[gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1902 bytes --]

Hello,

I think we're mostly aware what the use and benefits
of the *use.stable.mask files are.

They would be at least really useful in Python ebuilds, where we
have to either:

a) forcedly stabilize a particular Python implementation (like pypy),

b) don't support it all,

c) or just keep two package revisions around, one with 'stable' Python
implementations for stabilization and the other with all
implementations for testing users.


Therefore, having *use.stable.mask would be at least helpful to us. But
as far as I can see, the spec says we can use it only in profile dirs
with EAPI 5...

So far, I doubt anyone would want us to migrate our major profiles
to a newer EAPI, like EAPI 4, not to mention fresh 5. And of course,
that wouldn't follow our migration path practices.


Therefore, I see the following solutions:

1) duplicate most of the major profiles. Make an EAPI 5-enabled wrapper
profiles which will provide the *use.stable.mask files. Require users
to migrate to those profiles after getting an EAPI 5 capable package
manager (how?). Possibly mask the relevant flags completely in other
profiles.


2) change the rules. Make *use.stable.mask files not require EAPI 5
profile dirs but apply only to EAPI 5 packages. The outcome is similar
-- package managers without the feature ignore the ebuilds. If they
have EAPI 5, they should be able to handle stable unmasking as well.

Of course, it all falls apart because of package manager strictness.
We may want to change that retroactively and quickly release updated
package managers before the EAPI 5 support is spread wider (assuming
some transitional period before we will start using the files), or defer
it into EAPI 6.


Either way, I believe that *use.stable.mask would be very helpful if we
were able to use it. What are your thoughts?

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Ian Stakenvicius]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 10/12/12 04:27 PM, Michał Górny wrote:
> Hello,
> 
> I think we're mostly aware what the use and benefits of the
> *use.stable.mask files are.
> 
> They would be at least really useful in Python ebuilds, where we 
> have to either:
> 
> a) forcedly stabilize a particular Python implementation (like
> pypy),
> 
> b) don't support it all,
> 
> c) or just keep two package revisions around, one with 'stable'
> Python implementations for stabilization and the other with all 
> implementations for testing users.
> 
> 
> Therefore, having *use.stable.mask would be at least helpful to us.
> But as far as I can see, the spec says we can use it only in
> profile dirs with EAPI 5...
> 
> So far, I doubt anyone would want us to migrate our major profiles 
> to a newer EAPI, like EAPI 4, not to mention fresh 5. And of
> course, that wouldn't follow our migration path practices.
> 
> 
> Therefore, I see the following solutions:
> 
> 1) duplicate most of the major profiles. Make an EAPI 5-enabled
> wrapper profiles which will provide the *use.stable.mask files.
> Require users to migrate to those profiles after getting an EAPI 5
> capable package manager (how?). Possibly mask the relevant flags
> completely in other profiles.
> 
> 
> 2) change the rules. Make *use.stable.mask files not require EAPI
> 5 profile dirs but apply only to EAPI 5 packages. The outcome is
> similar -- package managers without the feature ignore the ebuilds.
> If they have EAPI 5, they should be able to handle stable unmasking
> as well.
> 
> Of course, it all falls apart because of package manager
> strictness. We may want to change that retroactively and quickly
> release updated package managers before the EAPI 5 support is
> spread wider (assuming some transitional period before we will
> start using the files), or defer it into EAPI 6.
> 
> 
> Either way, I believe that *use.stable.mask would be very helpful
> if we were able to use it. What are your thoughts?
> 

I wonder how (2) would really differ from the current situation -- ie,
if there's a use.stable.mask file in a profiles dir, and portage is
too old to support it, doesn't it just get ignored?
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iF4EAREIAAYFAlDGk/4ACgkQ2ugaI38ACPBkogEAsqOBZBa1n63+dkd/mz7XzFzy
XHoshXhY5kOMTMKz7NgBAI9JODGAp9VGlAZg2w7lOoAFTmvgQyElWY0AA/9Sn6h7
=rHGA
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 2661 bytes --]

