkit )"
>>>>> 
>>>>> But this just means that the ebuild has redundant USE
>>>>> flags, so one of them shouldn't be in IUSE, in the first
>>>>> place.
>>>> 
>>>> It serves to convey meaning, such that a user who has
>>>> disabled the qt USE flag will get a meaningful prompt if that
>>>> flag is required for webkit support. This kind of information
>>>> could be useful to some people, and it may be preferable to
>>>> having a separate webkit-qt flag.
>>> 
>>> If 'qt' flag is required for webkit support, it's 'webkit? ( qt
>>> )'.
>> 
>> What if '!webkit? ( !qt )' also applies though? As an alternative
>> to listing both constraints separately, you could combine them as
>> '^^ ( webkit !qt )', or add support for '== ( qt webkit )' to
>> make the expression easier to read.
> 
> Then it's pointless to have the 'webkit' flag which doesn't
> control anything.

Generalize the discussion to be about two abstract flags "x" and "y"
that have the same kind of relationship, where each one actually does
control something, but the two features are intertwined in a
particular package such that they must both be enabled or disabled in
unison.
- -- 
Thanks,
Zac
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk+omdkACgkQ/ejvha5XGaO4CQCdGwcuuk4usnDj25nrcmd7D697
/TgAn3vXcPzEX3jCLhBVPPbnnX+lLWDW
=G/eD
-----END PGP SIGNATURE-----

Re: [gentoo-dev] add global useflag: webkit

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1914 bytes --]

On Mon, 07 May 2012 20:58:18 -0700
Zac Medico <zmedico@gentoo.org> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 05/07/2012 08:50 PM, Michał Górny wrote:
> > On Mon, 07 May 2012 14:41:33 -0700 Zac Medico <zmedico@gentoo.org>
> > wrote:
> > 
> >> On 05/07/2012 01:43 PM, Michał Górny wrote:
> >>> On Mon, 07 May 2012 13:24:31 -0700 Zac Medico
> >>> <zmedico@gentoo.org> wrote:
> >>> 
> >>>> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
> >>>>>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
> >>>>> 
> >>>>>> I propose:
> >>>>> 
> >>>>>> REQUIRED_USE="== ( qt webkit )"
> >>>>> 
> >>>>> But this just means that the ebuild has redundant USE
> >>>>> flags, so one of them shouldn't be in IUSE, in the first
> >>>>> place.
> >>>> 
> >>>> It serves to convey meaning, such that a user who has
> >>>> disabled the qt USE flag will get a meaningful prompt if that
> >>>> flag is required for webkit support. This kind of information
> >>>> could be useful to some people, and it may be preferable to
> >>>> having a separate webkit-qt flag.
> >>> 
> >>> If 'qt' flag is required for webkit support, it's 'webkit? ( qt
> >>> )'.
> >> 
> >> What if '!webkit? ( !qt )' also applies though? As an alternative
> >> to listing both constraints separately, you could combine them as
> >> '^^ ( webkit !qt )', or add support for '== ( qt webkit )' to
> >> make the expression easier to read.
> > 
> > Then it's pointless to have the 'webkit' flag which doesn't
> > control anything.
> 
> Generalize the discussion to be about two abstract flags "x" and "y"
> that have the same kind of relationship, where each one actually does
> control something, but the two features are intertwined in a
> particular package such that they must both be enabled or disabled in
> unison.

Then please show me an example of that.

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 316 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

