| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
`requires` is not recommended in this context. I think I tried it out of
desperation trying to combat the issue that these units would be started
before they were able to resolve names in switch-to-configuration.
Unfortunately, network access during switch-to-configuration can't be
detected using network-online.target, it seems.
Change-Id: Ia98a0a3b505ffa56eb37fb58a5375a1215d6cb1b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12794
Autosubmit: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
|
|
Repositories (especially nixpkgs) WILL grow to ridiculous sizes
otherwise, killing cgit performance in the process.
Change-Id: I3bef3e5dc5a61152e89fd53e31f14e78193a4888
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12793
Autosubmit: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Reviewed-by: sterni <sternenseemann@systemli.org>
|
|
Mostly it's important (which I did this time at least) to remember that
with how things work at the moment, the initrd sshd is only reachable
via IPv4.
Change-Id: Ie9a87b6a38b2e128a8a2141d2221bbe7cfe24cdb
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12792
Autosubmit: sterni <sternenseemann@systemli.org>
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
|
|
nixos-generate-config claims I need this, so let's add it…
Change-Id: I3d852ffce5d0e7c65d9a1bbe887d3de15136698a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12791
Reviewed-by: sterni <sternenseemann@systemli.org>
Tested-by: BuildkiteCI
Autosubmit: sterni <sternenseemann@systemli.org>
|
|
I've discovered that it is possible for irccat to fail enough times to
run into the restart limit before network is online after booting.
Change-Id: Ia54a46d56bdc765a825fee50e7bdc8206718edc0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12790
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: sterni <sternenseemann@systemli.org>
|
|
* update wasm-bindgen in all wasm projects
* //users/wpcarro/website: declare missing dependency on
string-conversions. Presumably this was propagated
before from some other dependency which got updated now.
Change-Id: Ib93de576408974441d532196601e6e53d22cdafe
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12770
Tested-by: BuildkiteCI
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
Change-Id: I4800c542b2af0bc2a5e389342f3574b4928b1704
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12789
Autosubmit: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
|
|
I don't have this keyboard anymore, no need to build the firmware
anymore.
Change-Id: I1ab25dfaa71d03dc7006312ecab86453d6f90333
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12788
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
`buildSplitKeyboard`, as well as all the patching of `miryoku_zmk` is
independent of the specific keyboard used, so it can be moved one layer
up.
`config-flat` is now provided through a helper function, accepting the
name of the keymap to use when rendering the config.
This all makes the amount of code added for the new keyboard itself
pleasantly small.
Change-Id: I2216aa246502eddaf9bc4f4d126b0639d574ad87
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12787
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
Change-Id: Ia57768d30db4a841f74c4305cfbae45c9e26522e
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12767
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
Change-Id: If5296eb7aa7304801ade7a2c3a1aa6907b051e4f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12768
Autosubmit: tazjin <tazjin@tvl.su>
Reviewed-by: tazjin <tazjin@tvl.su>
Tested-by: BuildkiteCI
|
|
Change-Id: Ib5e518359e152553d0a9ebf6c674f1acc846800a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12785
Tested-by: BuildkiteCI
Reviewed-by: Profpatsch <mail@profpatsch.de>
Reviewed-by: lukegb <lukegb@tvl.fyi>
|
|
These are required to support certain nix's local-overlay store
operations, it's safer to return empty results for these
operations than failing with "operation not implemented" errors.
Change-Id: Ic9b69d75dd52af5a826bfb6a8b283b082a0f6bcf
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12766
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
In particular QueryPathFromHashPart, QueryValidPaths, QueryValidDerivers
Change-Id: Ie6ad83cec5ce9580044b85e201e4e23394f87075
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12762
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
Reviewed-by: flokli <flokli@flokli.de>
|
|
This change contains 2 improvements to the tvix-store copy command:
1. Allows reading the reference graph from stdin, using `-` argument
2. Supports json representation produced by `nix path-info --json`
command.
