1 |
On Wed, Aug 29, 2012 at 6:14 PM, Michał Górny wrote: |
2 |
> On Wed, 29 Aug 2012 18:05:19 -0400 Mike Frysinger wrote: |
3 |
>> On Wed, Aug 29, 2012 at 6:02 PM, Michał Górny wrote: |
4 |
>> > On Wed, 29 Aug 2012 17:50:16 -0400 Mike Frysinger wrote: |
5 |
>> >> On Wed, Aug 29, 2012 at 5:42 PM, Michał Górny wrote: |
6 |
>> >> > In other words, pkg-config is only used when no other criteria |
7 |
>> >> > allows it to classify the particular .la file as suitable for |
8 |
>> >> > removal or not. Sadly, it's rather, ehm, unfriendly to ebuild |
9 |
>> >> > developers who obviously don't even read the relevant part. |
10 |
>> >> > |
11 |
>> >> > Do you have any ideas how we can improve that? |
12 |
>> >> |
13 |
>> >> before the func executes pkg-config, run `has virtual/pkgconfig |
14 |
>> >> ${DEPEND}` and spit an eqawarn if it's not found |
15 |
>> > |
16 |
>> > Ciaran will shot at me for doing that. |
17 |
>> |
18 |
>> it isn't violating anything and can find real bugs. i don't see a |
19 |
>> problem here. |
20 |
> |
21 |
> It is violating the Holy PMS. |
22 |
|
23 |
does it actually ? are DEPEND variables not allowed to be expanded in |
24 |
pkg_* src_* funcs ? |
25 |
|
26 |
we could probably add a similar check to autotools.eclass: grep for |
27 |
PKG_PROG_PKG_CONFIG and check ${DEPEND} |
28 |
|
29 |
>> >> > One thing that comes into my mind is finally making pkgconfig |
30 |
>> >> > a required, implicit part of toolchain (or @system). Since we |
31 |
>> >> > have pkgconf now, this is more feasible than before. |
32 |
>> >> |
33 |
>> >> i don't think making it part of the toolchain makes sense. i'd |
34 |
>> >> rather not add it to @system simply to keep a few packages from |
35 |
>> >> sometimes failing. |
36 |
>> > |
37 |
>> > I'd add it to @system because a lot of packages actually need to |
38 |
>> > DEPEND on pkgconfig because they use libraries using .pc files. And |
39 |
>> > the number is going to increase, hopefully. |
40 |
>> |
41 |
>> sure, but keeping things in @system doesn't make much sense: |
42 |
>> - there's a penalty (as noted in old threads) |
43 |
>> - it isn't actually required at runtime, so it's bloat on reduced |
44 |
>> systems |
45 |
> |
46 |
> I think it's practically the same as compiler. |
47 |
|
48 |
that isn't a bad view point, but for the purposes of this discussion, |
49 |
i don't think it's relevant :) |
50 |
|
51 |
> Also, a quick look at !ddep shows over 7000 reverse dependencies. That |
52 |
> looks like a bigger penalty to me. |
53 |
|
54 |
if we had a @build-system, you might be able to convince me. but we |
55 |
don't. so the number of packages here doesn't matter as it's an |
56 |
invalid implicit RDEPEND. |
57 |
-mike |