On 05/07/2012 09:07 PM, Michał Górny wrote:
> On Mon, 07 May 2012 20:58:18 -0700
> Zac Medico <zmedico@gentoo.org> wrote:
> 
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> On 05/07/2012 08:50 PM, Michał Górny wrote:
>>> On Mon, 07 May 2012 14:41:33 -0700 Zac Medico <zmedico@gentoo.org>
>>> wrote:
>>>
>>>> On 05/07/2012 01:43 PM, Michał Górny wrote:
>>>>> On Mon, 07 May 2012 13:24:31 -0700 Zac Medico
>>>>> <zmedico@gentoo.org> wrote:
>>>>>
>>>>>> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
>>>>>>>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
>>>>>>>
>>>>>>>> I propose:
>>>>>>>
>>>>>>>> REQUIRED_USE="== ( qt webkit )"
>>>>>>>
>>>>>>> But this just means that the ebuild has redundant USE
>>>>>>> flags, so one of them shouldn't be in IUSE, in the first
>>>>>>> place.
>>>>>>
>>>>>> It serves to convey meaning, such that a user who has
>>>>>> disabled the qt USE flag will get a meaningful prompt if that
>>>>>> flag is required for webkit support. This kind of information
>>>>>> could be useful to some people, and it may be preferable to
>>>>>> having a separate webkit-qt flag.
>>>>>
>>>>> If 'qt' flag is required for webkit support, it's 'webkit? ( qt
>>>>> )'.
>>>>
>>>> What if '!webkit? ( !qt )' also applies though? As an alternative
>>>> to listing both constraints separately, you could combine them as
>>>> '^^ ( webkit !qt )', or add support for '== ( qt webkit )' to
>>>> make the expression easier to read.
>>>
>>> Then it's pointless to have the 'webkit' flag which doesn't
>>> control anything.
>>
>> Generalize the discussion to be about two abstract flags "x" and "y"
>> that have the same kind of relationship, where each one actually does
>> control something, but the two features are intertwined in a
>> particular package such that they must both be enabled or disabled in
>> unison.
> 
> Then please show me an example of that.

I don't see any offhand. I guess it's fairly uncommon, or non-existent.
-- 
Thanks,
Zac

Re: [gentoo-dev] add global useflag: webkit

[Samuli Suominen]

On 05/07/2012 11:24 PM, Zac Medico wrote:
> On 05/07/2012 12:18 PM, Ulrich Mueller wrote:
>>>>>>> On Mon, 7 May 2012, Ciaran McCreesh wrote:
>>
>>> I propose:
>>
>>>      REQUIRED_USE="== ( qt webkit )"
>>
>> But this just means that the ebuild has redundant USE flags, so one of
>> them shouldn't be in IUSE, in the first place.
>
> It serves to convey meaning, such that a user who has disabled the qt
> USE flag will get a meaningful prompt if that flag is required for
> webkit support. This kind of information could be useful to some people,
> and it may be preferable to having a separate webkit-qt flag.

ulm is right, it would still show up as an redudant USE flag as,
the preferable result is that even KDE/Qt4 users get webkit-gtk 
installed if it's the rendering engine provided by the package.
GTK+ is just graphical toolkit and not directly GNOME (like KDE users 
avoiding GNOME). By this same logic both USE="gtk qt4" are enabled in 
the desktop profile.
I don't think we have any packages with *both* webkit-gtk and webkit-qt 
supports in tree (do we?).

Re: [gentoo-dev] add global useflag: webkit

[Ulrich Mueller]

>>>>> On Mon, 07 May 2012, Zac Medico wrote:

> Another possible way to model this kind of relationship would be to
> us REQUIRED_USE to enforce relationships with the qt and gtk flags:

> REQUIRED_USE="webkit? ( qt ) !webkit? ( !qt ) qt? ( webkit ) !qt? (
> !webkit )"

This line just says that either both webkit and qt must be set, or
neither of them. In other words, that the ebuild has redundant flags,
which IMHO should be avoided.

> versus

> REQUIRED_USE="webkit? ( gtk ) !webkit? ( !gtk ) gtk? ( webkit ) !gtk? (
> !webkit )"

> It's pretty awkward with the existing operators, but we could extend
> the REQUIRED_USE syntax to support an equivalent operator in a
> future EAPI.

As far as I can see, it is equivalent to:

REQUIRED_USE="^^ ( webkit !qt )"

(or analog for the gtk case). But see above.

Ulrich

Re: [gentoo-dev] add global useflag: webkit

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 247 bytes --]

On Mon, 7 May 2012 20:26:08 +0200
Ulrich Mueller <ulm@gentoo.org> wrote:
> REQUIRED_USE="^^ ( webkit !qt )"

Please provide an algorithm that will turn that into an appropriate
error message for displaying to a user.

-- 
Ciaran McCreesh

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

On 05/07/2012 11:26 AM, Ulrich Mueller wrote:
>> REQUIRED_USE="webkit? ( gtk ) !webkit? ( !gtk ) gtk? ( webkit ) !gtk? (
>> !webkit )"
> 
>> It's pretty awkward with the existing operators, but we could extend
>> the REQUIRED_USE syntax to support an equivalent operator in a
>> future EAPI.
> 
> As far as I can see, it is equivalent to:
> 
> REQUIRED_USE="^^ ( webkit !qt )"
> 
> (or analog for the gtk case). But see above.