On Mon, 10 Dec 2012 21:01:34 -0500
Ian Stakenvicius <axs@gentoo.org> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
> 
> On 10/12/12 04:27 PM, Michał Górny wrote:
> > Hello,
> > 
> > I think we're mostly aware what the use and benefits of the
> > *use.stable.mask files are.
> > 
> > They would be at least really useful in Python ebuilds, where we 
> > have to either:
> > 
> > a) forcedly stabilize a particular Python implementation (like
> > pypy),
> > 
> > b) don't support it all,
> > 
> > c) or just keep two package revisions around, one with 'stable'
> > Python implementations for stabilization and the other with all 
> > implementations for testing users.
> > 
> > 
> > Therefore, having *use.stable.mask would be at least helpful to us.
> > But as far as I can see, the spec says we can use it only in
> > profile dirs with EAPI 5...
> > 
> > So far, I doubt anyone would want us to migrate our major profiles 
> > to a newer EAPI, like EAPI 4, not to mention fresh 5. And of
> > course, that wouldn't follow our migration path practices.
> > 
> > 
> > Therefore, I see the following solutions:
> > 
> > 1) duplicate most of the major profiles. Make an EAPI 5-enabled
> > wrapper profiles which will provide the *use.stable.mask files.
> > Require users to migrate to those profiles after getting an EAPI 5
> > capable package manager (how?). Possibly mask the relevant flags
> > completely in other profiles.
> > 
> > 
> > 2) change the rules. Make *use.stable.mask files not require EAPI
> > 5 profile dirs but apply only to EAPI 5 packages. The outcome is
> > similar -- package managers without the feature ignore the ebuilds.
> > If they have EAPI 5, they should be able to handle stable unmasking
> > as well.
> > 
> > Of course, it all falls apart because of package manager
> > strictness. We may want to change that retroactively and quickly
> > release updated package managers before the EAPI 5 support is
> > spread wider (assuming some transitional period before we will
> > start using the files), or defer it into EAPI 6.
> > 
> > 
> > Either way, I believe that *use.stable.mask would be very helpful
> > if we were able to use it. What are your thoughts?
> > 
> 
> I wonder how (2) would really differ from the current situation -- ie,
> if there's a use.stable.mask file in a profiles dir, and portage is
> too old to support it, doesn't it just get ignored?

Well, assuming the EAPI 5 support is applied at once, that portage
version will ignore EAPI 5 packages as well, making the file therefore
irrelevant.

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Zac Medico]

On 12/10/2012 01:27 PM, Michał Górny wrote:
> 1) duplicate most of the major profiles. Make an EAPI 5-enabled wrapper
> profiles which will provide the *use.stable.mask files. Require users
> to migrate to those profiles after getting an EAPI 5 capable package
> manager (how?). Possibly mask the relevant flags completely in other
> profiles.

I think this is the obvious solution. You can make users migrate by
adding "deprecated" files to the old profiles.
-- 
Thanks,
Zac

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1401 bytes --]

On Mon, 10 Dec 2012 22:35:07 -0800
Zac Medico <zmedico@gentoo.org> wrote:

> On 12/10/2012 01:27 PM, Michał Górny wrote:
> > 1) duplicate most of the major profiles. Make an EAPI 5-enabled wrapper
> > profiles which will provide the *use.stable.mask files. Require users
> > to migrate to those profiles after getting an EAPI 5 capable package
> > manager (how?). Possibly mask the relevant flags completely in other
> > profiles.
> 
> I think this is the obvious solution. You can make users migrate by
> adding "deprecated" files to the old profiles.

To be honest, I don't see much benefit from it compared to not having
the *stable.use.mask files at all and just adding separate stable
profiles.

AFAICS, that would have three advantages over the proposed solution:

1) the 'new' profiles wouldn't need to be EAPI=5 and therefore the
solution will work correctly even for quite an ancient package managers,

2) less users will have to switch profiles. Even if for safety we
wanted the unstable users to switch profiles,

3) package.accept_keywords will not magically switch masks. This one
probably is a matter of taste but if some arch testers actually use
package.accept_keywords to quickly test packages before stabilizing
them, EAPI=5 solution will automatically unmask the flags which won't
be present on a stable system.

