mirrors/arti.git/crates/tor-dirserver/src/database.rs, branch arti-v2.1.0mirror of https://gitlab.torproject.org/tpo/core/arti
http://git.dilluti0n.com/mirrors/arti.git/atom?h=arti-v2.1.02026-01-22T16:25:24Ztor-dirserver: Add comment on compression bug2026-01-22T16:25:24ZClara Englercve@cve.cx2026-01-22T14:42:47Zurn:sha1:7678f36d675ca762f3672c5bdc0cc3f007ee299e
This commit adds a comment explaining why we treat compression errors as
a bug.
tor-dirserver: Make compression failures a bug2026-01-22T16:25:15ZClara Englercve@cve.cx2026-01-21T17:53:24Zurn:sha1:bc59f1ef133eab5c70426388deb2bc95b8c6b8a4
This commit removes `DatabaseError::Compression` because it does not fit
in. Right now, this single variant makes the error to be call-site
oriented which is not nice for error handling. Instead, this error
should indicate that something was truly wrong with the database in
itself, such as an invalid schema, a low-level SQLite bug, etc.
Instead, we now map a compression error to `DatabaseError::Bug` because
there is no good reason on why it should fail, given that we compress
memory data to memory data. Probably because it uses the
`std::io::Writer` interface which itself demands use of
`std::io::Result`.
tor-dirserver: Remove FromStr for DocumentId2026-01-22T15:20:19ZClara Englercve@cve.cx2026-01-22T15:20:19Zurn:sha1:e17869b7bc0e7c674d5d770aec8c5cd120246a4c
This commit removes FromStr for DocumentId because it was only used in
testing anyways.
Instead, it replaces it with a simple From<[u8; 32]> only enabled in
test builds, which is sufficient for what we are trying to do.
An alternative would be to make the inner field pub, but this seems to
aggressive for a testing only thing.
tor-dirserver: Add `sha2` to `router_descriptor`2026-01-19T10:01:59ZClara Englercve@cve.cx2026-01-15T14:55:13Zurn:sha1:4dc3571fe6c6b7bd8a3aa9d15da3074a140b054e
This comit adds a `sha2` column to `router_descriptor` alongside a
`CHECK` to see whether it equals `docid`.
The reason for this is simple: Microdescriptors are the only kind of
documents that are queriable with a SHA2 hash. Previously, we would
have simply used the `docid` column for this, but in order to abstract
it better, a distinct column with this hash is better.
tor-dirserver: Rename `doc_id` to `docid`2026-01-19T10:01:59ZClara Englercve@cve.cx2026-01-15T14:53:39Zurn:sha1:0ba6540a84dc6ee15b79cdb49137fa659f4a4de8
Because we went with `docid` in the database (due to `rowid`), it is
only natural to call the code variables `docid` too.
tor-dirserver: Rename sha256 to docid in schema2026-01-19T10:01:59ZClara Englercve@cve.cx2026-01-15T14:50:28Zurn:sha1:30d43c6384c57c7f90dbac7f40d496143cec6102
This commit renames the sha256 column to docid for the reason that we
agreed upon making the schema visible to all modules, so if we were to
encapsulate docid properly, this change is only natural.
tor-dirserver: Rename database meta table2026-01-19T10:01:59ZClara Englercve@cve.cx2026-01-15T14:43:30Zurn:sha1:aa72981f33acc66312db73601c09958d6cfa6ba0
This commit renames arti_dirmirror_schema_version to
arti_dirserver_schema_version.
database: Introduce DocumentId2026-01-19T10:01:59ZClara Englercve@cve.cx2026-01-15T14:33:19Zurn:sha1:2e0bf76a2878e7f9979dcac4bebe5b986596d6aa
This commit introduces a new type in the database module named
`DocumentId` for abstracting the underlying content-addressable hash
algorithm we are using.
tor-dirserver: Small pre-merge fixes2026-01-15T10:59:00ZClara Englercve@cve.cx2026-01-15T10:59:00Zurn:sha1:4c2ce57374d88e4f34c60daac212247415a9d6d7tor-dirserver: Add TODO wrt hex::encode_upper2026-01-15T10:49:46ZClara Englercve@cve.cx2026-01-13T12:49:08Zurn:sha1:5bb2296c28fbb01238953264063ca1390752b4bc
Adding a comment suggesting to implement this as part of abstracting
`Sha256` behind a more generic identifier.