| 1 |
On Thu, Aug 30, 2012 at 3:41 AM, Michał Górny wrote: |
| 2 |
> On Wed, 29 Aug 2012 19:12:01 -0400 Mike Frysinger wrote: |
| 3 |
>> On Wed, Aug 29, 2012 at 6:37 PM, Ciaran McCreesh wrote: |
| 4 |
>> > On Wed, 29 Aug 2012 18:18:20 -0400 Mike Frysinger wrote: |
| 5 |
>> >> does it actually ? are DEPEND variables not allowed to be |
| 6 |
>> >> expanded in pkg_* src_* funcs ? |
| 7 |
>> > |
| 8 |
>> > Nope. We don't guarantee that the metadata variable gets exported |
| 9 |
>> > back to the ebuild environment. |
| 10 |
>> |
| 11 |
>> it's not a requirement, so if the PM doesn't export it, that's not a |
| 12 |
>> problem. |
| 13 |
>> |
| 14 |
>> if [[ ${DEPEND+set} == "set" ]] && ! has virtual/pkgconifg |
| 15 |
>> ${DEPEND} ; then eqawarn "..." |
| 16 |
>> fi |
| 17 |
> |
| 18 |
> It may have a semi-random value, I think. Then you could warn for |
| 19 |
> perfectly valid ebuilds. |
| 20 |
|
| 21 |
in my quick test it seemed to have a fine value. let's deploy it and |
| 22 |
see what kind of reports we get back. |
| 23 |
|
| 24 |
you could also make it do: |
| 25 |
libs=$("${pkgconf}" --libs "${tf}") || die |
| 26 |
-mike |
Archives