1 |
On 21/02/15 20:36, Michał Górny wrote: |
2 |
>> # @ECLASS-VARIABLE: EBO_EXTRA_ECONF # @DEFAULT_UNSET # |
3 |
>> @DESCRIPTION: # Extra config options passed to econf, similar to |
4 |
>> EXTRA_ECONF. |
5 |
> |
6 |
> Similar as in how? That it must not be used in ebuilds but in |
7 |
> make.conf? |
8 |
|
9 |
Probably missleading name. All the consuming packages have a >90% |
10 |
identical buildsystem. We thought this option to add an easy way to |
11 |
handle the rest. But this should be an eclass/ebuild variable and not |
12 |
used by the user. |
13 |
|
14 |
> |
15 |
>> |
16 |
>> case ${EAPI:-0} in 5) ;; *) die "this eclass doesn't support < |
17 |
>> EAPI 5" ;; esac |
18 |
>> |
19 |
>> if [[ -f "${FILESDIR}"/${P}_fix-build-system.patch ]]; then |
20 |
> |
21 |
> This looks terribly hacky. I'm not convinced you can rely on test |
22 |
> like this working properly in all cases, and I'm pretty convinced |
23 |
> it's a terribly unpredictable form of an API. |
24 |
|
25 |
I already thought that this isn't the best way. This means, always |
26 |
running eautoreconf. As said before, we can be nearly 100% sure that |
27 |
there always be a reconf needed. |
28 |
|
29 |
>> SRC_URI="ftp://emboss.open-bio.org/pub/EMBOSS/${EF}.tar.gz -> |
30 |
>> embassy-${EN}-${PVR}.tar.gz" |
31 |
> |
32 |
> Why PVR? Why not just ${EF}? :P |
33 |
|
34 |
Upstream sucks and updates packages without version bumping. so we |
35 |
need to have it more fine graded. |
36 |
>> # 2. Patches with "${FILESDIR}"/${PF}.patch, if present |
37 |
> |
38 |
> Ugly API, I dare say. Prefer explicit. |
39 |
|
40 |
Agreed. |
41 |
|
42 |
|
43 |
Thanks for the comments, |
44 |
|
45 |
Justin |