mirrors/arti.git/crates/tor-hsservice/src, branch arti-v1.1.10mirror of https://gitlab.torproject.org/tpo/core/arti
http://git.dilluti0n.com/mirrors/arti.git/atom?h=arti-v1.1.102023-10-31T11:40:36Ztor-hsservice: Remove an unused mut in a test2023-10-31T11:40:36ZIan Jacksonijackson@chiark.greenend.org.uk2023-10-30T15:53:08Zurn:sha1:be2db12c6f0381e0fdcc3f889f2f9a13f2ea95e4
I think retaining this lint even in tests is probably reasonable.
IptsPublisherView has interior mutability, so with this function
taking one, it won't ever need mut.
Merge branch 'timeout' into 'main'2023-10-30T18:28:34ZIan Jacksoniwj@torproject.org2023-10-30T18:28:34Zurn:sha1:8176dac8f597a29524db3c0b8f6e457ea9b2373f
timeout_track improvements
See merge request tpo/core/arti!1711Fix typo in doc comment2023-10-30T18:06:22Zgabi-250gabi@torproject.org2023-10-30T18:06:22Zurn:sha1:0a3f52a518da97bf5be5a64411648dd178c7d21ator-netdoc: Use the new HandshakeType enum to represent CREATE2 HTYPEs.2023-10-30T14:18:02ZGabriela Moldovangabi@torproject.org2023-10-30T13:56:58Zurn:sha1:28a0b24db240cc24230c80732c5f4ba3a0b0d461
Representing the supported HTYPEs as `HandshakeType`s instead of `u32`s
makes it more difficult to pass in wrong/invalid values to
`HsDescBuilder::create2_formats`.
This also fixes a descriptor publisher bug spotted by @jnewsome, where
the advertised CREATE2 HTYPEs included HTYPE `1`, which is actually
supposed to be a reserved value. The publisher now only advertises the
`NTOR` HTYPE (just like C Tor).
tor-hsservice: timeout_track: Test `update` functions2023-10-30T13:16:37ZIan Jacksonijackson@chiark.greenend.org.uk2023-10-30T12:36:03Zurn:sha1:54c1b038712d6951c5d748dbec21bdb41c387725tor-hsservice: timeout_track: Provide `update` via a trait2023-10-30T13:16:37ZIan Jacksonijackson@chiark.greenend.org.uk2023-10-30T12:57:12Zurn:sha1:a14cf6e843b3da0db5ded6787b8a5fad0411d76c
We have a profusion of similar methods with similar names and similar
documentation. Centralising the documentation in a trait is helpful.
This also means that callers can just naturally call `.update` rather
than needing to spell out `.update_abs` vs `.update_rel` etc.
tor-hsservice: timeout_track: Clarify edge cases2023-10-30T13:16:37ZIan Jacksonijackson@chiark.greenend.org.uk2023-10-30T12:38:38Zurn:sha1:9d41328ea603124d3a32664d3176045faf6e6e5a
Clarify this in the update functions, and in the general discussion.
We don't want zero timeouts.
tor-netdoc: Change return type of create_desc_sign_key_cert.2023-10-25T15:32:36ZGabriela Moldovangabi@torproject.org2023-10-25T14:30:43Zurn:sha1:aad987a47aa9584ce2ba6a16ab06710103737dc3
`Bug` wasn't necessarily the right error type here. Plus, with the new
error type adding new errors (i.e. `CertEncodeError` variants), is not a
breaking change.
tor-netdoc: Building a descriptor now only requires the public part of blinded_id.2023-10-25T15:32:36ZGabriela Moldovangabi@torproject.org2023-10-25T14:05:20Zurn:sha1:6420e5ef3460e1ff3ccb77beeb75af2c918f86d7tor-netdoc: Update HsDesc, HsDescOuter to accept the hs_desc_sign cert as an argument.2023-10-25T15:32:36ZGabriela Moldovangabi@torproject.org2023-10-23T17:56:33Zurn:sha1:7edade162c8dd5f45f8d69496084f32003357e55
This will enable us to (eventually) the load the descriptor signing key
cert from the keystore (as opposed to always recomputing it when
building the `HsDesc`).
Part of #1048