In general it makes is easier and faster to import arbitrary closures
from an existing nix store with e.g the following command:
```
nix path-info ./result --json --closure-size --recursive | \
jq -s '{closure: add}' | \
tvix-store copy -
```
Change-Id: Id6eea2993da233ecfbdc186f1a8c37735b686264
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12765
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
By manually writing out the async function, and moving the owned String
we can avoid for S to be Sync in the NixSerialize for StorePath<S> impl.
Co-Authored-By: edef@edef.eu
Change-Id: I8427b38d2bd61de2818088728cbad79cda69d17d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12764
Reviewed-by: Vladimir Kryachko <v.kryachko@gmail.com>
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: edef <edef@edef.eu>
Tested-by: BuildkiteCI
|
|
This describes the current composition system used for BlobService /
DirectoryService / PathInfoService, why it's hidden, how to expose it,
and adds some common examples to explain it.
Change-Id: I2ce7da40992cc988947c3e924499f8157c5e4937
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12749
Tested-by: BuildkiteCI
Reviewed-by: yuka <yuka@yuka.dev>
|
|
Implement PartialEq/Eq ourselves instead of deriving, by proxying to
name.as_ref() (and digest of course).
Also implement Hash on our own, clippy doesn't like this to be derived,
while Eq/PartialEq is not.
Change-Id: Idbe289a23ba3bc8dabf893d4d8752792ae2778c3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12744
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
Autosubmit: flokli <flokli@flokli.de>
|
|
NixString doesn't contain a `HashSet<BuildReference>` anymore, there's a
more packed internal data structure. We don't need to be overly specific
in the docs, just say we expose an API.
Change-Id: I13380c49293f9a86d1916909fdfeefbe64d9024b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12755
Tested-by: BuildkiteCI
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: edef <edef@edef.eu>
|
|
This conversion was a bit too magic, and we can just use
`NixString::new_context_from` without having to worry about the
distinction between an empty context or no context, as
NixString::new_context_from already deals with that internally.
Change-Id: I3e5d57ecfa0f7456aa6c526863e49f2523afaec3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12754
Tested-by: BuildkiteCI
Reviewed-by: edef <edef@edef.eu>
Autosubmit: flokli <flokli@flokli.de>
|
|
Change-Id: I07a13da0ae4aee4298025fca4345d738f40cfe5a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12757
Reviewed-by: Ilan Joselevich <personal@ilanjoselevich.com>
Reviewed-by: edef <edef@edef.eu>
Tested-by: BuildkiteCI
|
|
Change-Id: I0e75445a34d3ee11e535e6948c6d762f4fb15105
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12760
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: Domen Kožar <domen@cachix.org>
Reviewed-by: flokli <flokli@flokli.de>
|
|
Change-Id: Ia601e2eae24a2bc13d8851b2e8ed9d6c1808bb35
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12745
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: Vladimir Kryachko <v.kryachko@gmail.com>
Tested-by: BuildkiteCI
|
|
This change includes only the basic nix handshake protocol handling and
sets up a client session. The only supported operation at this point is
SetOptions.
Additional operations will be implemented in subsequent cls.
Change-Id: I3eccd9e0ceb270c3865929543c702f1491768852
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12743
Autosubmit: Vladimir Kryachko <v.kryachko@gmail.com>
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
Reviewed-by: edef <edef@edef.eu>
Reviewed-by: Brian Olsen <me@griff.name>
|
|
This prevents the field from getting printed to stderr, apparently.
Change-Id: Ia9860e4ff37224003154db88ee5f83103060e626
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12756
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: Ilan Joselevich <personal@ilanjoselevich.com>
|
|
Allow this to be overridden via the CLI, also getting rid of the unsafe
because this doesn't need to be const anymore.
Change-Id: I5e51b52e42522a21f59ef69628b464477c0764d1
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12753
Tested-by: BuildkiteCI
Autosubmit: flokli <flokli@flokli.de>
Reviewed-by: Ilan Joselevich <personal@ilanjoselevich.com>
|
|
When uploading a Store Path to a Nix HTTP Binary Cache, Nix first does a
HEAD request for $outhash.narinfo, and if that's not found, for
`{narhash}.nar[.compression_suffix]`.
