1 |
On Monday 03 February 2014 14:11:00 Michael Haubenwallner wrote: |
2 |
> On 02/02/2014 09:12 PM, Jeremy Olexa wrote: |
3 |
> > On Sat, Feb 1, 2014 at 8:31 PM, Ruud Koolen <redlizard@g.o |
4 |
> > <mailto:redlizard@g.o>> wrote: |
5 |
> > |
6 |
> > Hi all, |
7 |
> > |
8 |
> > I have recently been doing some cleanup work to the bootstrap script, |
9 |
> > getting rid of cruft here and there. My goals in this have been twofold: |
10 |
> > part of the |
11 |
> |
12 |
> One more thing for gcc-4.2.4: Currently, 4.2.4 is the only one I made |
13 |
> working for Prefix on ppc-aix. The next one I plan to make working is some |
14 |
> gcc-4.8.x - eventually I get one more round tuit [1] this year. |
15 |
|
16 |
I should clarify: my bootstrap script does away with the need to *temporarily* use an older gcc version during bootstrap than is available for the platform in |
17 |
prefix. If gcc 4-2-4 is simply the only gcc available for $platform at all, even outside the bootstrap, then my bootstrap approach isn't going to improve things. |
18 |
|
19 |
> > Now, the hard part is that one person does not have access to many |
20 |
> > platforms. |
21 |
> |
22 |
> Exactly. |
23 |
> |
24 |
> However, my current setup might help to some degree: I do have a buildbot |
25 |
> running to bootstrap current Gentoo Prefix instances (32bits only) with |
26 |
> LATEST_TREE=yes once a week, with additional sync+update once a day (iirc). |
27 |
> |
28 |
> The buildbot slave machines currently operating are (gcc-version as |
29 |
> available on the login-shell): |
30 |
> |
31 |
> ppc AIX 7.1 gcc-4.6.1, last success on 2013, Dec 29 |
32 |
> ppc AIX 6.1 gcc-4.6.1, last success on 2014, Jan 8 |
33 |
> ppc AIX 5.3 gcc-3.3.6, currently broken, not investigated yet, maybe going to become obsolete for us |
34 |
> |
35 |
> sparc Sun Solaris 10 gcc-3.4.6, currently broken, not investigated yet, going to become obsolete for us |
36 |
> |
37 |
> ia64 HP-UX 11.31 gcc-4.2.3, currently broken, obsolete for us |
38 |
> hppa2.0w HP-UX 11.31 gcc-4.4.3, currently broken, obsolete for us |
39 |
|
40 |
It would be amazingly helpful if I could test on those machines; or, if that's not possible, if you could test things for me (though that will probably involve |
41 |
quite a few tedious "patch - test again - still broken" cycles). The other systems I can acquire on my own (though of course I certainly wouldn't complain if |
42 |
you just have them available). |
43 |
|
44 |
-- Ruud |