[gentoo-project] Call for agenda items - Council meeting 2013-05-14

[gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Petteri Räty]

[-- Attachment #1: Type: text/plain, Size: 616 bytes --]

In two weeks from now, the council will meet again. This is the time
to raise and prepare items that the council should put on the agenda
to discuss or vote on.

Please respond to this email with agenda items. Please do not hesitate
to repeat your agenda item here with a pointer if you previously
suggested one (since the last meeting).

The agenda for the next meeting will be sent out on Tuesday
7th of May 2013.

Please respond to the gentoo-project list for the items and have the
actual discussion on the items on the best suited mailing list (usually
gentoo-dev or gentoo-project).

Petteri


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 866 bytes --]

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 1558 bytes --]

On Sat, 27 Apr 2013 19:59:20 +0300
Petteri Räty <betelgeuse@gentoo.org> wrote:

> Please respond to this email with agenda items. Please do not hesitate
> to repeat your agenda item here with a pointer if you previously
> suggested one (since the last meeting).

I would like to request the Council to vote on the topic of changing
the ${D} and ${ROOT} variables not to be terminated with a trailing
slash. To decrease confusion and improve forward compatibility, this
would benefit from retroactively removing the 'must end with a trailing
slash' guarantees from PMS, changing it to undefined and working on
fixing the ebuilds which relied on those.

Then, in a future EAPI a common guarantee will be added that all path
variables do not end with a trailing slash.

The topic has been brought on the ml [1] and there was some discussion
in the relevant bug [2] as well. I've written a simple plan on how to
handle it effectively [3], attached a PMS patch and two dedicated QA
checks for repoman.

The idea has met mostly positive feedback. Ciaran is opposed to it.
The number of ebuilds needing changing for the 'no trailing slash'
standard is estimated to be at least 30 times smaller than those which
incorrectly assume no trailing slash in ${D} -- which proves that
the current rule is unknown to most of the developers.

[1]:http://thread.gmane.org/gmane.linux.gentoo.devel/84989
[2]:https://bugs.gentoo.org/show_bug.cgi?id=465772
[3]:https://bugs.gentoo.org/show_bug.cgi?id=465772#c20

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 966 bytes --]

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Ulrich Mueller]

[-- Attachment #1: Type: text/plain, Size: 212 bytes --]

I'd like to ask for a vote on a clarification of PMS wording, namely
that econf arguments are passed to configure after default options.
Patch is here: http://article.gmane.org/gmane.linux.gentoo.pms/751

Ulrich

[-- Attachment #2: Type: application/pgp-signature, Size: 490 bytes --]

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Mike Gilbert]

On Sat, Apr 27, 2013 at 12:59 PM, Petteri Räty <betelgeuse@gentoo.org> wrote:
> In two weeks from now, the council will meet again. This is the time
> to raise and prepare items that the council should put on the agenda
> to discuss or vote on.
>
> Please respond to this email with agenda items. Please do not hesitate
> to repeat your agenda item here with a pointer if you previously
> suggested one (since the last meeting).
>
> The agenda for the next meeting will be sent out on Tuesday
> 7th of May 2013.
>
> Please respond to the gentoo-project list for the items and have the
> actual discussion on the items on the best suited mailing list (usually
> gentoo-dev or gentoo-project).
>

I would like to get the council's opinion on enabling the
"preserve-libs" portage feature by default in the stable portage
branch.

This feature compliments the slot-operator feature introduced with
EAPI 5 by keeping libraries installed until reverse dependencies can
be rebuilt.

Zac Medico seems to be ok with it, but requested that I ask the
council to discuss it.

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Zac Medico]

I'd like to get the Council's opinion on having FEATURES=preserve-libs
enabled by default in stable portage. About 6 months ago I mentioned on
my blog [1] that I might consider enabling it by default after
slot-operators and sub-slots became widely adopted. We seem to be
reaching a satisfactory level of adoption now.

I know that this feature has been questioned by some, especially by
people involved with Paludis (which doesn't implement preserve-libs). I
think that the main compliant is that preserve-libs doesn't preserve any
non-library dependencies (such as configuration files) that a library
may depend on.

