| 1 |
Yes, a generic INSTALL_PROTECT sounds good. I'd vote for that :) |
| 2 |
It should be used at the final merge stage and not before. |
| 3 |
|
| 4 |
For performance reasons it could be also used at the untar stage |
| 5 |
when using .tbz2 packages (with --exclude=FILE). |
| 6 |
|
| 7 |
On Fri, Oct 24, 2003 at 01:08:13AM +0200, Spider wrote: |
| 8 |
> On Thu, 23 Oct 2003 23:34:10 +0200 |
| 9 |
> bkhl@××××××××××××.se (Björn Lindström) wrote: |
| 10 |
> |
| 11 |
> |
| 12 |
> > And why couldn't those lines check the INSTALL_PROTECT (if that's what |
| 13 |
> > we'll call it) variable? Including the non-installed files in binary |
| 14 |
> > packages would work doesn't really have anything to do with whether |
| 15 |
> > this is implemented as a FEATURE flag or a separate variable. |
| 16 |
> > |
| 17 |
> |
| 18 |
> True, and then we could do some interesting default masking, like |
| 19 |
> walling off /usr/local per default... |
| 20 |
> |
| 21 |
|
| 22 |
-- |
| 23 |
gentoo-dev@g.o mailing list |
Archives