Re: [gentoo-user] Re: Gentoo install script

[Michael Mol <mikemol@gmail.com>]

On Wed, Jul 4, 2012 at 12:53 PM, Mark Knecht <markknecht@gmail.com> wrote:
> On Sun, Jul 1, 2012 at 5:28 PM, Michael Mol <mikemol@gmail.com> wrote:
>> On Wed, Jun 27, 2012 at 10:13 PM, Michael Mol <mikemol@gmail.com> wrote:
>>> Very rough, and very much a works-for-me thing, but I thought I'd share.
>>>
>>> https://github.com/mikemol/gentoo-install
>>>
>>> I wrote it to ease the pain of the "install-configure-build" cycle I
>>> was going through to figure out what was breaking glibc.
>>
>> Just a bit of a followup. I've got most of the bugs worked out, and
>> I'm very pleased with it. I've used it to get through most of the
>> install sequence for inara, and it's currently on package 113/158 of
>> its second pass of 'emerge -e @world'.
>>
>> If anyone else gets around to trying it, let me know. :)
>>
>> --
>> :wq
>>
>
> Hi Michael,
>    Looks interesting. From reading the code it looks like this was a
> (proper) reaction to rebuilding the two machines you recently had
> trouble with, right?

Indeed. Inara and kaylee are still incomplete, but the script at least
helped me get to a working, up-to-date chroot environment, and that's
a major improvement over where I was.

>
>    Not sure when I'll get to it but I'll likely give it a try building
> a VM in Virtualbox as a test.

I can tell you right now that you'll hit a circular dependency problem
toward the end, where it starts installing what I'd call "comfort
tools." The general(ish) solution is probably going to be a two-pass
emerge. One with USE="-gtk", followed by a one-two sequence without
that. It's a typical gtk->cups->avahi cyclic dependency, and since gtk
is the most-core element in the cycle, I figure it probably it'll be
part of most cyclic dependency problems that might crop up in the
future, so suppressing it likely covers the most ground. (By the same
reasoning, it's probably appropriate to include -qt, -gnome and -kde
in that first pass, too.)

>
>    Thanks for doing the heavy lifting and posting the work output.

Np. It beat the *pants* off of manually trying different CFLAGS to
figure out why glibc was puking. I only had to go through the manual
steps twice before I got sick of it and wrote the bulk of the script
in a couple hours. The rest has mostly been letting the thing run,
discover where I missed something, modify the script, and let it run
again.

-- 
:wq