| 1 |
On Sat, Oct 08, 2005 at 03:45:18PM -0700, Rob Cakebread wrote: |
| 2 |
> >How should Portage handle this? |
| 3 |
> I've been working on an eclass for these but its not ready. |
| 4 |
> I was hoping it'd be as easy as the Ruby gems eclass, because |
| 5 |
> they are similar, but eggs don't play nicely in a sandbox yet. |
| 6 |
|
| 7 |
But aren't eggs a bit against the Gentoo philosophy? I mean there are |
| 8 |
some eggs that contain precompiled C-extensions. Shouldn't it still be |
| 9 |
source builds that just somehow work with setuptools? |
| 10 |
|
| 11 |
> You're probably best off using the straight source for TurboGears |
| 12 |
> and its dependencies if you want to install via ebuild for now. |
| 13 |
|
| 14 |
The problem is that even straight source does not play nicely in |
| 15 |
ebuilds. Take SQLObject 0.7rc1, when you use the distutils eclass, it |
| 16 |
will still download setuptools as an egg and use that to do all |
| 17 |
dependency checks and resolutions. For SQLObject you can just delete the |
| 18 |
ez_setup dir after unpacking it, then it will use ordinary distutils |
| 19 |
with no dependency resolution, but with some packages such as |
| 20 |
FormEncode, you HAVE to use setuptools, or do everything manually in the |
| 21 |
ebuild. |
| 22 |
|
| 23 |
> Mail me off-list or join #gentoo-python if you're interested |
| 24 |
> in working on the easy_install eclass, please. |
| 25 |
|
| 26 |
Should we move this off-list? |
| 27 |
|
| 28 |
-- |
| 29 |
Anders |
| 30 |
-----BEGIN GEEK CODE BLOCK----- |
| 31 |
Version: 3.12 |
| 32 |
GCS/O d--@ s:+ a-- C++ UL+++$ P++ L+++ E- W+ N(+) o K? w O-- M- V |
| 33 |
PS+ PE@ Y+ PGP+ t 5 X R+ tv+ b++ DI+++ D+ G e- h !r y? |
| 34 |
------END GEEK CODE BLOCK------ |
| 35 |
PGPKey: http://random.sks.keyserver.penguin.de:11371/pks/lookup?op=get&search=0xD4DEFED0 |
| 36 |
-- |
| 37 |
gentoo-dev@g.o mailing list |
Archives