The rust-analyzer update will need more work as it seems to break rust
analysis on windows.
This was carried out using `cargo upgrade` from `cargo-edit`:
* getting exclusions options for rust-analyzer with
```bash
cargo upgrade -i --dry-run | grep -o 'ra_ap_\S\+' | sort -u | sed 's/^/--exclude=/' > /tmp/exclude
```
* running
```bash
cargo upgrade -i $(cat /tmp/exclude)
misc/bazel/3rdparty/update_cargo_deps.sh
```
* The ungram file is now taken from the rust-analyzer dependencies
pulled in by bazel
* the grammar parsing code is not published, so it must be taken
directly from rust-analyzer code. That part should be less prone to be
updated than the ungram file, so it does not necessarily need to be
in sync with the rust-analyzer version is used elsewhere.
* both need some patches. The former is patched during build, the latter
during loading in `MODULE.bazel`.
This splits the `ArrayExpr` class into `ArrayListExpr` and `ArrayRepeatExpr`.
This uses the `synth.from_class` machinery to integrate seamlessly into the
generated code, by hiding the extracted `ArrayExpr` behind an internal class
and replacing it with a hierarchy of those two classes under a new
`ArrayExpr` class.
This doesn't require `ql.name` and is simpler while we don't have
to write upgrade scripts. The `ql.name` mechanism might get useful
once we do have to write upgrade scripts, as that doesn't change the
dbscheme.
We've been observing some performance issues using crate_universe on CI.
Therefore, we're moving to vendor the auto-generated BUILD files
in our repository. This should provide a nice speed boost, while
getting rid of the complexity of the "rust cache" job we've been using
when we had a lot of git dependencies.
This PR includes a vendor script, and I'll put up a CI job internally
that runs that vendor script on Cargo.toml and Cargo.lock changes, to check
that the vendored files are in sync.