[1]
http://blogs.gentoo.org/zmedico/2012/09/21/preserve-libs-available-in-portage-2-1/
-- 
Thanks,
Zac

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 581 bytes --]

On Wed, 01 May 2013 08:18:50 -0700
Zac Medico <zmedico@gentoo.org> wrote:
> I know that this feature has been questioned by some, especially by
> people involved with Paludis (which doesn't implement preserve-libs).
> I think that the main compliant is that preserve-libs doesn't
> preserve any non-library dependencies (such as configuration files)
> that a library may depend on.

...and that it's utterly frickin' broken as a concept, and that
adopting it will slow down people switching to the proper solution to
the problem, which is slots.

-- 
Ciaran McCreesh

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

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 852 bytes --]

On Wed, 1 May 2013 16:22:47 +0100
Ciaran McCreesh <ciaran.mccreesh@googlemail.com> wrote:

> On Wed, 01 May 2013 08:18:50 -0700
> Zac Medico <zmedico@gentoo.org> wrote:
> > I know that this feature has been questioned by some, especially by
> > people involved with Paludis (which doesn't implement preserve-libs).
> > I think that the main compliant is that preserve-libs doesn't
> > preserve any non-library dependencies (such as configuration files)
> > that a library may depend on.
> 
> ...and that it's utterly frickin' broken as a concept, and that
> adopting it will slow down people switching to the proper solution to
> the problem, which is slots.

Convince the developers to split packages into proper parts, then we
can talk. Or even better, convince upstreams to split their packages.

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 966 bytes --]

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 1074 bytes --]

On Wed, 1 May 2013 17:28:02 +0200
Michał Górny <mgorny@gentoo.org> wrote:
> On Wed, 1 May 2013 16:22:47 +0100
> Ciaran McCreesh <ciaran.mccreesh@googlemail.com> wrote:
> > On Wed, 01 May 2013 08:18:50 -0700
> > Zac Medico <zmedico@gentoo.org> wrote:
> > > I know that this feature has been questioned by some, especially
> > > by people involved with Paludis (which doesn't implement
> > > preserve-libs). I think that the main compliant is that
> > > preserve-libs doesn't preserve any non-library dependencies (such
> > > as configuration files) that a library may depend on.
> > 
> > ...and that it's utterly frickin' broken as a concept, and that
> > adopting it will slow down people switching to the proper solution
> > to the problem, which is slots.
> 
> Convince the developers to split packages into proper parts, then we
> can talk. Or even better, convince upstreams to split their packages.

It's necessary, and the way to convince developers is to stop providing
a nasty hack as a not-really-working alternative.

-- 
Ciaran McCreesh

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

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Michał Górny]

[-- Attachment #1: Type: text/plain, Size: 473 bytes --]

On Sat, 27 Apr 2013 19:59:20 +0300
Petteri Räty <betelgeuse@gentoo.org> wrote:

> Please respond to this email with agenda items. Please do not hesitate
> to repeat your agenda item here with a pointer if you previously
> suggested one (since the last meeting).

And once again, I'd like to ask the Council to start voting on EAPI 6
items.

'Collecting items' thread:
  http://thread.gmane.org/gmane.linux.gentoo.devel/84704

-- 
Best regards,
Michał Górny

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 966 bytes --]

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Rich Freeman]

On Wed, May 1, 2013 at 11:22 AM, Ciaran McCreesh
<ciaran.mccreesh@googlemail.com> wrote:
> ...and that it's utterly frickin' broken as a concept, and that
> adopting it will slow down people switching to the proper solution to
> the problem, which is slots.

Unless we fully slot (not subslot) every version of everything that
gets linked there is still the issue of what happens between the time
package A is upgraded, and package B which depends on A is upgraded.
That will often be minutes, and could be hours or days even with slot
operator dependencies.  Some packages take a long time to build, and
if there is a failure it might take a user days to work it out.

Even if preserve-libs is imperfect, it improves usability.  Are there
any scenarios where it actually makes things worse?  Sure, maybe once
in a blue moon a lib no longer works because it is out of sync with a
config file, but would it work any better if it were just outright
deleted?

