1 |
On Fri, 2018-09-14 at 20:22 +0300, Alon Bar-Lev wrote: |
2 |
> On Fri, Sep 14, 2018 at 8:16 PM Richard Yao <ryao@g.o> wrote: |
3 |
> > |
4 |
> > On 09/14/2018 12:40 PM, Alon Bar-Lev wrote: |
5 |
> > > On Fri, Sep 14, 2018 at 12:34 AM Sergei Trofimovich <slyfox@g.o> wrote: |
6 |
> > > > |
7 |
> > > > On Tue, 11 Sep 2018 12:44:38 +0300 |
8 |
> > > > Alon Bar-Lev <alonbl@g.o> wrote: |
9 |
> > > > |
10 |
> > > > I'm personally in favour of not allowing -Werror |
11 |
> > > > to be in build system unconditionally. |
12 |
> > > > |
13 |
> > > > Maintainer is free to implement --enable-werror any way |
14 |
> > > > it's convenient to run on their own extended sanity checks |
15 |
> > > > and optionally expose it to users. Be it USE flag or |
16 |
> > > > EXTRA_ECONF option. |
17 |
> > > |
18 |
> > > This discussion is not for downstream to have a more strict policy |
19 |
> > > than upstream. People try to hijack discussion and introduce noise to |
20 |
> > > de-focus the discussion. |
21 |
> > > |
22 |
> > > Downstream policy cannot be more strict than upstream as then every |
23 |
> > > change upstream is doing downstream need to rebase and invest in even |
24 |
> > > more changes. |
25 |
> > > |
26 |
> > > This discussion is to follow upstream strict policy if upstream proves |
27 |
> > > that it stands behind it and downstream is willing to follow. |
28 |
> > |
29 |
> > I don't think we should do that unless we provide a USE flag for users |
30 |
> > to opt into the behavior. Forcing it on users is problematic for the |
31 |
> > reasons others stated. However, letting them opt into the behavior is |
32 |
> > reasonable. |
33 |
> > |
34 |
> > In the case of sys-fs/zfs, enabling -Werror (which includes -Wall) on |
35 |
> > USE=debug is following upstream's wishes to build debug builds with -Werror. |
36 |
> |
37 |
> Let's do this the other way around and be react based on facts and not |
38 |
> speculations. |
39 |
> Let's change the policy for a year for selected packages as I |
40 |
> outlined, monitor bugs and after a year see response times, affected |
41 |
> users and if downstream patches are accumulated. Then we can decide if |
42 |
> we need to patch upstream packages. |
43 |
> If we need to patch upstream package anyway, not follow upstream |
44 |
> policy and not accepting input for various of permutations and |
45 |
> architecture from all users, this discussion is nearly void. |
46 |
> |
47 |
|
48 |
...and for how long did you exactly ignore the standing policy that |
49 |
suddenly we need a new testing period? How about we do the opposite |
50 |
and you prove a *single* bug found downstream using this method so far? |
51 |
|
52 |
Because so far this discussion is not much different than "let's make |
53 |
the ebuild fail for some values of ${RANDOM}, and add extra values when |
54 |
users complain". Though the variant with random has probably a greater |
55 |
chance of failing when *actual* security issues happen. |
56 |
|
57 |
-- |
58 |
Best regards, |
59 |
Michał Górny |