1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA256 |
3 |
|
4 |
On 12/08/15 02:19 PM, Alexis Ballier wrote: |
5 |
> On Wed, 12 Aug 2015 20:00:42 +0200 Ulrich Mueller |
6 |
> <ulm@g.o> wrote: |
7 |
> |
8 |
>>>>>>> On Wed, 12 Aug 2015, Ciaran McCreesh wrote: |
9 |
>> |
10 |
>>>> pkg_pretend() { if use qt4; then required_use -qt5 else |
11 |
>>>> required_use qt5 fi } |
12 |
>> |
13 |
>>> And how would the PM understand that -qt5 is conditional upon |
14 |
>>> qt4? Such knowledge is required if it's supposed to |
15 |
>>> auto-resolve stuff... |
16 |
>> |
17 |
>> Right, the above was too simple (and wrong). It should have |
18 |
>> been: |
19 |
>> |
20 |
>> pkg_pretend() { use qt4 && use qt5 && required_use -qt5 use qt4 |
21 |
>> || use qt5 || required_use qt4 } |
22 |
> |
23 |
> what is the difference ? |
24 |
> |
25 |
> pkg_pretend still needs to be executed to guess what useflags |
26 |
> are enabled or not, which information is needed before |
27 |
> dependency calculation |
28 |
> |
29 |
> or are we talking about moving pkg_pretend into dependency |
30 |
> calculation? |
31 |
> |
32 |
|
33 |
|
34 |
pkg_pretend is already executed during dependency calculation in |
35 |
portage, although this doesn't seem to actually be specified in PMS: |
36 |
"The pkg_pretend function is called some unspecified time before a |
37 |
(possibly hypothetical) normal sequence." as per PMS sec.9.2 |
38 |
|
39 |
|
40 |
-----BEGIN PGP SIGNATURE----- |
41 |
Version: GnuPG v2 |
42 |
|
43 |
iF4EAREIAAYFAlXLj0YACgkQAJxUfCtlWe1NwgD/UlFiaP1Xyh4yfHlEH+nGIYN/ |
44 |
NTqMi1IFiKXpgKJjVWEA/1BdSHkivkFhodChHJ0jlYCX9xHg0Xc0VDzPN0S1aPRn |
45 |
=KQmI |
46 |
-----END PGP SIGNATURE----- |