Gentoo Archives: gentoo-dev

From: Michael Mol <mikemol@×××××.com>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] Portage FEATURE suggestion - limited-visibility builds
Date: Tue, 31 Jul 2012 19:25:35
Message-Id: CA+czFiDkdLsGgjBxnGRu6QEj==nTDPzi6hVvwUXFtx30DfhYTA@mail.gmail.com
In Reply to: Re: [gentoo-dev] Portage FEATURE suggestion - limited-visibility builds by Ian Stakenvicius
1 On Tue, Jul 31, 2012 at 10:56 AM, Ian Stakenvicius <axs@g.o> wrote:
2 > -----BEGIN PGP SIGNED MESSAGE-----
3 > Hash: SHA256
4 >
5 > On 31/07/12 10:55 AM, Michael Mol wrote:
6 >> On Tue, Jul 31, 2012 at 10:48 AM, "Paweł Hajdan, Jr."
7 >> <phajdan.jr@g.o> wrote:
8 >>> On 7/26/12 8:26 PM, Rich Freeman wrote:
9 >>>> I've been messing around with namespaces and some of what
10 >>>> systemd has been doing with them, and I have an idea for a
11 >>>> portage feature.
12 >>>>
13 >>>> But before doing a brain dump of ideas, how useful would it be
14 >>>> to have a FEATURE for portage to do a limited-visibility build?
15 >>>> That is, the build would be run in an environment where the
16 >>>> root filesystem appears to contain everything in a DEPEND
17 >>>> (including @system currently) and nothing else?
18 >>>
19 >>> I was thinking about something similar too. In my opinion it's a
20 >>> great feature. If/when there are any bugs to get this
21 >>> implemented, please let me know.
22 >>>
23 >>> A possible alternative implementation would be to make the
24 >>> sandbox deny access to anything outside DEPEND. One totally crazy
25 >>> idea to make that fast are extended attributes (portage would
26 >>> record which package a file belongs to when merging the file).
27 >>> Another possible solution is using a cache.
28 >>
29 >> We already have the ability to run commands like 'equery b
30 >> $somefile' to map a file back to a package, so the data for a
31 >> filesystem helper should already be available in whatever database
32 >> equery is using.
33 >>
34 >
35 > Although that is true, it would be -WAY- too slow to generate said
36 > list via equery/q* helpers; I think that's where the
37 > extended-attributes and/or cache idea comes into play.
38
39 Yeah, I was thinking you could use the equery database to initially
40 fill the cache. Spawning an equery instance for every file access
41 would be absolute madness. I have enough entropy problems on my
42 system.
43
44 --
45 :wq