Bug 463692 – [Future EAPI] Provide PATCHES array support in default phase of src_prepare

• Product: Gentoo Hosted Projects
• Status: RESOLVED FIXED

[Description Tomáš Chvátal (RETIRED) 2013-03-29 09:18:29 UTC]

As per summary this is the only thing left in base eclass that eapi5 can't do.
So prefferably with next gen we should be just able to use it everywhere.

Also some developers refuse to use the array even if they inherit the base eclass because it is not in eapi, so it would be nice to have it and used everywhere.

[Comment 1 Michał Górny 2013-03-29 09:26:38 UTC]

You've got my +1. That was the thing I forgot about!

[Comment 2 Tomáš Chvátal (RETIRED) 2013-03-29 09:31:06 UTC]

Also another bit wrt this bug might be support for epatch_user, but that thing must be redone a bit.

If it would be enabled on "all packages" then it must be obvious from starters if the patches are provided by user or by us. Something in emerge.log about those with huge "!!!! USER_PATCH: blabla-i-screwed-my-build.patch" could be enough... :-)

[Comment 3 Michał Górny 2013-03-29 09:47:58 UTC]

(In reply to comment #2)
> Also another bit wrt this bug might be support for epatch_user, but that
> thing must be redone a bit.
> 
> If it would be enabled on "all packages" then it must be obvious from
> starters if the patches are provided by user or by us. Something in
> emerge.log about those with huge "!!!! USER_PATCH:
> blabla-i-screwed-my-build.patch" could be enough... :-)

Well, I'd say it's pretty obvious right now that the code would have to be quite verbose.

Hmm but we don't have epatch in the EAPI. We should have a bug for that as well. I don't really want EAPI once again not to provide the tools it uses.

[Comment 4 Julian Ospald 2013-03-29 17:58:08 UTC]

epatch is not in PMS, so how does this compute?

[Comment 5 Michał Górny 2013-03-29 17:59:32 UTC]

(In reply to comment #4)
> epatch is not in PMS, so how does this compute?

I see two possibilities:

a) we add epatch ;),

b) we use something like 'patch -p1' and end the epatch mess.

[Comment 6 Julian Ospald 2013-03-29 18:00:09 UTC]

I am for a)

[Comment 7 Ulrich Müller 2013-08-18 12:30:03 UTC]

Should both an array and a whitespace separated list be supported, as in base.eclass? That would also be consistent with the DOCS variable.

[Comment 8 Tomáš Chvátal (RETIRED) 2013-08-18 15:13:00 UTC]

(In reply to Ulrich Müller from comment #7)
> Should both an array and a whitespace separated list be supported, as in
> base.eclass? That would also be consistent with the DOCS variable.

Indiferent, whatever is inlined with rest of the variables would be best.
OTOH question is how many packages do we have that still use non-array patches if it is really really needed.

[Comment 9 Ulrich Müller 2014-06-16 10:47:17 UTC]

We support whitespace separated lists for all other package manager variables, so I think for reasons of consistency we should support them also for PATCHES. Also it is trivial to implement (simply copy the test for the DOCS variable from src_install).

[Comment 10 Ulrich Müller 2015-11-08 21:49:46 UTC]

In EAPI 6:
https://gitweb.gentoo.org/proj/pms.git/commit/?id=628de448c6a0a7e9b6bac8d206d4288f5bb6933b