1 |
On Wed, 08 May 2019 12:01:21 +0200 |
2 |
Michał Górny <mgorny@g.o> wrote: |
3 |
|
4 |
> On Wed, 2019-05-08 at 11:54 +0200, Alexis Ballier wrote: |
5 |
> > On Wed, 08 May 2019 11:41:41 +0200 |
6 |
> > Michał Górny <mgorny@g.o> wrote: |
7 |
> > |
8 |
> > > > There's multilib that adds a lot of flags with a single eclass |
9 |
> > > > change, but I'd guess the number of packages and flags is |
10 |
> > > > constantly growing, so sooner or later you'll be hit by this |
11 |
> > > > again and no multilib killing will help you then. |
12 |
> > > > |
13 |
> > > > I think it is more future proof to use the addition of multilib |
14 |
> > > > flags to fix pkgcheck rather than actively reducing the number |
15 |
> > > > of multilib flags to cope with its limitations. |
16 |
> > > |
17 |
> > > Then please do it, by all means. The reality is simple. If the |
18 |
> > > tool is broken, you either fix it or stop doing what you know |
19 |
> > > that breaks it. Being unable to do the former, and having no good |
20 |
> > > replacement, I'd go for the latter. |
21 |
> > |
22 |
> > Well, why is it slow ? IO ? CPU ? Did you collect profiling data ? |
23 |
> |
24 |
> CPU definitely. More detail than that, I don't and I don't have time |
25 |
> to investigate. |
26 |
|
27 |
So you don't have time to change 3 lines to add cProfile but do have |
28 |
time to send various emails and rework the entire multilib system ? |
29 |
weird. |
30 |
|
31 |
|
32 |
> > > > Also, remember that multilib is not entirely about skype or |
33 |
> > > > slack, this was made with multibin in mind too: for example an |
34 |
> > > > ABI may perform better than another one on specific workflows |
35 |
> > > > (x32) and it may make sense to use this abi for a specific |
36 |
> > > > binary (which would be manually built for now). |
37 |
> > > |
38 |
> > > No, it weren't. Just because someone found it convenient to use |
39 |
> > > the design beyond its original purpose doesn't mean it had a |
40 |
> > > different purpose. Besides, how many 'multibin' packages do we |
41 |
> > > have right now? |
42 |
> > |
43 |
> > It was, because portage multilib was doing this and claimed to have |
44 |
> > a use for this. |
45 |
> |
46 |
> I am not talking of 'portage multilib'. I am talking of |
47 |
> multilib-build implementation. If you are only concerned about the |
48 |
> former, we can surely drop those abi_* flags that are irrelevant to |
49 |
> it, can't we? |
50 |
|
51 |
I am also talking about multilib-build and its way of doing cleanly and |
52 |
replacing portage multilib. |
53 |
|
54 |
|
55 |
> > Its original purpose was generic enough to allow multibin |
56 |
> > to be added easily. The number of multibin packages is only |
57 |
> > relevant to show that this is not important enough for someone |
58 |
> > to step up and do it: Everybody can easily build anything themselves |
59 |
> > with the current implementation. |
60 |
> |
61 |
> Exactly. And that's why I believe having fast CI is more important |
62 |
> than DIY multibin. If someone can 'easily build anything |
63 |
> themselves', they can also add needed flags locally. |
64 |
|
65 |
|
66 |
It's a "little" more complex than just adding one flag to an eclass and |
67 |
cannot really be done with an overlay; in the end, this requires to fork |
68 |
gentoo.git more or less. |