1 |
Ciaran McCreesh wrote the following on 03.10.2006 14:26 : |
2 |
> On Tue, 03 Oct 2006 02:56:42 +0200 Lionel Bouton |
3 |
> <lionel-dev@××××××.name> wrote: |
4 |
> | Being able to tune the CFLAGS is part of one of the core principles of |
5 |
> | Gentoo: let the user be in control. |
6 |
> |
7 |
> What? No it isn't. |
8 |
> |
9 |
|
10 |
Maybe it depends on what you mean by 'in control'. What I mean is that |
11 |
you have a good stable base from which to work on, but nothing prevents |
12 |
you to tweak things like you want: Gentoo doesn't get in your way. |
13 |
http://www.gentoo.org/main/en/about.xml mentions "Extreme |
14 |
Configurabiliy" and the main picture states "Larry the Cow was in |
15 |
control. And he liked it.". |
16 |
|
17 |
> | <li>nss_ldap stopped working with <c>-ffast-math</c> (reported to |
18 |
> | break many packages changing with the actual gcc version)</li> |
19 |
> |
20 |
> Uh, -ffast-math has never been and will never be a safe thing to stick |
21 |
> in CFLAGS. |
22 |
> |
23 |
|
24 |
I agree (how could I say otherwise after spending several days with a |
25 |
hole in my foot finally finding that I had a gun named fast-math in my |
26 |
hand :-) ). |
27 |
Apparently many developpers think that it might be in CFLAGS though (see |
28 |
the amount of 'filter-flags -ffast-math' in ebuilds) so a reminder might |
29 |
not be wasted for some users. |
30 |
|
31 |
> |
32 |
> | Users with unsupported CFLAGS (see the <uri |
33 |
> | link='http://gentoo-wiki.com/CFLAGS_matrix'>CFLAGS matrix</uri> for |
34 |
> | example) might want to return to safe CFLAGS (see <uri |
35 |
> | link='http://gentoo-wiki.com/Safe_Cflags'>Safe CFLAGS</uri>) if recent |
36 |
> | updates caused them stability problems. On the other hand, more |
37 |
> | adventurous users might want to experiment with CFLAGS that didn't |
38 |
> | work properly with gcc-3.4.6... As always, the user is in control. |
39 |
> |
40 |
> Linking to that is a very bad idea. The wiki is in control of the |
41 |
> minority ricer fringe. |
42 |
> |
43 |
> |
44 |
|
45 |
Ok. Anyway I'm now convinced that a dev-proofed version of its content |
46 |
in the GWN would be far better. |
47 |
|
48 |
> GWN shouldn't be advocating this kind of thing at all. Here's a better |
49 |
> paragraph: |
50 |
> |
51 |
> <p>We would like to remind you that using anything beyond -O2 |
52 |
> -fomit-frame-pointer -march/-mcpu/-mtune in CFLAGS or CXXLFAGS (and |
53 |
> -mieee, -mabi etc on selected archs that tell you to do this), and using |
54 |
> anything at all in LDFLAGS or ASFLAGS, is pointless and will lead to a |
55 |
> broken system. Your penis length is not proportional to the size of your |
56 |
> CFLAGS.</p> |
57 |
> |
58 |
> |
59 |
|
60 |
Hum, I'll leave out the last sentence or rephrase it... I'd prefer to be |
61 |
more soft-spoken: pointless might be a little too much too. Let's say |
62 |
that the cost-risk/benefit ratio is not worth it for the vast majority |
63 |
of users. CFLAGS tuning should probably only be used by people with very |
64 |
specific needs (gcc devs/testers, HPTC people with extensive |
65 |
knowledge/experience of the problems involved). For LDFLAGS and ASFLAGS |
66 |
I'll take your word for it (I never even tried modifying them myself). |
67 |
|
68 |
Lionel. |
69 |
-- |
70 |
gentoo-dev@g.o mailing list |