1 |
Kevin F. Quinn wrote: |
2 |
> Mike Kelly <pioto@×××××.org> wrote: |
3 |
> |
4 |
>> Well, the main benefit I get by not using an eclass hook is that I can |
5 |
>> properly clean up after myself if a build fails, and, I think, better |
6 |
>> handle the addition of users in the case where the build and install |
7 |
>> machines are different. |
8 |
> |
9 |
> If the main issue is cleaning up after an abort, perhaps it would be |
10 |
> useful to add a pkg_abort() phase to the package manager; then that can |
11 |
> be implemented in an eclass. |
12 |
|
13 |
I'd also be interested in a pkg_abort() phase, or another way to script actions |
14 |
triggered by a compile failure. If it could tell the difference between build |
15 |
error and user abort, even better. And if it could pass the phase it was in |
16 |
when the abort occurred to the eclass, that would rock. |
17 |
|
18 |
--de. |
19 |
|
20 |
-- |
21 |
gentoo-portage-dev@g.o mailing list |