1 |
On Fri, 29 Jul 2016 17:42:40 +0200 Michał Górny wrote: |
2 |
> On Fri, 29 Jul 2016 18:31:58 +0300 |
3 |
> Andrew Savchenko <bircoph@g.o> wrote: |
4 |
> |
5 |
> > diff --git a/eclass/subversion.eclass b/eclass/subversion.eclass |
6 |
> > index c45c725..49040a4 100644 |
7 |
> > --- a/eclass/subversion.eclass |
8 |
> > +++ b/eclass/subversion.eclass |
9 |
> > @@ -434,6 +434,7 @@ subversion_src_unpack() { |
10 |
> > # Default src_prepare. Bootstrap. |
11 |
> > subversion_src_prepare() { |
12 |
> > subversion_bootstrap || die "${ESVN}: unknown problem occurred in subversion_bootstrap." |
13 |
> > + [[ ${EAPI:-0} -ge 6 ]] && eapply_user |
14 |
> |
15 |
> EAPI is not always a number, so this is wrong. |
16 |
|
17 |
Then as QA you should go and fix xorg-2.eclass :) |
18 |
|
19 |
Due to changes below this code is not needed in the |
20 |
subversion.eclass. |
21 |
|
22 |
> Furthermore, since it didn't work in EAPI 6 anyway, please take a look |
23 |
> into improving this eclass for EAPI 6. In particular: |
24 |
> |
25 |
> - add a proper EAPI check preventing it from implicitly being enabled |
26 |
> in new EAPIs, |
27 |
|
28 |
Done. |
29 |
|
30 |
> - disable subversion_bootstrap() along with all the relevant facilities |
31 |
> in EAPI 6 and later, therefore also stopping exporting src_prepare(). |
32 |
|
33 |
Done, no need for eappy_user in this case. |
34 |
|
35 |
Please in future explain such requirements to save fellow |
36 |
developer's time, since subversion_bootstrap doesn't contradict |
37 |
directly with EAPI 6 (ESVN_PATCHES is not needed with PATCHES |
38 |
available in EAPI 6, but this is another matter). The real reason |
39 |
for removal is that nobody uses it in the main tree anymore. |
40 |
|
41 |
Best regards, |
42 |
Andrew Savchenko |