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Zac Medico]

On 12/11/2012 01:45 PM, Michał Górny wrote:
> On Mon, 10 Dec 2012 22:35:07 -0800
> Zac Medico <zmedico@gentoo.org> wrote:
> 
>> On 12/10/2012 01:27 PM, Michał Górny wrote:
>>> 1) duplicate most of the major profiles. Make an EAPI 5-enabled wrapper
>>> profiles which will provide the *use.stable.mask files. Require users
>>> to migrate to those profiles after getting an EAPI 5 capable package
>>> manager (how?). Possibly mask the relevant flags completely in other
>>> profiles.
>>
>> I think this is the obvious solution. You can make users migrate by
>> adding "deprecated" files to the old profiles.
> 
> To be honest, I don't see much benefit from it compared to not having
> the *stable.use.mask files at all and just adding separate stable
> profiles.

The main use case for *use.stable.mask that I'm aware of is that it's
handy for masking flags to pass repoman checks. For example,
sys-apps/portage could use it for the pypy1_9 flag. Otherwise, we have
to mask that flag for a given portage version before we can add stable
keywords.
-- 
Thanks,
Zac

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1453 bytes --]

On Tue, 11 Dec 2012 16:44:25 -0800
Zac Medico <zmedico@gentoo.org> wrote:

> On 12/11/2012 01:45 PM, Michał Górny wrote:
> > On Mon, 10 Dec 2012 22:35:07 -0800
> > Zac Medico <zmedico@gentoo.org> wrote:
> > 
> >> On 12/10/2012 01:27 PM, Michał Górny wrote:
> >>> 1) duplicate most of the major profiles. Make an EAPI 5-enabled wrapper
> >>> profiles which will provide the *use.stable.mask files. Require users
> >>> to migrate to those profiles after getting an EAPI 5 capable package
> >>> manager (how?). Possibly mask the relevant flags completely in other
> >>> profiles.
> >>
> >> I think this is the obvious solution. You can make users migrate by
> >> adding "deprecated" files to the old profiles.
> > 
> > To be honest, I don't see much benefit from it compared to not having
> > the *stable.use.mask files at all and just adding separate stable
> > profiles.
> 
> The main use case for *use.stable.mask that I'm aware of is that it's
> handy for masking flags to pass repoman checks. For example,
> sys-apps/portage could use it for the pypy1_9 flag. Otherwise, we have
> to mask that flag for a given portage version before we can add stable
> keywords.

Yes, and having 'stable' and 'unstable' profiles will work just
the same. Except for the fact that it will be a bit cleaner, not require
EAPI=5 at all and probably make arch testing a bit easier for a few
people.

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Zac Medico]

On 12/12/2012 01:32 AM, Michał Górny wrote:
> On Tue, 11 Dec 2012 16:44:25 -0800
> Zac Medico <zmedico@gentoo.org> wrote:
> 
>> On 12/11/2012 01:45 PM, Michał Górny wrote:
>>> On Mon, 10 Dec 2012 22:35:07 -0800
>>> Zac Medico <zmedico@gentoo.org> wrote:
>>>
>>>> On 12/10/2012 01:27 PM, Michał Górny wrote:
>>>>> 1) duplicate most of the major profiles. Make an EAPI 5-enabled wrapper
>>>>> profiles which will provide the *use.stable.mask files. Require users
>>>>> to migrate to those profiles after getting an EAPI 5 capable package
>>>>> manager (how?). Possibly mask the relevant flags completely in other
>>>>> profiles.
>>>>
>>>> I think this is the obvious solution. You can make users migrate by
>>>> adding "deprecated" files to the old profiles.
>>>
>>> To be honest, I don't see much benefit from it compared to not having
>>> the *stable.use.mask files at all and just adding separate stable
>>> profiles.
>>
>> The main use case for *use.stable.mask that I'm aware of is that it's
>> handy for masking flags to pass repoman checks. For example,
>> sys-apps/portage could use it for the pypy1_9 flag. Otherwise, we have
>> to mask that flag for a given portage version before we can add stable
>> keywords.
> 
> Yes, and having 'stable' and 'unstable' profiles will work just
> the same. Except for the fact that it will be a bit cleaner, not require
> EAPI=5 at all and probably make arch testing a bit easier for a few
> people.

Sounds good to me.
-- 
Thanks,
Zac

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Andreas K. Huettel]

[-- Attachment #1: Type: Text/Plain, Size: 1496 bytes --]

Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
> > Yes, and having 'stable' and 'unstable' profiles will work just
> > the same. Except for the fact that it will be a bit cleaner, not require
> > EAPI=5 at all and probably make arch testing a bit easier for a few
> > people.
> 
> Sounds good to me.

