Re: [gentoo-project] Call for Agenda Items -- Council Meeting 2016-02-14

["Michał Górny" <mgorny@gentoo.org>]

[-- Attachment #1: Type: text/plain, Size: 2340 bytes --]

On Mon, 1 Feb 2016 20:25:20 +0100
"Justin Lecher (jlec)" <jlec@gentoo.org> wrote:

> Dear all,
> 
> the Gentoo Council will meet again on Sunday, February 14 at 19:00 UTC
> in #gentoo-council on FreeNode.
> 
> Please reply to this message with any items you would like us to discuss
> or vote on.

Since the agenda hasn't been sent yet, I would like the Council to
consider helping out with removing cases of 'use*' function calls in
global scope.

The issue is known for quite some time already, and the offending
ebuilds and eclasses are currently tracked in bug #566518 [1].

Those functions can not be used in global scope as their results depend
on configuration. This causes two major issues:

1. it breaks metadata invariancy -- the ebuild metadata may change
dependening on value of USE flags, therefore either invalidating caches
or causing the package manager to obtain incorrect results from cache,

2. it causes circular dependencies in configuration -- USE flags can be
applied to specific SLOTs, SLOTs may depend on USE flags...

The global scope use calls were always forbidden by the PMS. Portage
bans it in EAPI 6 but we'd like to extend the ban to the remaining
EAPIs. However, at the moment we can't because that would cause
existing violations to prevent users from installing toolchain.

So far most of the developers understood the issue and fixed their
violations. However, for years toolchain is actively refusing to do so,
and either closing our requests or blocking them with request to
provide support for USE-dynamic SLOTs. Which is quite unlikely to be
fulfilled since it would still cause the second issue listed above.

Sadly, our inability to ban this is causing new violations to be
committed accidentally by developers.

Therefore, I would like to ask for Council's help on this issue. I see
the possible following actions that would help the cause:

a. officially refusing the 'dynamic slot' request [2] so that it would
not be used to block fixing the QA violations indefinitely,

b. setting an official deadline on fixing the remaining violations
and making the calls fatal.

[1]:https://bugs.gentoo.org/show_bug.cgi?id=566518
[2]:https://bugs.gentoo.org/show_bug.cgi?id=174407

-- 
Best regards,
Michał Górny
<http://dev.gentoo.org/~mgorny/>

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 949 bytes --]