If the NAR is already present, only the NARInfo is uploaded.
Even though we don't have a service allowing to globally look up from
NARHash to root node, `root_nodes` in `AppState` at least contains
recently uploaded NARHashes.
If we find it in there, we can prevent Nix unnecessarily uploading NARs
if the same contents have already been recently uploaded.
We also promote this key, chances are high Nix will subsequently upload
a NARInfo referring to this NARHash.
Change-Id: I34e3fd9b334b695abe945e64cd291e30f303c2a2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12752
Tested-by: BuildkiteCI
Reviewed-by: Ilan Joselevich <personal@ilanjoselevich.com>
Autosubmit: flokli <flokli@flokli.de>
|
|
Align this naming with CachePathInfoService.
Change-Id: Ib9a0d73b8ca57a93e9fc027ae907fc6ed370842a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12751
Tested-by: BuildkiteCI
Reviewed-by: yuka <yuka@yuka.dev>
|
|
This being in combinators makes it harder to find.
Change-Id: If7984bdbd43f164c670548639bb4846d859f6695
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12750
Reviewed-by: yuka <yuka@yuka.dev>
Tested-by: BuildkiteCI
|
|
Extend the docstrings of `add_default_services`, and add one for
`addrs_to_configs` as well as the module-wide one at
`tvix_store::composition`.
Change-Id: Ie9b449988eb210cd65b19b174094bbe0c4af2fd6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12748
Tested-by: BuildkiteCI
Reviewed-by: yuka <yuka@yuka.dev>
|
|
tvix-castore already supports composition without any additional feature
flags, the only thing that can be explicitly enabled is referring to
other stores via an anonymous url. Rename that feature flag to
"xp-composition-url-refs".
tvix-store effectively only controls the CLI surface, so rename this to
"xp-composition-cli".
The "store" in the feature name was dropped, as it's already apparent
from being in the tvix-[ca]store crate.
Change-Id: I1175dc6280cbba4cbcdfb7fd4b35fce713b45fc4
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12747
Tested-by: BuildkiteCI
Reviewed-by: yuka <yuka@yuka.dev>
|
|
This becomes the root of the composition. `default` implies we can
directly access anything else, which we cannot. `root` makes this more
understandable, and it's all internal only anyways.
Change-Id: I297511bc05a7c32c59510b9d192b40d1bd937b5f
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12746
Reviewed-by: yuka <yuka@yuka.dev>
Tested-by: BuildkiteCI
|
|
This groups most `wire` feature gated logic into a single module.
The nix_daemon module will be gated by a feature that adds
nix-compat-derive as a dependency.
All of this is a way to break the crate2nix dependency cycle between
nix-compat and nix-compat-derive(which depends on nix-compat for its
doctests).
Change-Id: I95938a6f280c11967371ff21f8b5a19e6d3d3805
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12761
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
Change-Id: Icd81415c208c26ae606673f3476bfdf0525eec89
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12759
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
Currently it is not possible to distinguish between tracing of the same
*Service type whenever there are multiple of them. Now the instance_name
of ServiceBuilder is passed into the *Service and used in the existing
instrument as the `instance_name` field.
Places that did not already have a instance_name in its context use
`"default"`. In tests I used `"test"`.
Change-Id: Ia20bf2a7bb849a781e370d087ba7ddb3be79f654
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12739
Tested-by: BuildkiteCI
Autosubmit: Bob van der Linden <bobvanderlinden@gmail.com>
Reviewed-by: flokli <flokli@flokli.de>
|
|
This will be used by tvix-daemon to write tracing data into the active
client's connection socket.
Change-Id: I8889dd0a638e004ee2c8cb312946b029c9779313
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12734
Tested-by: BuildkiteCI
Autosubmit: Vladimir Kryachko <v.kryachko@gmail.com>
Reviewed-by: flokli <flokli@flokli.de>
|
|
This makes it quite hard to press Alt+Number otherwise.