Except that it completely breaks stabilization procedures, since packages are 
then not only tested with a larger range of useflags, but with an entirely 
different profile. Not such a great idea. 

The whole point of the stable masking was to keep the changes minimal when 
going from a "testing" to a "stable" state - by only restricting the use flag 
choices, and nothing else. This means most of the testing done with ~arch 
packages is still valid and provides meaningful feedback to maintainers and 
arch teams for stabilization.

In general, using a separate set of profiles, however, whill not help you 
enabling the stable mask files, since these will then only be allowed inside 
the new profiles. Not in the base profile or in the main profile directory, 
which still follows the old EAPI. In the sense of easy handling, noone will 
probably want to edit 
profiles/highly_unstable/next_version/package.stable.mask. 

<rant>I have basically given up that this "feature" will ever become useful 
for the main tree. Long live inertia.</rant>

-- 

Andreas K. Huettel
Gentoo Linux developer 
dilfridge@gentoo.org
http://www.akhuettel.de/


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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 2142 bytes --]

On Thu, 13 Dec 2012 21:33:50 +0100
"Andreas K. Huettel" <dilfridge@gentoo.org> wrote:

> Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
> > > Yes, and having 'stable' and 'unstable' profiles will work just
> > > the same. Except for the fact that it will be a bit cleaner, not require
> > > EAPI=5 at all and probably make arch testing a bit easier for a few
> > > people.
> > 
> > Sounds good to me.
> 
> Except that it completely breaks stabilization procedures, since packages are 
> then not only tested with a larger range of useflags, but with an entirely 
> different profile. Not such a great idea. 
> 
> The whole point of the stable masking was to keep the changes minimal when 
> going from a "testing" to a "stable" state - by only restricting the use flag 
> choices, and nothing else. This means most of the testing done with ~arch 
> packages is still valid and provides meaningful feedback to maintainers and 
> arch teams for stabilization.

Well, it's all a question of decisions, I believe. If we make sure that
the new 'unstable' profiles differ from the 'stable' ones only by
additional masked/unmasked USE flags, I don't think it'd be an issue.

> In general, using a separate set of profiles, however, whill not help you 
> enabling the stable mask files, since these will then only be allowed inside 
> the new profiles. Not in the base profile or in the main profile directory, 
> which still follows the old EAPI. In the sense of easy handling, noone will 
> probably want to edit 
> profiles/highly_unstable/next_version/package.stable.mask. 
> 
> <rant>I have basically given up that this "feature" will ever become useful 
> for the main tree. Long live inertia.</rant>

I'm thinking of making all the current profiles 'testing'. As in, we
mask the 'unstable' flags in base profile completely.

The new profiles will just add a common 'testing' profile which would
unmask those flags. ~arch users could still use the regular profiles
but would have to switch if they wanted the additional flags
automatically unmasked.

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Zac Medico]

On 12/13/2012 12:43 PM, Michał Górny wrote:
> On Thu, 13 Dec 2012 21:33:50 +0100
> "Andreas K. Huettel" <dilfridge@gentoo.org> wrote:
> 
>> Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
>>>> Yes, and having 'stable' and 'unstable' profiles will work just
>>>> the same. Except for the fact that it will be a bit cleaner, not require
>>>> EAPI=5 at all and probably make arch testing a bit easier for a few
>>>> people.
>>>
>>> Sounds good to me.
>>
>> Except that it completely breaks stabilization procedures, since packages are 
>> then not only tested with a larger range of useflags, but with an entirely 
>> different profile. Not such a great idea. 
>>
>> The whole point of the stable masking was to keep the changes minimal when 
>> going from a "testing" to a "stable" state - by only restricting the use flag 
>> choices, and nothing else. This means most of the testing done with ~arch 
>> packages is still valid and provides meaningful feedback to maintainers and 
>> arch teams for stabilization.
> 
> Well, it's all a question of decisions, I believe. If we make sure that
> the new 'unstable' profiles differ from the 'stable' ones only by
> additional masked/unmasked USE flags, I don't think it'd be an issue.

Yeah, should be fine.