Rich

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Zac Medico]

On 05/01/2013 08:35 AM, Ciaran McCreesh wrote:
> On Wed, 1 May 2013 17:28:02 +0200
> Michał Górny <mgorny@gentoo.org> wrote:
>> On Wed, 1 May 2013 16:22:47 +0100
>> Ciaran McCreesh <ciaran.mccreesh@googlemail.com> wrote:
>>> On Wed, 01 May 2013 08:18:50 -0700
>>> Zac Medico <zmedico@gentoo.org> wrote:
>>>> I know that this feature has been questioned by some, especially
>>>> by people involved with Paludis (which doesn't implement
>>>> preserve-libs). I think that the main compliant is that
>>>> preserve-libs doesn't preserve any non-library dependencies (such
>>>> as configuration files) that a library may depend on.
>>>
>>> ...and that it's utterly frickin' broken as a concept, and that
>>> adopting it will slow down people switching to the proper solution
>>> to the problem, which is slots.
>>
>> Convince the developers to split packages into proper parts, then we
>> can talk. Or even better, convince upstreams to split their packages.
> 
> It's necessary, and the way to convince developers is to stop providing
> a nasty hack as a not-really-working alternative.

As downstream packagers, do we or can we really expect to have that much
influence of upstream developers? If the big binary distros are willing
to package these things without complaints, then how likely is it that
upstream developers will change their ways?
-- 
Thanks,
Zac

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 564 bytes --]

On Wed, 01 May 2013 10:27:27 -0700
Zac Medico <zmedico@gentoo.org> wrote:
> > It's necessary, and the way to convince developers is to stop
> > providing a nasty hack as a not-really-working alternative.
> 
> As downstream packagers, do we or can we really expect to have that
> much influence of upstream developers? If the big binary distros are
> willing to package these things without complaints, then how likely
> is it that upstream developers will change their ways?

Big binary distros already deal with this problem...

-- 
Ciaran McCreesh

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

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Rich Freeman]

On Wed, May 1, 2013 at 1:30 PM, Ciaran McCreesh
<ciaran.mccreesh@googlemail.com> wrote:
> On Wed, 01 May 2013 10:27:27 -0700
> Zac Medico <zmedico@gentoo.org> wrote:
>> > It's necessary, and the way to convince developers is to stop
>> > providing a nasty hack as a not-really-working alternative.
>>
>> As downstream packagers, do we or can we really expect to have that
>> much influence of upstream developers? If the big binary distros are
>> willing to package these things without complaints, then how likely
>> is it that upstream developers will change their ways?
>
> Big binary distros already deal with this problem...

Well, if you only support one version of every library and change them
all in lockstep a few times a year, then this becomes a complete
non-issue.  However, if you're going to do that then a big part of the
value that Gentoo creates is lost.

Honestly, this seems a bit like hurting our users so that developers
feel sorry for them and jump through hoops.  We might as well threaten
to kill kittens anytime the bug wrangler list goes over a certain
threshold, or a GLSA stays open too long.

Sure, I'd prefer a perfect world to a better world, but I'm not going
to get a perfect world, so I'd prefer a better world to the one I live
in today.

Rich

Re: [gentoo-project] Call for agenda items - Council meeting 2013-05-14

[Ciaran McCreesh]

[-- Attachment #1: Type: text/plain, Size: 876 bytes --]

On Wed, 1 May 2013 13:35:25 -0400
Rich Freeman <rich0@gentoo.org> wrote:
> > Big binary distros already deal with this problem...
> 
> Well, if you only support one version of every library and change them
> all in lockstep a few times a year, then this becomes a complete
> non-issue.  However, if you're going to do that then a big part of the
> value that Gentoo creates is lost.

No, they also tend to get their developers to split out libraries
where necessary, allowing multiple versions of the libs to be installed
in parallel, correctly. Granted, the way most of them do it is ugly,
but their developers do put in the work.

> Sure, I'd prefer a perfect world to a better world, but I'm not going
> to get a perfect world, so I'd prefer a better world to the one I live
> in today.

"Barely adequate" is the enemy of "good".

-- 
Ciaran McCreesh

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