| 1 |
On 12/05/14 20:47, Peter Stuge wrote: |
| 2 |
> Rich Freeman wrote: |
| 3 |
>>> Longterm, this makes it year after year more difficult to develop |
| 4 |
>>> software for "Linux". |
| 5 |
>> I'm with you here, but what is the solution? |
| 6 |
>> |
| 7 |
>> If we say we stick to upstream then we don't provide pkg-config files |
| 8 |
>> at all (in these cases). |
| 9 |
> I think this is a sane default. |
| 10 |
|
| 11 |
Except having pkg-config is the only way to fix some of the build issues |
| 12 |
we are seeing |
| 13 |
today, like getting 'Libs.private: ' for static linking, there has been |
| 14 |
multiple bugs lately, |
| 15 |
and we are in middle of process of obsoleting every custom foo-config |
| 16 |
due to same |
| 17 |
reasons, so having pkg-config files is an absolute requirement. |
| 18 |
Some binary-only distros might get away without them, but we won't. |
| 19 |
|
| 20 |
> |
| 21 |
> |
| 22 |
>> Then when Debian does the other upstreams use them and then those |
| 23 |
>> packages break on Gentoo. |
| 24 |
> I like Gentoo to stay very close to upstream. |
| 25 |
> |
| 26 |
> If upstream pkg A depends on $distro-specific foo of pkg B then that |
| 27 |
> will obviously not work in an environment only following upstreams, |
| 28 |
> and will require effort to untie gentoo pkg A from $distro specifics. |
| 29 |
|
| 30 |
pkg-config by design works without .pc files if needed, by exporting |
| 31 |
FOO_LIBS |
| 32 |
and FOO_CFLAGS, so if this is the only problem with them, it's really no |
| 33 |
problem |
| 34 |
at all compared to the problems caused by lacking the pkg-config files |
| 35 |
|
| 36 |
|
| 37 |
(Are we seriously discussing banning something useful as pkg-config |
| 38 |
files?! That's retarded. Must be some joke.) |
Archives