>> In general, using a separate set of profiles, however, whill not help you 
>> enabling the stable mask files, since these will then only be allowed inside 
>> the new profiles. Not in the base profile or in the main profile directory, 
>> which still follows the old EAPI. In the sense of easy handling, noone will 
>> probably want to edit 
>> profiles/highly_unstable/next_version/package.stable.mask. 
>>
>> <rant>I have basically given up that this "feature" will ever become useful 
>> for the main tree. Long live inertia.</rant>
> 
> I'm thinking of making all the current profiles 'testing'. As in, we
> mask the 'unstable' flags in base profile completely.
> 
> The new profiles will just add a common 'testing' profile which would
> unmask those flags. ~arch users could still use the regular profiles
> but would have to switch if they wanted the additional flags
> automatically unmasked.

That makes sense. The currently supported profile types for
profiles.desc are 'stable', 'dev', and 'exp'. By default, repoman only
checks the 'stable' ones. If you want your new 'testing' profiles to be
checked by default, we could add a new category for them.
-- 
Thanks,
Zac

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Markos Chandras]

On 13 December 2012 21:46, Zac Medico <zmedico@gentoo.org> wrote:
> On 12/13/2012 12:43 PM, Michał Górny wrote:
>> On Thu, 13 Dec 2012 21:33:50 +0100
>> "Andreas K. Huettel" <dilfridge@gentoo.org> wrote:
>>
>>> Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
>>>>> Yes, and having 'stable' and 'unstable' profiles will work just
>>>>> the same. Except for the fact that it will be a bit cleaner, not require
>>>>> EAPI=5 at all and probably make arch testing a bit easier for a few
>>>>> people.
>>>>
>>>> Sounds good to me.
>>>
>>> Except that it completely breaks stabilization procedures, since packages are
>>> then not only tested with a larger range of useflags, but with an entirely
>>> different profile. Not such a great idea.
>>>
>>> The whole point of the stable masking was to keep the changes minimal when
>>> going from a "testing" to a "stable" state - by only restricting the use flag
>>> choices, and nothing else. This means most of the testing done with ~arch
>>> packages is still valid and provides meaningful feedback to maintainers and
>>> arch teams for stabilization.
>>
>> Well, it's all a question of decisions, I believe. If we make sure that
>> the new 'unstable' profiles differ from the 'stable' ones only by
>> additional masked/unmasked USE flags, I don't think it'd be an issue.
>
> Yeah, should be fine.

How are you engoing to ensure that? And how are you going to monitor
them so they will not get out-of-sync in future? We have plenty of
examples of stale profile entries
all over the profiles/arch directory so I think that the stable
*use.stable.mask will also end up
unmaintained in the near future.

-- 
Regards,
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 2126 bytes --]

On Fri, 14 Dec 2012 12:38:24 +0000
Markos Chandras <hwoarang@gentoo.org> wrote:

> On 13 December 2012 21:46, Zac Medico <zmedico@gentoo.org> wrote:
> > On 12/13/2012 12:43 PM, Michał Górny wrote:
> >> On Thu, 13 Dec 2012 21:33:50 +0100
> >> "Andreas K. Huettel" <dilfridge@gentoo.org> wrote:
> >>
> >>> Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
> >>>>> Yes, and having 'stable' and 'unstable' profiles will work just
> >>>>> the same. Except for the fact that it will be a bit cleaner, not require
> >>>>> EAPI=5 at all and probably make arch testing a bit easier for a few
> >>>>> people.
> >>>>
> >>>> Sounds good to me.
> >>>
> >>> Except that it completely breaks stabilization procedures, since packages are
> >>> then not only tested with a larger range of useflags, but with an entirely
> >>> different profile. Not such a great idea.
> >>>
> >>> The whole point of the stable masking was to keep the changes minimal when
> >>> going from a "testing" to a "stable" state - by only restricting the use flag
> >>> choices, and nothing else. This means most of the testing done with ~arch
> >>> packages is still valid and provides meaningful feedback to maintainers and
> >>> arch teams for stabilization.
> >>
> >> Well, it's all a question of decisions, I believe. If we make sure that
> >> the new 'unstable' profiles differ from the 'stable' ones only by
> >> additional masked/unmasked USE flags, I don't think it'd be an issue.
> >
> > Yeah, should be fine.
> 
> How are you engoing to ensure that? And how are you going to monitor
> them so they will not get out-of-sync in future? We have plenty of
> examples of stale profile entries
> all over the profiles/arch directory so I think that the stable
> *use.stable.mask will also end up
> unmaintained in the near future.

