1 |
On Sun, Sep 07, 2014 at 09:03:00PM -0400, Rich Freeman wrote: |
2 |
> Right now the general policy is that we don't allow unmasked (hard or |
3 |
> via keywords) ebuilds in the tree if they use an scm to fetch their |
4 |
> sources. There are a bunch of reasons for this, and for the most part |
5 |
> they make sense. |
6 |
> |
7 |
> I was wondering if this policy still made sense in the case of scm |
8 |
> ebuilds that pull a particular git commit. While portage can't check |
9 |
> the Manifest, the fact is that git will in this case, and since we're |
10 |
> pointed at a content-hashed commit we can ensure that the package |
11 |
> never changes. We of course can't mirror it with the current setup |
12 |
> (there is no real reason we couldn't mirror git, but this is a |
13 |
> different problem). |
14 |
> |
15 |
> Tying ebuilds to a git commit has pros and cons, but I'm hard-pressed |
16 |
> to think of any actual QA issues. That is, something that would make |
17 |
> our tree inconsistent, or create a security vulnerability. |
18 |
> |
19 |
> Am I just not thinking of something? It would probably be most useful |
20 |
> for packages that track a backport branch or something along those |
21 |
> lines - where upstream does not regularly update their tarballs so |
22 |
> we're constant creating patchsets. In this case all we'd have to do |
23 |
> is bump the commit ID in the ebuild. |
24 |
> |
25 |
|
26 |
What you're saying makes perfect sense at the developer->infra side, |
27 |
but as you've conceded, not so much at the infra->mirror side. |
28 |
|
29 |
Clearly instead of every user downloading a git clone, it's better to |
30 |
pin to that commit and have the push script hook with infra to automate |
31 |
a tarball, since the dev->infra channel is secure, and the SHA1 is |
32 |
current at that point. That would enable you to submit a pinned ebuild |
33 |
with the relevant keywords, users to download standard tarballs and |
34 |
verify if needed (since the commit id is in the ebuild, and that has |
35 |
a strong manifest); the saving in bandwidth is something sponsors |
36 |
might go for. |
37 |
|
38 |
I'm sure many users would be happy to contribute in various fashions, |
39 |
too; though I'd ask patrick to consult and perhaps provide backend if |
40 |
infra is stretched. He runs hosts for quite a lot of gentoo people |
41 |
already, as well as git and trac machines, and the tinderbox, on |
42 |
gentooexperimental.org. |
43 |
|
44 |
Regards |
45 |
steveL. |
46 |
-- |
47 |
#friendly-coders -- We're friendly, but we're not /that/ friendly ;-) |