| 1 |
On Tue, 3 Mar 2015 22:12:07 +0100 |
| 2 |
Ulrich Mueller <ulm@g.o> wrote: |
| 3 |
|
| 4 |
> Hi, |
| 5 |
> We currently have two mechanisms to influence building of packages |
| 6 |
> that cannot be binary redistributed. Some 150 packages have |
| 7 |
> RESTRICT=bindist (unconditional, or USE conditional), whereas |
| 8 |
> 27 packages have the bindist flag in their IUSE. Until very recently, |
| 9 |
> the two sets were practically disjoint, with only one single package |
| 10 |
> (www-client/chromium) in both sets. |
| 11 |
> |
| 12 |
> As a consequence, neither building with the well-known USE=bindist nor |
| 13 |
> with ACCEPT_RESTRICT="* -bindist" (presumably less known and Portage |
| 14 |
> only?) will guarantee that only redistributable binaries will be |
| 15 |
> built [1]. |
| 16 |
> |
| 17 |
> So a few days ago I filed bug 541408 and dependent bugs, with the goal |
| 18 |
> that ebuilds with the bindist flag in their IUSE should also add |
| 19 |
> RESTRICT="!bindist? ( bindist )". However, it turned out that some 10 |
| 20 |
> packages are using the bindist flag only in REQUIRED_USE but nowhere |
| 21 |
> else [2] (with foo being some feature flag): |
| 22 |
> |
| 23 |
> IUSE="bindist foo" |
| 24 |
> REQUIRED_USE="bindist? ( !foo )" |
| 25 |
> |
| 26 |
> IMHO, the bindist USE flag is redundant in these cases. So we should |
| 27 |
> get rid of the REQUIRED_USE and add a restriction instead: |
| 28 |
> |
| 29 |
> IUSE="foo" |
| 30 |
> RESTRICT="foo? ( bindist )" |
| 31 |
> |
| 32 |
> What do you think? Should we proceed in this direction? |
| 33 |
> |
| 34 |
|
| 35 |
obviously, +1 for this |
| 36 |
|
| 37 |
|
| 38 |
from the rest of your email, I understand that RESTRICT=bindist is |
| 39 |
eapi0 (or even not covered by eapi?), right? |
| 40 |
|
| 41 |
is there any information on RESTRICT=bindist ? with these 150 packages |
| 42 |
using it, this sounds like something well known that I had never heard |
| 43 |
of :/ |
| 44 |
|
| 45 |
|
| 46 |
Alexis. |
Archives