<https://github.com/ruma/ruma/pull/2050>
We could just set None now, but the ruma changelog states that
> Servers are encouraged to keep sending it for compatibility with
> clients that required it.
The only change to the internal interface for this one is removing the
current_third_party_invite argument from state_res::auth_check. Ruma now
fetches the event using the fetch_state closure. This is convenient for
us, because we previously didn't bother to implement it.
The significant change is 26edd40a704040e7104161da81c9bae91b7ddcaa,
which removes the global compat feature, so that each compat feature
must now be enabled individually. We're using the slightly later
1387667d because it has a bugfix that ruma needs to compile.
There are a few ruma compat features that were not previously part of
the global compat feature:
- compat-arbitrary-length-ids
- compat-upload-signature
- compat-encrypted-stickers
I have not enabled these here, to avoid a behavior change.
There are two new default features:
'system-proxy' for reqwest, which we are not enabling because it only
does anything on windows and mac, which we don't support.
'prefer-post-quantum' for rustls, which we are not enabling because it
only does anything for aws-lc-rs, which we don't use.
Previously we were only using trust-dns for resolving SRV records in
server discovery, and then for resolving the hostname from the SRV
record target if one exists. With the previous behavior, admins need to
ensure that both their system resolver and trust-dns are working
correctly in order for outgoing traffic to work reliably. This can be
confusing to debug, because it's not obvious to the admin if or when
each resolver are being used. Now, everything goes through trust-dns and
outgoing federation DNS should fail/succeed more predictably.
I also expect some performance improvement from having an in-process DNS
cache, but haven't taken measurements yet.
Ruma dropped a couple dependencies and includes a stateres performance
improvement. May as well pull in everything else (except OTel) while
we're at it.
And some supporting changes:
* crane: It removed its dependency on nixpkgs and made overrideToolchain
take a function for splicing reasons, but we're doing splicing
ourselves so we can just ignore the function argument. These changes
are in `flake.nix`.
* [NixOS/nixpkgs#347228][0]: linkerFor* were removed because the linker
no longer needs to be different in some edge cases. Based on the
diff of the PR that introduced this change, ccFor* are the proper
replacements. These changes are in `cross-compilation-env.nix` in the
compiler-and-linker-choosing section.
* [NixOS/nixpkgs#350299][1]: buildPlatform isn't at the top level
anymore, we have to go through stdenv now. These changes are in
`nix/shell.nix`.
* rocksdb: nixpkgs has 9.6.1 now so we need to upgrade our rust
library to use the matching version. These changes are in
`Cargo.toml`, `Cargo.lock`, `nix/pkgs/default/default.nix`, and
`cross-compilation-env.nix` in the linker flags section.
[0]: https://github.com/NixOS/nixpkgs/pull/347228
[1]: https://github.com/NixOS/nixpkgs/pull/350299
Flake lock file updates:
• Updated input 'attic':
'github:zhaofengli/attic/4dbdbee45728d8ce5788db6461aaaa89d98081f0' (2024-03-29)
→ 'github:zhaofengli/attic/48c8b395bfbc6b76c7eae74df6c74351255a095c' (2024-10-30)
• Updated input 'attic/crane':
'github:ipetkov/crane/7195c00c272fdd92fc74e7d5a0a2844b9fadb2fb' (2023-12-18)
→ 'github:ipetkov/crane/4c6c77920b8d44cd6660c1621dea6b3fc4b4c4f4' (2024-08-06)
• Updated input 'attic/flake-compat':
'github:edolstra/flake-compat/35bb57c0c8d8b62bbfd284272c928ceb64ddbde9' (2023-01-17)
→ 'github:edolstra/flake-compat/0f9255e01c2351cc7d116c072cb317785dd33b33' (2023-10-04)
• Added input 'attic/flake-parts':
'github:hercules-ci/flake-parts/8471fe90ad337a8074e957b69ca4d0089218391d' (2024-08-01)
