1 |
On Sun, Nov 11, 2018 at 12:31 PM M. J. Everitt <m.j.everitt@×××.org> wrote: |
2 |
> |
3 |
> Binpkgs are also a popular component of a few downstream distro's based on |
4 |
> Gentoo (thinking pentoo right now as an easy example). |
5 |
> |
6 |
> So we don't want to break existing users of this format without considering |
7 |
> the ramifications for these scenarios, as you'll have some very grumpy devs... |
8 |
> |
9 |
|
10 |
I'd argue that they'd be more important for Gentoo if they were more |
11 |
useful. IMO the main limitation with them is the inability to |
12 |
auto-download them from a repository, detecting the binpkg USE flags |
13 |
BEFORE downloading. This is why I suggested hashing the USE flags or |
14 |
similar and sticking that in the filename. |
15 |
|
16 |
Obviously you can't host a repository with all the USE combinations. |
17 |
However, you could have a reference repo and the package manager could |
18 |
check it before doing a build. If you get a hit then you can install |
19 |
the binpkg. If you don't then you can do a source build. |
20 |
|
21 |
Portage already checks the USE flags inside the binpkg before merging |
22 |
it and by default doesn't use a non-matching binpkg. The problem with |
23 |
the current approach is: |
24 |
1. You have to download the package to check this (could be a big file). |
25 |
2. You can't host multiple versions of a binpkg with different USE |
26 |
flags since the filenames collide. |
27 |
|
28 |
I suggested a content hash because you can use it for an arbitrary |
29 |
amount of metadata, vs having to cram arch/USE/multilib and I'm sure |
30 |
something I'm missing into a filename. Make the hash as short as is |
31 |
economical - it isn't like we have THAT many permutations, the PM can |
32 |
still check the internal metadata, and this isn't a security feature. |
33 |
|
34 |
-- |
35 |
Rich |