1 |
Markos Chandras <hwoarang@g.o> said: |
2 |
> On Tue, Aug 10, 2010 at 06:31:52PM -0400, Mike Frysinger wrote: |
3 |
> > On Tue, Aug 10, 2010 at 5:53 PM, Markos Chandras wrote: |
4 |
> > >> It seems like few of our fellow developers don't know how to track |
5 |
> > >> down |
6 |
> > >> packages that don't respect LDFLAGS. Adding -Wl,--hash-style=gnu |
7 |
> > >> is a good way |
8 |
> > >> to do that. I would like to see this linker flag enabled by |
9 |
> > >> default on LDFLAGS |
10 |
> > >> (or at least for the dev/ profiles for now). Do you agree? |
11 |
> > > |
12 |
> > > I would really really *really* appreciated if our beloved arch |
13 |
> > > testers ( at least for linux amd64/x86 because they are the first |
14 |
> > > who stabilize a package ) make this default on their build boxes. |
15 |
> > |
16 |
> > sounds like someone needs to update/extend the arch testing |
17 |
> > documentation. random e-mails posted to random dev lists are quickly |
18 |
> > forgotten. new arch testers however should be reading the arch |
19 |
> > tester documnt. |
20 |
> |
21 |
> I will update the guide for amd64 HT and I will strongly advice the |
22 |
> rest of the arches to do that as well. Using my QA powerzzz I will be |
23 |
> quite strict from now on with arches making such stabilizations. |
24 |
|
25 |
i agree on this. |
26 |
packages on which portage complains about stuff like |
27 |
|
28 |
dohtml: bla file not found |
29 |
|
30 |
should not be marked stable. arch testers should not let stuff like this |
31 |
pass by. of course, neither should developers. |
32 |
|
33 |
but then again, we need better documentation of all of this. |
34 |
|
35 |
lyckily, the wiki effort has been killed off by a recent cabal</sarcasm> |
36 |
|
37 |
kind regards |
38 |
Thilo |
39 |
|
40 |
> |
41 |
> > > It is annoying to mark a package stable when it has *clear* QA |
42 |
> > > problems. |
43 |
> > |
44 |
> > please dont blow this out of proportion. two points: |
45 |
> > - stabilizing newer versions of a package when there is no QA |
46 |
> > |
47 |
> > regression is fine. |
48 |
> |
49 |
> Fair enough, still those QA need fixing. The fact that these QA probs |
50 |
> are not regressions doesn't mean it is ok to ignore them |
51 |
> |
52 |
> > - ignoring LDFLAGS, while incorrect, is rarely going to lead to |
53 |
> > |
54 |
> > broken packages being emerged on end users' systems. ignoring |
55 |
> > CFLAGS/CXXFLAGS however is much more likely to result in problems for |
56 |
> > end users when working with multilib or cross builds. |
57 |
> > -mike |
58 |
> |
59 |
> Of course. Respecting any *FLAGS is vital and definitely ony of the |
60 |
> many reasons we use Gentoo. |