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