[gentoo-dev] Python 3.10.0rc1

2021-08-03 Thread Michał Górny
Hi, everyone.

Just a quick note: I've pushed Python 3.10.0rc1 today.  It has many last
minute changes that can break packages that were ported to previous
3.10.0 betas before.

For practical reasons, we're going to support only >=3.10.0_rc1 going
forward.  If your package fails with >=3.10.0_rc1, feel free to apply
fixes without caring for backwards compatibility with betas.  If you see
failures with 3.10.0_beta4, please try upgrading to _rc1 first.

Notably, the newest releases of dev-python/django and dev-python/sphinx
that I've pushed today were updated for _rc1 and will have failures with
_beta4.

-- 
Best regards,
Michał Górny





Re: [gentoo-dev] [PATCH v2] xdg.eclass: add EAPI 8 support

2021-08-03 Thread Mart Raudsepp
Ühel kenal päeval, K, 21.07.2021 kell 17:08, kirjutas Florian Schmaus:
> Note that this removes the export of src_prepare in EPAI 8 as
> requested by ionen:
> 
>   1. remove src_prepare export in EAPI-8
> 
>   While "some" packages need xdg_environment_reset, most don't
> because
>   the eclass is often only inherited to handle icons/.desktop and
> this
>   just needlessly overwrite the src_prepare of other eclasses
> requiring
>   more careful inherit ordering (e.g. inherit xdg cmake).
> 
>   I'd prefer it was clear when a package need this by calling
>   xdg_environment_reset directly. Unless there is a non-trivial
> amount
>   of packages that need it (e.g. for tests) that I'm not aware of.

asturm asked me to reply here if I think the changes are in the spirit
of xdg.eclass, so..

I'm not sure about the spirit, but I personally am fine with the
changes, after the technical nitpicks are figured out in the thread
here.

I do not like at all that we'll need to remember about calling
xdg_environment_reset sometimes, but the status quo of clashing with
other eclasses src_prepare export is probably worse.
And you needing this reset call or not is not at all immediately
obvious - it causing trouble may only happen during test suite run, for
example, when XDG_RUNTIME_DIR is written to, or something else ends up
causing writes to it. Basically ENV_UNSET was a great start, but does
not completely address our XDG env reset needs and it'd be great if
someone championed a full solution for EAPI-9 or something.

I'm starting to think the easiest is to just make the
xdg_environment_reset call unconditionally in meson.eclass
src_configure like cmake.eclass does, and forget all about this for
almost all possible cases.

That rant aside, I'm happy in spirit with the changes as proposed here.


Mart


signature.asc
Description: This is a digitally signed message part


Re: [gentoo-dev] [PATCH] metadata/install-qa-check.d: add 60tmpfiles-path QA check

2021-08-03 Thread Sam James


> On 4 Aug 2021, at 03:39, Thomas Deutschmann  wrote:
> 
> On 2021-08-01 01:56, Sam James wrote:
>> 1) Verify packages don't install tmpfiles to /etc/tmpfiles.d, which
>> is a deprecated location;
> 
> This location is _not_ deprecated.
> 
> This is the location for the local system administrator to override 
> tmpfiles.d specifications installed by packages which should install below 
> /usr.
> 
> 

Sure, thanks for the clarification. It's deprecated in the sense of ebuilds 
installing to it though, right?

best,
sam


signature.asc
Description: Message signed with OpenPGP


Re: [gentoo-dev] [PATCH] metadata/install-qa-check.d: add 60tmpfiles-path QA check

2021-08-03 Thread Thomas Deutschmann

On 2021-08-01 01:56, Sam James wrote:

1) Verify packages don't install tmpfiles to /etc/tmpfiles.d, which
is a deprecated location;


This location is _not_ deprecated.

This is the location for the local system administrator to override 
tmpfiles.d specifications installed by packages which should install 
below /usr.



--
Regards,
Thomas Deutschmann / Gentoo Linux Developer
fpr: C4DD 695F A713 8F24 2AA1 5638 5849 7EE5 1D5D 74A5



OpenPGP_signature
Description: OpenPGP digital signature