| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA1 |
| 3 |
|
| 4 |
On Monday 22 March 2004 23:37, Paul de Vrieze wrote: |
| 5 |
> On Wednesday 17 March 2004 01:31, Olivier Crête wrote: |
| 6 |
> > If we want to support a very large number of hardware/software |
| 7 |
> > combinations we have to either decide that we want to keep the same |
| 8 |
> |
| 9 |
> I don't see this as much as a problem. Supporting many hardware software |
| 10 |
> combinations is hard in general. Even a very widely supported package |
| 11 |
> like gcc does not support more than two dozen configurations. Each |
| 12 |
> package must be checked for every configuration any way. What might be |
| 13 |
> an option is to have some level of hinting. Say have the x86-linux flag |
| 14 |
> mean that it will probably work on x86 machines and that it will |
| 15 |
> probably work on linux machines. Adding something like :sunos: (a hint |
| 16 |
> between colons) would then automatically mark this package as testing |
| 17 |
> under solaris-x86. If that package then builds on sparc-linux it is also |
| 18 |
> likely to work on sparc-solaris etc. It would allow testing to be more |
| 19 |
> flexible, while still having the same requirements for stability. |
| 20 |
|
| 21 |
I won't say much or I'll get myself in trouble again, but I think you've hit |
| 22 |
it on the head. If this could be tied in with profiles and their parents, I |
| 23 |
think you couldn't really get a better solution. |
| 24 |
|
| 25 |
+1 |
| 26 |
|
| 27 |
Regards, |
| 28 |
Jason Stubbs |
| 29 |
-----BEGIN PGP SIGNATURE----- |
| 30 |
Version: GnuPG v1.2.4 (GNU/Linux) |
| 31 |
|
| 32 |
iQCVAwUBQF8A6FoikN4/5jfsAQI0RAP9HPbp29gzj1ld4lgqVU2mS4ZtB+m5xoeU |
| 33 |
JMR0Ma1jkvcCbTGYcO/5ZpyoBsmjkwOZ9cI+QUfwzdG1lW6PTg3mOOiJVreSh1lF |
| 34 |
hbWT6oMLr6fBt4S6ZVZIapLhZCeJzqnmGZ1UKEiaRlgcHvzDz+1E7H7CBs9GYGDx |
| 35 |
Y0JAU92H78c= |
| 36 |
=JFQ7 |
| 37 |
-----END PGP SIGNATURE----- |
| 38 |
|
| 39 |
-- |
| 40 |
gentoo-dev@g.o mailing list |
Archives