What is your solution then? Keeping two revisions of most ebuilds so
that one could be stabilized? I don't see how that is more
maintainable, except for a few days who will easily stay out of it
and pretend that the issue doesn't exist.

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Markos Chandras]

On 14 December 2012 14:29, Michał Górny <mgorny@gentoo.org> wrote:
> On Fri, 14 Dec 2012 12:38:24 +0000
> Markos Chandras <hwoarang@gentoo.org> wrote:
>
>> On 13 December 2012 21:46, Zac Medico <zmedico@gentoo.org> wrote:
>> > On 12/13/2012 12:43 PM, Michał Górny wrote:
>> >> On Thu, 13 Dec 2012 21:33:50 +0100
>> >> "Andreas K. Huettel" <dilfridge@gentoo.org> wrote:
>> >>
>> >>> Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
>> >>>>> Yes, and having 'stable' and 'unstable' profiles will work just
>> >>>>> the same. Except for the fact that it will be a bit cleaner, not require
>> >>>>> EAPI=5 at all and probably make arch testing a bit easier for a few
>> >>>>> people.
>> >>>>
>> >>>> Sounds good to me.
>> >>>
>> >>> Except that it completely breaks stabilization procedures, since packages are
>> >>> then not only tested with a larger range of useflags, but with an entirely
>> >>> different profile. Not such a great idea.
>> >>>
>> >>> The whole point of the stable masking was to keep the changes minimal when
>> >>> going from a "testing" to a "stable" state - by only restricting the use flag
>> >>> choices, and nothing else. This means most of the testing done with ~arch
>> >>> packages is still valid and provides meaningful feedback to maintainers and
>> >>> arch teams for stabilization.
>> >>
>> >> Well, it's all a question of decisions, I believe. If we make sure that
>> >> the new 'unstable' profiles differ from the 'stable' ones only by
>> >> additional masked/unmasked USE flags, I don't think it'd be an issue.
>> >
>> > Yeah, should be fine.
>>
>> How are you engoing to ensure that? And how are you going to monitor
>> them so they will not get out-of-sync in future? We have plenty of
>> examples of stale profile entries
>> all over the profiles/arch directory so I think that the stable
>> *use.stable.mask will also end up
>> unmaintained in the near future.
>
> What is your solution then? Keeping two revisions of most ebuilds so
> that one could be stabilized? I don't see how that is more
> maintainable, except for a few days who will easily stay out of it
> and pretend that the issue doesn't exist.
>
> --
> Best regards,
> Michał Górny

By keeping multiple ebuilds around you are transfering the maintenance
responsibility to the invdividual developer/herd.
By adding the *use.stable.mask to each architecture, you are
transferring this responsibility to the arch maintainers.
We already have plenty of understaffed arches, I don't think it is
wise to throw more responsibilities to them. Unless of course all
developers are allowed to touch these *stable* profiles which
personally I don't like because arches will lose
control of their stable trees.

-- 
Regards,
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Ian Stakenvicius]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Remind me and the list again pls, why is this necessary rather than
just using use.stable.mask and EAPI5 ebuilds in the regular profiles?
This shouldn't break the tree with a non-EAPI5 portage as the files
would just be ignored, as would the EAPI5 ebuilds.....

For some core stuff (like portage) i could see this as being an issue
but we aren't going to need to use.stable.mask flags on core packages
are we?


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iF4EAREIAAYFAlDLOjwACgkQ2ugaI38ACPAaugD/frwJ+kv9w49o1vPGQfLD0uQT
nj2pVXrks/RYUZp+PL8A/1JYcKdzlAup+LIpY/uQzcGwqmtS3U34ZzM7vG+CRQ70
=nXpr
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 4242 bytes --]

On Fri, 14 Dec 2012 14:36:49 +0000
Markos Chandras <hwoarang@gentoo.org> wrote:

> On 14 December 2012 14:29, Michał Górny <mgorny@gentoo.org> wrote:
> > On Fri, 14 Dec 2012 12:38:24 +0000
> > Markos Chandras <hwoarang@gentoo.org> wrote:
> >
> >> On 13 December 2012 21:46, Zac Medico <zmedico@gentoo.org> wrote:
> >> > On 12/13/2012 12:43 PM, Michał Górny wrote:
> >> >> On Thu, 13 Dec 2012 21:33:50 +0100
> >> >> "Andreas K. Huettel" <dilfridge@gentoo.org> wrote:
> >> >>
> >> >>> Am Mittwoch, 12. Dezember 2012, 11:30:17 schrieb Zac Medico:
> >> >>>>> Yes, and having 'stable' and 'unstable' profiles will work just
> >> >>>>> the same. Except for the fact that it will be a bit cleaner, not require
> >> >>>>> EAPI=5 at all and probably make arch testing a bit easier for a few
> >> >>>>> people.
> >> >>>>
> >> >>>> Sounds good to me.
> >> >>>
> >> >>> Except that it completely breaks stabilization procedures, since packages are
> >> >>> then not only tested with a larger range of useflags, but with an entirely
> >> >>> different profile. Not such a great idea.
> >> >>>
> >> >>> The whole point of the stable masking was to keep the changes minimal when
> >> >>> going from a "testing" to a "stable" state - by only restricting the use flag
> >> >>> choices, and nothing else. This means most of the testing done with ~arch
> >> >>> packages is still valid and provides meaningful feedback to maintainers and
> >> >>> arch teams for stabilization.
> >> >>
> >> >> Well, it's all a question of decisions, I believe. If we make sure that
> >> >> the new 'unstable' profiles differ from the 'stable' ones only by
> >> >> additional masked/unmasked USE flags, I don't think it'd be an issue.
> >> >
> >> > Yeah, should be fine.
> >>
> >> How are you engoing to ensure that? And how are you going to monitor
> >> them so they will not get out-of-sync in future? We have plenty of
> >> examples of stale profile entries
> >> all over the profiles/arch directory so I think that the stable
> >> *use.stable.mask will also end up
> >> unmaintained in the near future.
> >
> > What is your solution then? Keeping two revisions of most ebuilds so
> > that one could be stabilized? I don't see how that is more
> > maintainable, except for a few days who will easily stay out of it
> > and pretend that the issue doesn't exist.
> >
> > --
> > Best regards,
> > Michał Górny
> 
> By keeping multiple ebuilds around you are transfering the maintenance
> responsibility to the invdividual developer/herd.
> By adding the *use.stable.mask to each architecture, you are
> transferring this responsibility to the arch maintainers.

Yes and no. Although we have a few arches which believe nobody should
touch their files (but instead wait a few weeks till they actually
notice someone asked them to mask a flag), I think you shouldn't
overreact on this.

Let's talk on examples. A good example is pypy which is not stable
on any arch. I don't really see a problem with Python team actually
maintaining use.mask entry for that implementation. And even if arch
teams preferred to handle this on their own, I don't think that's much
work as long as it's clear what the goal is.

A quick look at dev-python gives me almost 800 packages; a quick ugly
grep gives around 350 packages with stable keywords. This means that in
the near time there will be around 250-300 additional ebuilds to
maintain just because we can't mask the pypy flag on stable arches.

> We already have plenty of understaffed arches, I don't think it is
> wise to throw more responsibilities to them. Unless of course all
> developers are allowed to touch these *stable* profiles which
> personally I don't like because arches will lose
> control of their stable trees.

I'd like to point out that my proposal implies that the *current*
arches become the stable arches, and new sub-arches would be
the testing ones. Therefore, everyone will be allowed to touch like
everyone is allowed to touch the *stable* profiles today.

In other words, we mask python_targets_pypy* in the base profiles,
and unmask them in the testing sub-profiles for amd64 & x86.

-- 
Best regards,
Michał Górny

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

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On Fri, 14 Dec 2012 09:39:56 -0500
Ian Stakenvicius <axs@gentoo.org> wrote:

> Remind me and the list again pls, why is this necessary rather than
> just using use.stable.mask and EAPI5 ebuilds in the regular profiles?
> This shouldn't break the tree with a non-EAPI5 portage as the files
> would just be ignored, as would the EAPI5 ebuilds.....

PMS requires that profiles directory is EAPI=5 before we can use those
files. In other words, we would be forced to make non-EAPI5 portage
ignore the whole profiles.

