| 1 |
On Mon, Jun 08, 2009 at 07:42:17PM -0400, Mike Frysinger wrote: |
| 2 |
> > One of the reasons to move stuff OUT of /lib are all the profiles where |
| 3 |
> > SYMLINK_LIB is disabled AND LIBDIR_${arch} != "lib". On non-multilib |
| 4 |
> > systems (so there is no lib23/64) or multilib systems where /lib is the |
| 5 |
> > correct location, then any test against /lib/rc/version would be fine. |
| 6 |
> > On anything else, it's not. |
| 7 |
> > |
| 8 |
> > Having it in a different location from upstream (OpenRC), means that any |
| 9 |
> > other distributions using OpenRC's /libexec/rc/version location would |
| 10 |
> > need to patch all their init.d scripts. |
| 11 |
> the proposed /sbin/functions.sh check would makes this issue moot |
| 12 |
And why the hell didn't you come forward on bug 270646 with this? |
| 13 |
|
| 14 |
I'm concerned that absence tests like that will not be as useful as |
| 15 |
OpenRC spreads. Sure in Gentoo, baselayout1 provides /sbin/functions.sh, |
| 16 |
but other users won't. |
| 17 |
|
| 18 |
Since /libexec and /lib are out for the discussed reasons, this leaves |
| 19 |
us coming back to /etc/openrc-version, with the same file in |
| 20 |
CONFIG_PROTECT_MASK. |
| 21 |
|
| 22 |
-- |
| 23 |
Robin Hugh Johnson |
| 24 |
Gentoo Linux Developer & Infra Guy |
| 25 |
E-Mail : robbat2@g.o |
| 26 |
GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 |
Archives