| Commit message (Collapse) | Author | Age | Files | Lines |
| ... | |
| | | | | | | |
|
| | | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This is part of #2657. The goal here is to use DirMgr as a
temporary backend for tor-dirserver, so that we can have a
sort-of-working directory cache that can be used for testing guards
before tor-dirserver is complete.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit removes the getters in database.rs. If the overall
tor-netdoc refactoring has taught us something, then that the use of
getters can be quite annoying.
Therefore, this commit removes the getters and marks the respective
fields as pub. Given that these data types are pub(crate) only right
now, it does not matter a lot anyways.
|
| |\ \ \ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
arti: Fix compilation of http-connect without rpc
Closes #2653
See merge request tpo/core/arti!4296
|
| | |/ / / /
| | | | |
| | | | |
| | | | | |
Closes #2653.
|
| |/ / / / |
|
| | | | |
| | | |
| | | |
| | | | |
Needed for tor-dirserver testing.
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
ci,maint: Update container image versions
Closes #2618
See merge request tpo/core/arti!4280
|
| | |/ / / |
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Use an IPv6 loopback address instead of the unspecified address when
creating test servers. The address returned by binding to [::]:0 is not
valid as a connection target on OpenBSD.
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
tor-proto: Upgrade statrs dev-dependency and remove default features
See merge request tpo/core/arti!4283
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This doesn't appear to be needed, and removes an nalgebra dev-dependency
which should help build times.
|
| | |/ / /
| | | |
| | | |
| | | |
| | | |
| | | | |
```bash
cargo upgrade --incompatible allow -p statrs
```
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
test(dirclient): invalid utf-8 bytes in response body
See merge request tpo/core/arti!4277
|
| | | | | | |
|
| |\ \ \ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
tor-netdoc: Fix left-over incompletes
See merge request tpo/core/arti!4273
|
| | |/ / / /
| | | | |
| | | | |
| | | | |
| | | | | |
This removes two left over incomplete feature flags.
Follow-up to c20df8c5f879806f7044c6e26d9e69404c88cfa3.
|
| |\ \ \ \ \
| |_|_|_|/
|/| | | |
| | | | |
| | | | | |
tor-netdoc: Remove outdated top-level RouterDesc comment
See merge request tpo/core/arti!4274
|
| | |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This feature is no longer present and we now always compile with
router descriptors enabled, deferring the optimization to the
Rust compiler and linker respectively.
Follow-up to arti!3972
|
| | |/ /
|/| |
| | | |
This reverts merge request !4264
|
| |\ \ \
| | | |
| | | |
| | | |
| | | | |
Upgrade to weak-table 0.4.0
See merge request tpo/core/arti!4278
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Weak-table 0.4.0 replaces the backend with hashbrown for improved
performance and correctness. It adds a bunch of APIs for parity
with std::collections, and deprecates the risky `ahash` feature.
(I'm the maintainer at this point.)
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
Make RouterDesc item docs consistent
See merge request tpo/core/arti!4279
|
| | | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Right now, the documentation for the items is inconsistent in the sense
that some items contain specification links whereas others contain
copies of the specification.
This has historic reasons during the development, but ultimately links
to the specification are superior, as they provide up-to-date
information and are the canonical source.
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Disintegrate types::misc::routerdesc
Closes #2555
See merge request tpo/core/arti!4276
|
| | | | | | |
|
| | | | | |
| | | | |
| | | | |
| | | | | |
No functional change, purely formatting.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit moves the remainder of misc::routerdesc into a new module
called types::descriptor, which contains the shared types needed in
router descriptors as well as extra-info documents.
Please review with --color-moved --color-moved-ws=allow-indentation-change,
as this is mainly code motion, except for parts where inputs and type
aliases had to be fixed.
|
| | |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This commit moves all non-signature related parts from
types::misc::routerdesc into doc::routerdesc, because they are only
required for that specific netdoc.
The signature related parts will be moved in the next commits, as we
will need those for extra info documents as well.
Review with --color-moved --color-moved-ws=allow-indentation-change.
This is mostly code motion with small import/type name fixes being the
notable exception.
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
tor-dirauth: Calculate microdescriptors
See merge request tpo/core/arti!4264
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
| | | | | |
| | | | |
| | | | |
| | | | | |
This is actually quite simple now.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
It doesn't actually use the value, but its call site is going to have
one of these, not a SupportedConsensusMethod.
|
| | | | | | |
|
| | |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
No longer request "experimental" - we don't need it.
We do need "testing" but only for our own tests, so that can go in
dev-dependencies. Enable retain-unknown, which we'll need.
|
| |\ \ \ \
| |_|/ /
|/| | |
| | | |
| | | |
| | | |
| | | | |
tor-proto: Fix cc vegas counter updates on clock stall
Closes #2651
See merge request tpo/core/arti!4256
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The pseudocode says:
```text
if next_cc_event:
next_cc_event--
if next_cwnd_event:
next_cwnd_event--
if clock_stalled_or_jumped:
inflight -= 'cc_sendme_inc'
return
```
but we were previously decrementing these after the return rather than
before.
|
| | | | |
| | | |
| | | |
| | | | |
We now have `add_relay_ent()` for this, so we can remove the TODO.
|
| | |/ /
|/| |
| | |
| | |
| | | |
I think this might have been c&p from the half-stream code ("hs" =
"half-stream"). "hc" ("half-circuit") seems like a better name.
|
| |\ \ \
| | | |
| | | |
| | | |
| | | | |
tor-netdoc: Use TimeRangeBoundBuilder for verification
See merge request tpo/core/arti!4265
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | |
| | | | |
No functional change, please review with
--minimal --color-moved --color-moved-ws=allow-indentation-change.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This commit utilizes TimeRangeBoundBuilder for the verification of
router descriptors by replacing all calls of
`.dangerously_unwrap_timely()` with `.unwrap_with()`, while also
wrapping the actual logic inside a closure which is then called by
`TimeRangeBound::build_intersect()` in the end.
Keep in mind, that the replacement also leads to the removal of the
use of tuples, as `.unwrap_with()` only returns the actual item.
|
| | | | | |
|
| |\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
arti-client: Update release date
See merge request tpo/core/arti!4269
|