Arun Isaac <[email protected]> writes:

> Hi Phong,
>
>>> Users would see progress bars like "Resolving deltas:" [...] Nix
>>> already solved the client-side problem. The package manager fetches
>>> expressions as tarballs via channels,
>>
>> Authentication of Guix channel tarballs would be nice, of course it's
>> not as good as full history authentication, but from an user PoV speed
>> could be an important tradeoff.
>
> Something like this would be very nice. It would be much preferable if
> every user didn't have to hit the git repo.
>
>>> The [nixpkgs] repository totals 83GB with half a million tree objects
>>> and 20,000 forks.  A local clone is only 2.5GB.  The rest is GitHub’s
>>> fork network storing every pull request branch and merge commit.
>>
>> Sounds like a GitHub problem to me /s
>>
>> But really, this is not a problem inherent to Git.  Something like AGit
>> could help with garbage collectibility on Codeberg's side,
>> and this was not an issue when we were using git:send-email.
>
> True. Although Codeberg would probably run into the same problems as
> GitHub, and possibly sooner considering GitHub must have put a lot of
> engineering into their infrastructure.
>

Codeberg doesn’t have (yet) the same deduplication that Github has, with
our 389 forks and ~500M per repository, that’s almost 200G!

>>> Databases have CHECK constraints and UNIQUE constraints; git has
>>> nothing, so every package manager builds its own validation layer.
>>> Databases have locking; git doesn’t. Databases have indexes for
>>> queries like "all packages depending on X"; with git you either
>>> traverse every file or build your own index.
>>
>> This is funny because it precisely describe Guix d-;
>
> :-)
>
>> The problem with developing directly on a database though is there's
>> no colaborative tooling for it AFAIK.
>
> Agreed.
>
>> In brief, aside from guix pull it does not seem to me like there is
>> much can be done without going back to the drawing board, but I'm no
>> expert on this and wish to be corrected.
>
> At some point, sooner or later, we might have to go back to the drawing
> board, and that process will be painful. But, I am no expert on this
> topic either. Hopefully, someone like Ludo could comment.
>
> Regards,
> Arun

Attachment: signature.asc
Description: PGP signature

Reply via email to