Age | Commit message (Collapse) | Author | Files | Lines |
|
There was a NixHash::new() before, which didn't perform any validation
of the digest length. We had some length validation when parsing nix
hashes or SRI hashes, but some places didn't perform validation and/or
constructed the struct directly.
Replace NixHash::new() with a
`impl TryFrom<(HashAlgo, Vec<u8>)> for NixHash`, which does do this
validation, and update constructing code to use that, rather than
populating structs directly. In some rare cases where we're sure the
digest length is correct we still populate the struct manually.
Fixes b/291.
Change-Id: I7a323c5b18d94de0ec15e391b3e7586df42f4229
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9109
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Autosubmit: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
Previously, Output deserialization would silence validation errors and
provide `None` for `hash_with_mode` as soon as a validation error would
happen inside of the `NixHashWithMode` deserialization, e.g. invalid
hash length would not provide an validation error but a silent `None`
value.
This is problematic, we workaround a serde limitation here by writing
our own Deserializer.
As you can see, we write some boilerplate code unfortunately, as, for
example:
- `#[serde(fail_if_unparsed_as="Option::is_none")]` is not a thing,
otherwise, we could have been able to just bubble up errors in case of
"not fully parsed" (and not missing) values.
- `From<&serde_json::Value> for serde::de::Unexpected` is not a thing,
otherwise, we could just map invalid type errors and reuse the
existing types instead of doing extremely bizarre things with
`serde::de::Unexpected::Other`, note: this is a not problem for
expected, we know what we expect, we don't know what we received in
practice.
I decided to write a `NixHashWithMode::from_map` which will eat a map
deserialized via `serde_json`, so our serde magic is not totally "data
model" agnostic.
I wanted to go for data model agnosticity and enable maximal
performance, e.g. building the structure as the map values are streamed
in the Visitor, this is needlessly painful because `Output` and
`NixHashWithMode` are in different files and this really makes sense
only if we write the full implementation in one file, indeed, otherwise,
you end up duplicating the work or having strange coupling.
So, for now, we will allocate a full map of the fields inside the
`Output`, i.e. if any "unknown field" is in that map, we will
deserialize it for no reason.
Doing it properly, as I repeat it in the code and to flokli at C3Camp
2023, requires to patch serde upstream IMHO.
Change-Id: I46fe6ccb8c390c48d6934fd3e3f02a0dfe59557b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9107
Tested-by: BuildkiteCI
Reviewed-by: flokli <flokli@flokli.de>
|
|
Change-Id: I1e52b752408a70fdb27e3e5c6f3bef5417b3e922
Reviewed-on: https://cl.tvl.fyi/c/depot/+/9085
Tested-by: BuildkiteCI
Reviewed-by: raitobezarius <tvl@lahfa.xyz>
Autosubmit: flokli <flokli@flokli.de>
|
|
HashAlgo implements Copy, no need to clone here.
Change-Id: Ief11d2cfbd4fd0cd44224c7ddd575f518edbbd55
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8989
Tested-by: BuildkiteCI
Reviewed-by: tazjin <tazjin@tvl.su>
Autosubmit: flokli <flokli@flokli.de>
|
|
Before there was code scattered about (e.g. text hashing module and
derivation output computation) constructing store paths from low level
building blocks --- there was some duplication and it was easy to make
nonsense store paths.
Now, we have roughly the same "safe-ish" ways of constructing them as
C++ Nix, and only those are exposed:
- Make text hashed content-addressed store paths
- Make other content-addressed store paths
- Make input-addressed fixed output hashes
Change-Id: I122a3ee0802b4f45ae386306b95b698991be89c8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8411
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|
|
They can go under `nixhash`
Change-Id: Ia15835c57130b66d58f5df80ae9595dceee00941
Reviewed-on: https://cl.tvl.fyi/c/depot/+/8408
Reviewed-by: flokli <flokli@flokli.de>
Tested-by: BuildkiteCI
|