1 |
> On 4 Jan 2022, at 21:54, Michał Górny <mgorny@g.o> wrote: |
2 |
> |
3 |
> On Tue, 2022-01-04 at 11:17 -0500, Mike Gilbert wrote: |
4 |
>> On Tue, Jan 4, 2022 at 5:23 AM Sam James <sam@g.o> wrote: |
5 |
>>> |
6 |
>>> It's already an implicit IUSE, so we don't need this. |
7 |
>> |
8 |
>> I think it is better to declare it explicitly rather than relying on |
9 |
>> the IUSE_IMPLICIT setting in profiles. |
10 |
> |
11 |
> I agree. Perhaps we should go even further and remove them from |
12 |
> implicit flags. |
13 |
> |
14 |
|
15 |
(Note: I did push this change already to coalesce this |
16 |
with FreeBSD and other removals.) |
17 |
|
18 |
I don't think declaring it explicitly is helpful given |
19 |
in most cases (and even more now after recent |
20 |
cleanups), it's relied upon without declaring it anyway. |
21 |
|
22 |
Use within IUSE while it's implicit also means |
23 |
IUSE seems to regularly get out of sync with |
24 |
actual use within the ebuilds. |
25 |
|
26 |
As for removing implicit use entirely: while you might |
27 |
argue this could be okay for the Linux kernel, it would cause |
28 |
unnecessary rebuilds indefinitely whenever we add a new dep |
29 |
to an ebuild for e.g. Darwin, or if we did it for libcs, |
30 |
for say, musl/non-glibc. |
31 |
|
32 |
best, |
33 |
sam |