| 1 |
On Tuesday 26 June 2007 17:15:45 Boyd Stephen Smith Jr. wrote: |
| 2 |
> > '[gentoo-user] FEATURES="test" -- Should this work?': |
| 3 |
> > I recently enabled the "test" feature on Portage |
| 4 |
|
| 5 |
It varies quite a bit between packages and maintainers. If you are sure there |
| 6 |
doesn't exist a bug for those packages already then I think it's ok to file |
| 7 |
some bugs about it. Don't file too many at once but wait and see what happens |
| 8 |
with the first reports before filing more. I.e. unless of course you have a |
| 9 |
fix. :) |
| 10 |
|
| 11 |
> I use paludis,. which does testing by default, and I've seen a number of |
| 12 |
> packages fail tests. I simply mask those specific versions. |
| 13 |
|
| 14 |
So I guess you've masked packages such as binutils, autoconf and automake? Who |
| 15 |
needs those anyway? *grin* ;) |
| 16 |
|
| 17 |
> Last time I checked, bug reports coming from alternative package manager |
| 18 |
> users were closed "fairly quickly". |
| 19 |
|
| 20 |
It depends very much on who the maintainer is. Some maintainers do support |
| 21 |
Paludis users.. The official stance is, however, that filing bugs using |
| 22 |
Paludis is currently discouraged. |
| 23 |
|
| 24 |
> Since you are using portage, your bugs may get a little more attention, but |
| 25 |
> I think the general consensus is that "tests aren't important" so they |
| 26 |
> generally get dropped to low priority, ignored until the next release, then |
| 27 |
> closed with a "refile if it affects the current release" message. |
| 28 |
|
| 29 |
Test failures are much more important if they are new. I.e. if the previous |
| 30 |
release didn't fail... |
| 31 |
|
| 32 |
-- |
| 33 |
Bo Andresen |
Archives