1 |
On Tue, Nov 29, 2016 at 6:24 PM, Mike Pagano <mpagano@g.o> wrote: |
2 |
> 2nd round. As per mgorny's advice, fix eapply_user as per PMS spec and |
3 |
> execute in src_prepare. Support older EAPIs with epatch_user. |
4 |
> |
5 |
> --- |
6 |
> eclass/kernel-2.eclass | 25 ++++++++++++++++++------- |
7 |
> 1 file changed, 18 insertions(+), 7 deletions(-) |
8 |
> |
9 |
> diff --git a/eclass/kernel-2.eclass b/eclass/kernel-2.eclass |
10 |
> index 3f5fb3b..e516aff 100644 |
11 |
> --- a/eclass/kernel-2.eclass |
12 |
> +++ b/eclass/kernel-2.eclass |
13 |
> @@ -95,7 +95,7 @@ has "${EAPI:-0}" 0 1 2 3 4 5 || die "kernel-2.eclass |
14 |
> is unsupported for EAPI ${E |
15 |
> PYTHON_COMPAT=( python{2_6,2_7} ) |
16 |
> |
17 |
> inherit eutils toolchain-funcs versionator multilib python-any-r1 |
18 |
> -EXPORT_FUNCTIONS pkg_setup src_unpack src_compile src_test src_install |
19 |
> pkg_preinst pkg_postinst pkg_postrm |
20 |
> +EXPORT_FUNCTIONS pkg_setup src_prepare src_unpack src_compile src_test |
21 |
> src_install pkg_preinst pkg_postinst pkg_postrm |
22 |
|
23 |
I don't think calling "EXPORT_FUNCTIONS src_prepare" is legal in EAPI |
24 |
0 or 1. You should probably move this behind an EAPI check. |
25 |
|
26 |
Also, for EAPI, 2, 3, 4, and 5, this change might break existing |
27 |
ebuilds; they may define src_prepare themselves, or inherit some other |
28 |
eclass that also exports src_prepare. |
29 |
|
30 |
Otherwise, this seems like a good next step. |