Ah, that's right. This would another good example for the dev manual,
like the one from bug 399069 [1].

[1] https://bugs.gentoo.org/show_bug.cgi?id=399069
-- 
Thanks,
Zac

Re: [gentoo-dev] add global useflag: webkit

[Stelian Ionescu]

[-- Attachment #1: Type: text/plain, Size: 2073 bytes --]

On Mon, 2012-05-07 at 11:11 -0700, Zac Medico wrote:
> On 05/06/2012 05:47 PM, Arfrever Frehtes Taifersar Arahesis wrote:
> > 2012-05-06 02:34:26 hasufell napisał(a):
> >> # grep :webkit use.local.desc | wc -l
> >> 33
> >>
> >> I would vote to make this a global useflag:
> >>
> >> webkit - Adds support for the webkit library/module
> > 
> > I suggest to use separate qt-webkit (or webkit-qt) and webkit-gtk USE flags.
> 
> Another possible way to model this kind of relationship would be to us
> REQUIRED_USE to enforce relationships with the qt and gtk flags:
> 
> REQUIRED_USE="webkit? ( qt ) !webkit? ( !qt ) qt? ( webkit ) !qt? (
> !webkit )"
> 
> versus
> 
> REQUIRED_USE="webkit? ( gtk ) !webkit? ( !gtk ) gtk? ( webkit ) !gtk? (
> !webkit )"
> 
> It's pretty awkward with the existing operators, but we could extend the
> REQUIRED_USE syntax to support an equivalent operator in a future EAPI.

Isn't it the time to make a new EAPI which no longer has USE "flags" but
USE "values" ? Many of the really weird USE flags combinations would be
much more clearly expressed if the possible types for a USE variable
were:
1) member-of: for choosing the backend of certain functionality; e.g.
all combinations of USE="ssl openssl gnutls nss" would become e.g.
in the ebuild:
USE="ssl=[member-of: openssl,gnutls,nss]"
in make-conf:
USE: ssl=openssl, as synonymous of old USE="ssl openssl", or
USE: ssl=none, as synonymous of old USE="-ssl"
2) subset-of: for selecting a number of modules to build. This would
replace USE_EXPAND quite neatly
3) boolean: as alias for member-of: [true,false]
4) unsigned int: IIRC some (few) packages can take optional uint at
configure-time. for example with dev-lisp/sbcl one can customize some
hardcoded GC parameters suchs as the default heap size, generation size,
etc...

In the above case, one could have a USE variable named "webkit" of type
member-of: [qt,gtk]

-- 
Stelian Ionescu a.k.a. fe[nl]ix
Quidquid latine dictum sit, altum videtur.
http://common-lisp.net/project/iolib


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [gentoo-dev] add global useflag: webkit

[Zac Medico]

On 05/07/2012 12:33 PM, Stelian Ionescu wrote:
> Isn't it the time to make a new EAPI which no longer has USE "flags" but
> USE "values" ? Many of the really weird USE flags combinations would be
> much more clearly expressed if the possible types for a USE variable
> were:
> 1) member-of: for choosing the backend of certain functionality; e.g.
> all combinations of USE="ssl openssl gnutls nss" would become e.g.
> in the ebuild:
> USE="ssl=[member-of: openssl,gnutls,nss]"
> in make-conf:
> USE: ssl=openssl, as synonymous of old USE="ssl openssl", or
> USE: ssl=none, as synonymous of old USE="-ssl"
> 2) subset-of: for selecting a number of modules to build. This would
> replace USE_EXPAND quite neatly
> 3) boolean: as alias for member-of: [true,false]
> 4) unsigned int: IIRC some (few) packages can take optional uint at
> configure-time. for example with dev-lisp/sbcl one can customize some
> hardcoded GC parameters suchs as the default heap size, generation size,
> etc...
> 
> In the above case, one could have a USE variable named "webkit" of type
> member-of: [qt,gtk]

We can get similar results using booleans with REQUIRED_USE, and your
approach seems to introduce needless complexity.
-- 
Thanks,
Zac