The first part replaces a loop which is probably a performance
improvement, and the second part addresses a shellcheck complaint about
not declaring and assigning separately.
It's no longer necessary since it's equivalent to the default-features
one now. The `DIRENV_DEVSHELL` thing is left in place however in case
that becomes useful again in the future.
This is okay because we only allow fast-forward merges, changes always
go through MRs, and MRs are subject to, e.g., the code quality checks.
This change is desirable because it should save some time and energy.
GitLab doesn't seem to have built-in support for this because of course
it doesn't.
To do this, we move the job scripts to a different file to make it
possible to share code between job scripts.
Without this, we're building a static rocksdb inside the rust-rocksdb
build script every time. As far as I know this doesn't change clippy's
behavior, but it does take a *long* time.
This fixes the problem where some artifacts were not being cached when
they should have been. The secret sauce is the `nix-store` command.
Also stops emitting artifacts to GitLab. Automatic build scheduling via
Nix is too convenient. Maybe I'll figure out a way to do both later on.
Also pins the remaining unpinned dependencies, namely direnv and
nix-direnv.
This drastically reduces the maintenance burden.
You may notice the `CODE_OF_CONDUCT.md` is deleted here. This is because
I don't feel like updating the relevant parts and I don't think this
will ever actually have a community build around it. If that changes for
some reason, I'm not opposed to adding a code of conduct again.
And also the public key so that pulling from the new endpoint will work.
This allows other people to host their own attic instances and configure
their (CI) environment to override the default endpoint so e.g. they can
take advantage of a binary cache without having write access to the
official one.
I didn't actually test this change but I think it should work.
Also why'd I format the script like that, ew lol
This will also include attic, so we don't need to explicitly do this
in `./bin/nix-build-and-cache` anymore, which is good because that
script gets called a good number of times and doing that repeatedly was
a bit of a waste.
Well, kinda. It crashed on me after 10 minutes because the tests timed
out like in <https://github.com/matrix-org/complement/issues/394>.
Sounds like this means it's a them problem though.
I want to use Nix to build this image instead in the future but this
will at least make it work for now and give me a reference for while I'm
porting it. I also want to make Conduit natively understand Complement's
requirements instead of `sed`ing a bunch of stuff and needing a reverse
proxy in the container. Should be more reliable that way.
I'm not making this run in CI until the above stuff is addressed and
until I can decide on a way to pin the revision of Complement being
tested against.
This will allow most Nix users to use the `default` package and without
having to build from source. And also allows any weirdos to get the OCI
image from the Nix binary cache if they want. No idea why that would be
desireable though lol