• Added input 'attic/flake-parts/nixpkgs-lib':
follows 'attic/nixpkgs'
• Removed input 'attic/flake-utils'
• Updated input 'attic/nixpkgs':
'github:NixOS/nixpkgs/07262b18b97000d16a4bdb003418bd2fb067a932' (2024-03-25)
→ 'github:NixOS/nixpkgs/159be5db480d1df880a0135ca0bfed84c2f88353' (2024-09-11)
• Updated input 'attic/nixpkgs-stable':
'github:NixOS/nixpkgs/44733514b72e732bd49f5511bd0203dea9b9a434' (2024-03-26)
→ 'github:NixOS/nixpkgs/797f7dc49e0bc7fab4b57c021cdf68f595e47841' (2024-08-22)
• Added input 'attic/nix-github-actions':
'github:nix-community/nix-github-actions/e04df33f62cdcf93d73e9a04142464753a16db67' (2024-10-24)
• Added input 'attic/nix-github-actions/nixpkgs':
follows 'attic/nixpkgs'
• Updated input 'crane':
'github:ipetkov/crane/109987da061a1bf452f435f1653c47511587d919' (2024-05-24)
→ 'github:ipetkov/crane/498d9f122c413ee1154e8131ace5a35a80d8fa76' (2024-10-27)
• Removed input 'crane/nixpkgs'
• Updated input 'fenix':
'github:nix-community/fenix/b6fc5035b28e36a98370d0eac44f4ef3fd323df6' (2024-05-22)
→ 'github:nix-community/fenix/87b4d20f896c99018dde4702a9c6157b516f2a76' (2024-11-01)
• Updated input 'fenix/rust-analyzer-src':
'github:rust-lang/rust-analyzer/21ec8f523812b88418b2bfc64240c62b3dd967bd' (2024-05-19)
→ 'github:rust-lang/rust-analyzer/0ba893e1a00d92557ac91efb771d72eee36ca687' (2024-10-31)
• Updated input 'flake-utils':
'github:numtide/flake-utils/b1d9ab70662946ef0850d488da1c9019f3a9752a' (2024-03-11)
→ 'github:numtide/flake-utils/c1dfcf08411b08f6b8615f7d8971a2bfa81d5e8a' (2024-09-17)
• Updated input 'nix-filter':
'github:numtide/nix-filter/3342559a24e85fc164b295c3444e8a139924675b' (2024-03-11)
→ 'github:numtide/nix-filter/776e68c1d014c3adde193a18db9d738458cd2ba4' (2024-10-29)
• Updated input 'nixpkgs':
'github:NixOS/nixpkgs/5710852ba686cc1fd0d3b8e22b3117d43ba374c2' (2024-05-21)
→ 'github:NixOS/nixpkgs/807e9154dcb16384b1b765ebe9cd2bba2ac287fd' (2024-10-29)
The rustls version bump in c24f79b79b
introduced a panic when serving listeners with 'tls = true':
> thread 'main' panicked at /nix/store/eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee-vendor-cargo-deps/c19b7c6f923b580ac259164a89f2577984ad5ab09ee9d583b888f934adbbe8d0/rustls-0.23.13/src/crypto/mod.rs:265:14:
> no process-level CryptoProvider available -- call CryptoProvider::install_default() before this point
This commit fixes this by setting the default provider to ring. I chose
ring (the old rustls default) over aws-lc-rs (the new default) for a few
reasons:
- Judging by github issues, aws-lc-rs seems to have a lot of build problems.
We don't need more of that.
- The "motivation" section in the aws-lc-rs docs only talks about FIPS,
which we do not care about.
- My past experience with things that start with "aws-" has been very
negative.
Someone contributed opentelemetry-prometheus support for v0.24 and this
version also doesn't put stupid requirements on the tokio version. This
version of the OTel ecosystem also fixes an apparent bug with some hacks
I plan on doing in the future...
We want to be able to disable colors for complement logs (since they
are likely to be opened in a text editor). There's no pressing need for
alternative log formats, but I'm interested in whether the 'pretty'
format will be easier for debugging.
I chose to add 'log_*' options rather than making a separate 'log'
section for now. There's been some discussion about trying to separate
the tracing/logging stuff into more structured sections, but that can
happen later.
Both the hand-rolled parser and serialization were wrong in countless
ways. The current Ruma parser is much better, and the Ruma serialization
will be fixed by https://github.com/ruma/ruma/pull/1830.
Unfortunately we need to pull tracing-opentelemetry from git because
there hasn't been a release including the dependency bump on the other
opentelemetry crates.
