1 |
On Wed, 29 Aug 2012 17:50:16 -0400 |
2 |
Mike Frysinger <vapier@g.o> wrote: |
3 |
|
4 |
> On Wed, Aug 29, 2012 at 5:42 PM, Michał Górny wrote: |
5 |
> > In other words, pkg-config is only used when no other criteria |
6 |
> > allows it to classify the particular .la file as suitable for |
7 |
> > removal or not. Sadly, it's rather, ehm, unfriendly to ebuild |
8 |
> > developers who obviously don't even read the relevant part. |
9 |
> > |
10 |
> > Do you have any ideas how we can improve that? |
11 |
> |
12 |
> before the func executes pkg-config, run `has virtual/pkgconfig |
13 |
> ${DEPEND}` and spit an eqawarn if it's not found |
14 |
|
15 |
Ciaran will shot at me for doing that. |
16 |
|
17 |
> > One thing that comes into my mind is finally making pkgconfig |
18 |
> > a required, implicit part of toolchain (or @system). Since we have |
19 |
> > pkgconf now, this is more feasible than before. |
20 |
> |
21 |
> i don't think making it part of the toolchain makes sense. i'd rather |
22 |
> not add it to @system simply to keep a few packages from sometimes |
23 |
> failing. |
24 |
|
25 |
I'd add it to @system because a lot of packages actually need to DEPEND |
26 |
on pkgconfig because they use libraries using .pc files. And the number |
27 |
is going to increase, hopefully. |
28 |
|
29 |
Also, some people are probably going to try to get some pkgconf support |
30 |
directly into gcc, in form of '-something libfoo' to make it grab |
31 |
everything magically, I think. |
32 |
|
33 |
-- |
34 |
Best regards, |
35 |
Michał Górny |