- -- 
Best regards,
Michał Górny
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)

iJwEAQEIAAYFAlDLPygACgkQfXuS5UK5QB02zAP/bu9eANQrQ01PlKdp/IDDuPxm
VgXi18K1sVUYXNA3n2K46t+7pkX/YewPRBqN8Bn7pHJwsfo/7m/FOfcOvdsdDCoM
sCySNcAKSXq65qG8rKuFK50yWUTNFfSLKC+nFgKlm+Zz8qPuHYAZnkAu5B9idlxV
Fr0DDriG5oRhduqWlEM=
=2Ftf
-----END PGP SIGNATURE-----

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Markos Chandras]

On 14 December 2012 14:59, Michał Górny <mgorny@gentoo.org> wrote:
>
>> We already have plenty of understaffed arches, I don't think it is
>> wise to throw more responsibilities to them. Unless of course all
>> developers are allowed to touch these *stable* profiles which
>> personally I don't like because arches will lose
>> control of their stable trees.
>
> I'd like to point out that my proposal implies that the *current*
> arches become the stable arches, and new sub-arches would be
> the testing ones. Therefore, everyone will be allowed to touch like
> everyone is allowed to touch the *stable* profiles today.
>
> In other words, we mask python_targets_pypy* in the base profiles,
> and unmask them in the testing sub-profiles for amd64 & x86.
>
> --
> Best regards,
> Michał Górny

I fear that the stable and testing profiles will diverge way too much
as time passes. But if you feel that maintainers and
herds will be able to keep the 'diff' between them as minimum as
possible, then I have no objections.

-- 
Regards,
Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2

Re: [gentoo-dev] Getting EAPI 5 *use.stable.mask to work in gx86?

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1385 bytes --]

On Fri, 14 Dec 2012 15:08:24 +0000
Markos Chandras <hwoarang@gentoo.org> wrote:

> On 14 December 2012 14:59, Michał Górny <mgorny@gentoo.org> wrote:
> >
> >> We already have plenty of understaffed arches, I don't think it is
> >> wise to throw more responsibilities to them. Unless of course all
> >> developers are allowed to touch these *stable* profiles which
> >> personally I don't like because arches will lose
> >> control of their stable trees.
> >
> > I'd like to point out that my proposal implies that the *current*
> > arches become the stable arches, and new sub-arches would be
> > the testing ones. Therefore, everyone will be allowed to touch like
> > everyone is allowed to touch the *stable* profiles today.
> >
> > In other words, we mask python_targets_pypy* in the base profiles,
> > and unmask them in the testing sub-profiles for amd64 & x86.
> >
> > --
> > Best regards,
> > Michał Górny
> 
> I fear that the stable and testing profiles will diverge way too much
> as time passes. But if you feel that maintainers and
> herds will be able to keep the 'diff' between them as minimum as
> possible, then I have no objections.

Well, my hope is that we will be able to do it mostly via a common
'testing' profile (or per-arch testing profiles) which will be parents
to other sub-profiles. 

-- 
Best regards,
Michał Górny

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

[gentoo-dev] Re: Getting EAPI 5 *use.stable.mask to work in gx86?

[Duncan]

Michał Górny posted on Fri, 14 Dec 2012 16:15:05 +0100 as excerpted:

> On Fri, 14 Dec 2012 15:08:24 +0000 Markos Chandras <hwoarang@gentoo.org>
> wrote:
> 
>> > I'd like to point out that my proposal implies that the *current*
>> > arches become the stable arches, and new sub-arches would be the
>> > testing ones. Therefore, everyone will be allowed to touch like
>> > everyone is allowed to touch the *stable* profiles today.
>> >
>> > In other words, we mask python_targets_pypy* in the base profiles,
>> > and unmask them in the testing sub-profiles for amd64 & x86.
>> 
>> I fear that the stable and testing profiles will diverge way too much
>> as time passes. But if you feel that maintainers and herds will be able
>> to keep the 'diff' between them as minimum as possible, then I have no
>> objections.
> 
> Well, my hope is that we will be able to do it mostly via a common
> 'testing' profile (or per-arch testing profiles) which will be parents
> to other sub-profiles.

Yes.  Divergence over time is a worry, but we've had cascading profiles 
for quite some time now, so in theory, all that would be needed here is a 
a set of cascading "testing" profiles that simply inherit the stable 
profiles.  Then the testing profiles can simply inherit the stable 
profiles, with the only difference being the new EAPI-5 files in the 
testing profiles (and perhaps eventually the deprecation and in a year or 
two the final removal of the old profiles, with everything from them then 
moved to the new ones).

As long as that's KEPT the only difference...

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman