1 |
On Fri, 6 Nov 2015 18:35:54 -0500 |
2 |
Mike Gilbert <floppym@g.o> wrote: |
3 |
|
4 |
> On Fri, Oct 30, 2015 at 1:39 AM, Michał Górny <mgorny@g.o> |
5 |
> wrote: |
6 |
> > On Thu, 29 Oct 2015 21:06:33 -0700 |
7 |
> > Brian Dolbec <dolsen@g.o> wrote: |
8 |
> > |
9 |
> >> On Thu, 29 Oct 2015 17:37:26 -0400 |
10 |
> >> Mike Frysinger <vapier@g.o> wrote: |
11 |
> >> |
12 |
> >> > On 22 Oct 2015 12:54, Paul Varner wrote: |
13 |
> >> > > Mike, I know you're busy with other stuff, but if you ever |
14 |
> >> > > want to see a new gentoolkit/gentoolkit-dev release, consider |
15 |
> >> > > this your authorization to just do it. The README.dev files |
16 |
> >> > > state how to make releases. |
17 |
> >> > |
18 |
> >> > thanks, i think this will help a lot |
19 |
> >> > |
20 |
> >> > > Since, the tools have dwindled down in gentoolkit-dev, I do |
21 |
> >> > > think it does make sense to keep it in the same repo and merge |
22 |
> >> > > the packages together behind a USE flag. I will revert the |
23 |
> >> > > commit, that emptied the genttolkit-dev branch and ask mgorny |
24 |
> >> > > to nuke the new gentoolkit-dev repository. |
25 |
> >> > > |
26 |
> >> > > As I get time, I will work towards moving the gentoolkit-dev |
27 |
> >> > > tools into gentoolkit and putting them behind a USE flag in |
28 |
> >> > > the ebuild. |
29 |
> >> > |
30 |
> >> > i'm no distutils expert, and every time i try to do something |
31 |
> >> > "fancy", i get frustrated by the module :). do people know of |
32 |
> >> > examples where you can do optional installs with a flag ? a |
33 |
> >> > cookbook sort of entry here would help and i could take care of |
34 |
> >> > merging in say ekeyword. -mike |
35 |
> >> |
36 |
> >> Have a look at layman's setup.py. It parses IUSE to set the |
37 |
> >> installed files via setup.py. It may not be the best method, but |
38 |
> >> it does work. |
39 |
> >> |
40 |
> >> The layman ebuild sets deps acording to the ISUE flags and setup.py |
41 |
> >> sets the installed modules on the python side. |
42 |
> > |
43 |
> > Sorry, what?! That's a huge QA violation. There is *NO* guarantee |
44 |
> > that USE will be exported. In fact, it is only exported because of |
45 |
> > poor design inside Portage that could result in the variable |
46 |
> > getting lost otherwise. |
47 |
> |
48 |
> PMS says that USE will be exported to the environment. So, this should |
49 |
> not break with any PMS-compliant package manager. |
50 |
> |
51 |
> Personally, it makes me cringe, but I suppose it doesn't make a lot of |
52 |
> sense to invent an entirely new envvar for it. |
53 |
> |
54 |
|
55 |
I remember discussing this with you at the time. setup.py is capable |
56 |
of taking args for this, but the distutils eclass and or distutils |
57 |
itself was not capable of doing that from an ebuild. So, this was my |
58 |
only option at the time. If that has changed, it'll be news to me :) |
59 |
|
60 |
|
61 |
Make it capable to pass options like automake and I'll be glad to change |
62 |
it. |
63 |
|
64 |
|
65 |
-- |
66 |
Brian Dolbec <dolsen> |