| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
Fabian Groffen wrote: |
| 5 |
> Just to have it clear and to be sure: |
| 6 |
> |
| 7 |
> On 04-11-2007 03:33:41 +0000, vapier@g.o wrote: |
| 8 |
>> Modified: diffutils-2.8.7-r2.ebuild |
| 9 |
>> Log: |
| 10 |
>> do *not* include userland_GNU in IUSE |
| 11 |
>> (Portage version: 2.1.3.16) |
| 12 |
>> |
| 13 |
>> |
| 14 |
>> Index: diffutils-2.8.7-r2.ebuild |
| 15 |
> |
| 16 |
>> -IUSE="nls static userland_GNU" |
| 17 |
>> +IUSE="nls static" |
| 18 |
> |
| 19 |
> |
| 20 |
> On 04-11-2007 08:10:29 +0000, Zac Medico wrote: |
| 21 |
>> Author: zmedico |
| 22 |
>> Date: 2007-11-04 08:10:29 +0000 (Sun, 04 Nov 2007) |
| 23 |
>> New Revision: 8420 |
| 24 |
>> |
| 25 |
>> Modified: |
| 26 |
>> main/trunk/pym/portage/dbapi/bintree.py |
| 27 |
>> Log: |
| 28 |
>> When evaluating *DEPEND conditionals for the Packages metadata |
| 29 |
>> index, do not use IUSE to filter USE since there is currently |
| 30 |
>> no guarantee that IUSE properly defines all of the necessary |
| 31 |
>> flags. |
| 32 |
> |
| 33 |
> These two changes now mean that without having "userland_GNU" in IUSE |
| 34 |
> |
| 35 |
> DEPEND="!userland_GNU? ( some/package )" |
| 36 |
> |
| 37 |
> will correctly end up in the Packages file, such that Portage will |
| 38 |
> properly calculate dependencies when reading from a binhost, right? |
| 39 |
|
| 40 |
Well, I consider my change to be a workaround for people behaving |
| 41 |
like Mike and refusing to declare certain conditionals in IUSE. The |
| 42 |
way that I see it, userland_GNU is a USE conditional, so it belongs |
| 43 |
in IUSE just like any other USE conditional. Maybe we would be |
| 44 |
better off if things like userland_GNU weren't in the USE |
| 45 |
conditional space, but they are. |
| 46 |
|
| 47 |
I don't understand why people refuse to declare certain conditionals |
| 48 |
in IUSE. IUSE conveys important information about which flags the |
| 49 |
package responds to. If we're not going to record them in IUSE |
| 50 |
explicitly, then every package, regardless of whether or not it uses |
| 51 |
those conditionals, will have to inherit them implicitly. If that's |
| 52 |
what we are going to do, then we should invent a way to declare in |
| 53 |
the profile which flags will behave that way (maybe we can just use |
| 54 |
USE_EXPAND_HIDDEN, but that's only designed for USE_EXPAND flags). |
| 55 |
|
| 56 |
Zac |
| 57 |
-----BEGIN PGP SIGNATURE----- |
| 58 |
Version: GnuPG v2.0.7 (GNU/Linux) |
| 59 |
|
| 60 |
iD8DBQFHLfk1/ejvha5XGaMRAqqIAKDiUJDeIXU9n+ikrRbk8rZNJ4Y0CwCbBSwH |
| 61 |
abNAp0Xoe8w5irwcZChgYEw= |
| 62 |
=PvYZ |
| 63 |
-----END PGP SIGNATURE----- |
| 64 |
-- |
| 65 |
gentoo-portage-dev@g.o mailing list |
Archives