| 1 |
i had that problem; grobian helped me fix it, its easy. i changed |
| 2 |
my uid to match my gentoo linux file server for nfs mounts. look at: |
| 3 |
|
| 4 |
etc/make.globals:PORTAGE_INST_UID |
| 5 |
usr/lib/portage/pym/portage/const_autotool.py:rootuid |
| 6 |
|
| 7 |
matt |
| 8 |
|
| 9 |
On Tue, 24 Jun 2008, Peter Abrahamsen wrote: |
| 10 |
|
| 11 |
> Hi all, |
| 12 |
> |
| 13 |
> We have a prefixed Gentoo disk image here at our office that we |
| 14 |
> distribute to developers on OS X boxes so they can run full |
| 15 |
> development environments locally. It has postgresql, memcached, and |
| 16 |
> other goodies pre-installed with the right USE flags. |
| 17 |
> |
| 18 |
> Presently, the images are pretty much fixed once they're distributed |
| 19 |
> -- any updates have to come from the person who made the image. It |
| 20 |
> seems that their username is hard-coded in some scripts and config |
| 21 |
> files here and there, and emerge is unhappy if run by a different |
| 22 |
> user. I've changed those entries manually before, but I'm wondering if |
| 23 |
> it's really necessary that the username be hard-coded anywhere? Would |
| 24 |
> it be easy to make the code user-agnostic? |
| 25 |
> |
| 26 |
> It would be even more awesome to keep installs agnostic with respect |
| 27 |
> to where they are mounted, but I'm guessing this would be a much |
| 28 |
> larger undertaking, requiring more fundamental changes to portage. |
| 29 |
> |
| 30 |
> Neither of these issues is a huge deal, as we're happy just |
| 31 |
> distributing new images from time to time and not making our devs deal |
| 32 |
> with system management, but there've been several times when I've |
| 33 |
> wanted these flexibilities from portage. |
| 34 |
> |
| 35 |
> Thanks, |
| 36 |
> Peter |
| 37 |
> -- |
| 38 |
> gentoo-alt@l.g.o mailing list |
| 39 |
> |
| 40 |
> |
| 41 |
-- |
| 42 |
gentoo-alt@l.g.o mailing list |
Archives