This makes magrathea use the repostory version of nix and git. This is
done in the pursuit of enforcing guaranteed unified experience of
magrathea tool across all users of a `tvl-kit`-based
repository. Especially among ubuntu users with uncontrolled set of
packages and versions installed on their system.
Not having this was giving build problems for `mg build` as one of the
users has 2.17 version of nix that had inconsistent hash computation
with 2.3.
Change-Id: I3182faf4c545ac61f6cc1cc862dc23d51c1cd397
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9892
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Delays allocation (through cloning) of the values to be compared
until *after* the keys have been compared.
Change-Id: I7d68c27d7a0fbcdcc387db7c092bce50ca4b94ea
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9900
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Similar to `into_iter_sorted`, add a marker function for call sites
that want *borrowed* sorted iteration.
Change-Id: I7c6f14e1ac43fdb14b861b3da183eb5d12bba139
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9899
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
builtins.intersectAttrs is used a _lot_ in nixpkgs eval, for whatever
reason. We previously had a very inefficient implementation that would
allocate for each comparison. It stuck out like a sore thumb in perf
analysis.
This moves to a custom algorithm with two iterators, one for the left
and one for the right side, advancing them along the (borrowed) map
keys until a match is found and allocation is required.
I've not made any effort to reduce the verbosity of this code, I don't
think it's worth it.
On my machine this reduces the mean runtime of evaluating
`nixpkgs.emacs.outPath` by ~8%.
Change-Id: Ie506d82cb8d5f45909628f771a6b73e0eca16b27
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9898
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
As described in https://b.tvl.fyi/issues/313, the `nix_oracle` tests
currently fail if run without Nix 2.3 in `$PATH`.
This is not a problem for people running Nix 2.3 on their systems, and
neither is it for the tests running inside Nix derivations itself (as we
hardcode Nix 2.3 there), but for interactive `cargo test` runs, either
using `mg shell //tvix:shell`, or `nix-shell` (from the tvix workspace)
it currently is broken.
Adding Nix 2.3 to the shell fixes that.
Change-Id: Idbe01aa92d63eb1a3dab5797b9be5eba1e8d7f7a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9907
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
The comment about what fixes have been applied was outdated (more was
cherry-picked), and all our backports have been upstreamed to the 2.3-
maintenance branch.
Switch to there, which will also give us support for zstd, which was
merged separately.
Fixes https://b.tvl.fyi/issues/328
Change-Id: Ice25ebaaf8250a97f032a40a4dd0c7fad2ebda28
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9906
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
The purpose of black_box is to actually prevent the compiler from being
able to optimize computation of the benchmarked function away.
To accomplish this, we need to actually *use* black_box to blackbox the
input data away, rather than the return type.
Change-Id: I5438982f57509fbf7b85034346a2739d76aef1fa
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9902
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
This can apparently work around some of the CPU throttling bugs on
~modern~ computers.
Change-Id: I807ece85d3eba53857a1cb1e73a33f7924538e96
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9895
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Together with temporarily pointing nixpkgs past
b10994c38c61038970a19fa60bfbec21a61755cc, this now fixes cross-
compilation for tvix.
This incorporates the changes proposed in https://cl.tvl.fyi/9888 and
https://cl.tvl.fyi/9889, but by fixing it in crate2nix, and using the
(re-)generated version of it.
Changes were sent upstream at
https://github.com/nix-community/crate2nix/pull/309,
this pulls in a minimal patch for now.
Change-Id: I70bb6f003bbc3e89de9c4eb4985ea4708ac3a9fd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9890
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Tested-by: BuildkiteCI
These are still a bit WIP, but should already explain the idea better
than not having any documentation about how we intend to implement this
at all.
Let's put this into the repo, as a working document.
Co-Authored-By: edef <edef@edef.eu>
Change-Id: Ic0c5195cdf0a33c8382af1b5cff9064fe6ef9ffb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9894
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
These describe the castore data model, so it should live in the castore
crate.
Also, some minor edits to //tvix/store/docs/api.md, to honor the move of
the castore bits to tvix-castore.
Change-Id: I1836556b652ac0592336eac95a8d0647599f4aec
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9893
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Apply the patch that didn't land using `patches`.
Change-Id: Id87675cd44123eea8cc2175efe04f61421f26f37
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9886
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: edef <edef@edef.eu>
Tested-by: BuildkiteCI
* //3p/nixpkgs: use py3status from stable for //users/grfn
* //3p/nixpkgs: use electrum from stable for //users/tazjin
* //3p/overlays: update crate2nix vendor hash
not sure why this hash is not stable if the package isn't updated,
but whatever
Change-Id: Ia7a13ac35b9c5534eefa53eff66e47ea1c32c62a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9830
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
This is required because Go 1.18 is actually being deleted. I've
applied the formatting breakage that it introduces (such as breaking
comment formatting), because I can't be bothered to try and work
around broken Go stuff.
Change-Id: Ica7cee0d01228845d6a766079fef36df99a3da96
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9832
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Expose `deps` separately, add a direnv with PATH_add for it to bring
tooling into $PATH.
Change-Id: I432cd2b082cad89e08bef78dc4653e10e137cd6b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9842
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Avoid having to re-enter the shell whenever the config is changed.
Change-Id: Ib9f6bb4075e29acaeb4863d64c017695ca85b60b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9841
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This allows cross-compiling depot targets to other architectures, so
`nix-build --argstr crossSystem aarch64-linux -A tvix.nar-bridge`
will emit a cross-compiled aarch64-linux binary.
There's still some other cross-related issues in crate2nix to sort out
for crate2nix builds, but this CL can already land.
Co-Authored-By: raitobezarius <tvl@lahfa.xyz>
Change-Id: I467d49d125dd707a4142bfde18eea4a1c3afaf70
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9755
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Tested-by: BuildkiteCI
This add the EC2 box config to the repo.
Change-Id: Id7a888a2cfbf1454cd9f9465018df377e14b4e9f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9836
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
This adds a deploy-archeology script.
I tried getting morph to work first, but passing it a
depot.ops.nixos.nixosFor seems to be very hard - the NixOS module system
doesn't like the arguments it's called with.
Replace morph with a 3 line bash script, which assumes your ssh_config
contains config for an `archeology` host.
Change-Id: I2bf694c60ded39c201efbbb899f3b5512aa4d0f2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9835
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
We also switch the MissingField error to &'static str, since we only
parse a fixed set of fields.
Together, this makes the performance impact of error handling
negligible in batch happy-path parsing.
Change-Id: I2bd0ef2f5b35fcaced56b32d238eca75ac199ef1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9867
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: edef <edef@edef.eu>
Tested-by: BuildkiteCI
We primarily want to measure the speed of the happy path.
Change-Id: Iad0146dde86fc262e2a4b8295bde4eb297b8bf30
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9866
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Autosubmit: edef <edef@edef.eu>
This provides more info about where a NARInfo failed to parse, rather
than just returning None and leaving a library user to manually debug.
Change-Id: I9a28ddd8e5712101483ebe686fdc474c7bbc8e4e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9831
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
