1 |
Michal, not opposed to splitting the patch into three parts. |
2 |
|
3 |
I'd rather use the env/ mechanism instead of the package.env one as it is |
4 |
more flexible. |
5 |
It also feels better as ebuild.sh will walk the profiles to source the |
6 |
bashrc script so a bashrc from a |
7 |
"low priority" profile may override a package.env definition from a high |
8 |
priority profile. |
9 |
|
10 |
I'll remove the with spaces and the Change-Id (gerrit specific tag I |
11 |
believe) |
12 |
|
13 |
Bertrand |
14 |
|
15 |
On Wed, Sep 17, 2014 at 2:43 PM, Zac Medico <zmedico@g.o> wrote: |
16 |
|
17 |
> On 09/17/2014 02:28 PM, Michał Górny wrote: |
18 |
> >> diff --git a/pym/portage/repository/config.py |
19 |
> b/pym/portage/repository/config.py |
20 |
> >> index 5e0d055..ef8054e 100644 |
21 |
> >> --- a/pym/portage/repository/config.py |
22 |
> >> +++ b/pym/portage/repository/config.py |
23 |
> >> @@ -40,7 +40,7 @@ if sys.hexversion >= 0x3000000: |
24 |
> >> _invalid_path_char_re = re.compile(r'[^a-zA-Z0-9._\-+:/]') |
25 |
> >> |
26 |
> >> _valid_profile_formats = frozenset( |
27 |
> >> - ['pms', 'portage-1', 'portage-2']) |
28 |
> >> + ['pms', 'portage-1', 'portage-2', 'profile-env']) |
29 |
> > |
30 |
> > I'm not sure if dedicated profile format is the best way to go. If I |
31 |
> > understand your patch correctly, this means that 'profile-env' has only |
32 |
> > features of PMS profile but not of 'portage-*' formats. This could mean |
33 |
> > that some people would have to choose between features of one or |
34 |
> > the other. Since both are compatible, i suggest 'portage-3' instead. |
35 |
> |
36 |
> The profile-formats field is designed to allow mixing of formats. So, |
37 |
> it's legal to mix profile-env with any of the other formats. |
38 |
> -- |
39 |
> Thanks, |
40 |
> Zac |
41 |
> |
42 |
> |