1 |
I have me a bug, but before I finish weighing in with my own pros and cons, I |
2 |
thought I'd ask -dev, in particular since I know the creators of things like |
3 |
catalyst and such are on here, and the results of this particular bug |
4 |
(feature request) might be of interest to them. |
5 |
|
6 |
A brave (perhaps bored, perhaps just frustrated) user has come up with an |
7 |
ebuild ( http://bugs.gentoo.org/show_bug.cgi?id=95038 ) that should (he says |
8 |
it does, but I haven't had a chance to set up a chroot to test it in) build |
9 |
an extremely minimal perl that's capable of some core perl functionality and |
10 |
litte else out of the box. Now for the perl herd, this would imply that we |
11 |
would need to come up with some more perl-core/* ebuilds to fill in any gaps |
12 |
that the minimal perl install has. |
13 |
|
14 |
Cons are the addition of a bunch of ebuilds to cover the functionality that |
15 |
the minimal install lacks. |
16 |
|
17 |
Pros, though, are pretty big, once getting past the first hurtle of pain. We |
18 |
would be able to eliminate the conflicts between perl coming with module foo |
19 |
and ebuild for foo, which is usually at or greater than the same version as |
20 |
the one that came with perl (usually greater) A lot of our collision protect |
21 |
bugs would dissipate rapidly I think. It would save space (with his use |
22 |
flags/architecture, perl went from 12,300K to 930K) on livecd's. I have a |
23 |
weak unfounded suspicion that it might help with the whole openssl->perl loop |
24 |
that expresses itself when some folks are building a stage1. |
25 |
|
26 |
I'm completely on the fence here, so any feedback welcome :) |
27 |
|
28 |
-- |
29 |
|
30 |
-----o()o--------------------------------------------- |
31 |
Michael Cummings | #gentoo-dev, #gentoo-perl |
32 |
Gentoo Perl Dev | on irc.freenode.net |
33 |
-----o()o--------------------------------------------- |