| 1 |
On 10-03-2009 17:17:00 +0100, Michael Haubenwallner wrote: |
| 2 |
> On Tue, 2009-03-10 at 10:15 +0100, Fabian Groffen wrote: |
| 3 |
> > Currently we have eprefixify as function provided by Portage, and hence |
| 4 |
> > ebuilds that use eprefixify cannot be straight ported to gentoo-x86. |
| 5 |
> > Most notable are the eselect-* ebuilds that some gentoo-x86 devs like to |
| 6 |
> > give full Prefix support, but simply can't because eprefixify (which we |
| 7 |
> > would use) cannot be used in gentoo-x86. |
| 8 |
> |
| 9 |
> Maybe indeed you're right, eprefixify is just an utility to help fixing |
| 10 |
> package's code and should not be part of an EAPI. |
| 11 |
|
| 12 |
I think I'll draft up a proposal tonight. |
| 13 |
|
| 14 |
> > So, should we start an eclass with for now eprefixify in there, and for |
| 15 |
> > every ebuild where we use it, start inheriting prefix and nuke the |
| 16 |
> > function from Portage? Somehow I think this is a good idea. |
| 17 |
> |
| 18 |
> In gentoo-x86, this indeed can help to integrate prefix ebuilds, even |
| 19 |
> with EAPI 0. Just set EPREFIX/ED/EROOT if unset, in global scope. |
| 20 |
|
| 21 |
I had not thought about that one, actually. That's really cool. Sort |
| 22 |
of makes our ebuilds backwards compatible... |
| 23 |
|
| 24 |
> But the keyword issue still stands... |
| 25 |
|
| 26 |
True, but that's just politics. |
| 27 |
|
| 28 |
|
| 29 |
-- |
| 30 |
Fabian Groffen |
| 31 |
Gentoo on a different level |
Archives