Change-Id: Id1cc4fbfa1575cc213014106f734bec035d46f2c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12737
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Autosubmit: flokli <flokli@flokli.de>
|
|
This change implements the serialization part that is needed to
implement the nix daemon protocol. Previously was add deserialization
and derivers for that and this then adds the other part of that equation
so that you can write types that can then be read using deserialization.
Change-Id: I2917de634980a93822a4f5a8ad38897b9ce16d89
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12729
Autosubmit: Brian Olsen <me@griff.name>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
This adds a span to default handling so that any error message
gives a more precise location as a hint for why the error occurred
instead of just pointing to the type.
Change-Id: I7bf6cf38e0284f9726d670ea50a94a6b1edd8a94
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12728
Reviewed-by: Vladimir Kryachko <v.kryachko@gmail.com>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Autosubmit: Brian Olsen <me@griff.name>
|
|
Change-Id: I4d89663eb9ac772ce1008ed5ee218bc7016164e7
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12733
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
Currently Value::Json is used in combination with VMRequest::ToJson to
recursively convert tvix Value to serde_json::Value. This functionality
is used in builtins.toJSON as well as derivation __structuredAttrs.
Both Value::Json and VMRequest::ToJson were removed in this commit.
Related functionality in vm.rs is also removed: vm.rs does not know
about JSON anymore.
Recursively converting to serde_json now happens without going through
the VM.
Thrown errors that are part of the value of toJSON are now directly
propagated as ErrorKind, were-as previously there was a split between
CatchableErrorKind and ErrorKind, where eventually CatchableErrorKind
would be converted to ErrorKind::Catchable.
Change-Id: I066f064926c491e4c087a984f07af43d19124cfe
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12732
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
Change-Id: Ia4a9a04c7e157b6add94dc8901ffab35486fe344
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12731
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
Change-Id: I65e3a8e8284ffb4986970ea67ed8d6b216e1eb66
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12730
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
Some references to nixrs had sneaked in while writing the original
CL so this just removes those.
Change-Id: I635adbd2de46c7c1f31e1d449396253a78e1b762
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12727
Reviewed-by: Vladimir Kryachko <v.kryachko@gmail.com>
Autosubmit: Brian Olsen <me@griff.name>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
When running tests for buildservice on MacOS it would fail
because of the oci test. I also got some clippy warnings on
MacOS because of disabled code.
Change-Id: I235739fa4270a4ef46e54d3e2b8cbb55eb20bdda
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12726
Reviewed-by: flokli <flokli@flokli.de>
Autosubmit: Brian Olsen <me@griff.name>
Tested-by: BuildkiteCI
|
|
Before this commit only deriving of tuple enums where tested.
This adds a few tests to cover the other types of enums.
Change-Id: I62f311e4db98cab84bd4ac164d50051e9aab0d4d
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12725
Autosubmit: Brian Olsen <me@griff.name>
Reviewed-by: Vladimir Kryachko <v.kryachko@gmail.com>
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
I found some more operations that are obsolete and a few missing
fields. I also did some small changes to make handshake description
more consistent.
Change-Id: I9f853a37e9d50621491f31ea71879d6a600046e2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12724
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
Change-Id: Id47c1e3f98cb0bf53006b002417b69cfa49b7410
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12721
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
Reviewed-by: Ilan Joselevich <personal@ilanjoselevich.com>
|
|
Previously such errors showed up as:
error[E006]: expected value of type 'sha256', but found a 'not a sha256'
Now they show up as:
error[E041]: Invalid hash: invalid encoded digest length '31' for algo
sha256
This is consistent with the errors of `builtins.fetchurl`.
Change-Id: Id11b26fc7951778640cc4e41b3bf23203eaf07df
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12719
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
The nixhash errors were wrapped in a generic TvixError. Now it has its
own TvixError with unique error code. The nixhash error is passed along
as a string.
The errors looked like:
error[E997]: invalid encoded digest length '51' for algo sha256
Now they look like:
error[E041]: Invalid hash: invalid encoded digest length '51' for algo
sha256
Change-Id: I5c420815538ba4c6567c95f5d44d60c4d48f43fd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/12718
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
