| Commit message (Collapse) | Author | Age | Files | Lines |
| |\
| |
| |
| |
| |
| |
| | |
Implement proposal 360 (limit HsDesc inflation opportunities)
Closes #2046
See merge request tpo/core/arti!3070
|
| | | |
|
| | |
| |
| |
| | |
(maxima are as amended in torspec!411)
|
| |/
|
|
| |
Made with https://crates.io/crates/typos-cli
|
| |
|
|
| |
See #2060.
|
| |
|
|
|
|
|
| |
On one hand, this can be quite noisy when something *is* out of range.
On the other hand, a clamped value shouldn't normally happen, and is
likely to result in surprising and otherwise difficult-to-debug
behavior.
|
| | |
|
| |\
| |
| |
| |
| | |
tor-netdir: decrease minimum hsdir_interval from 30 to 5
See merge request tpo/core/arti!3042
|
| | |
| |
| |
| |
| |
| |
| |
| | |
This matches the spec change:
<https://gitlab.torproject.org/tpo/core/torspec/-/merge_requests/406>
Progress on:
<https://gitlab.torproject.org/tpo/core/chutney/-/issues/40038>
|
| |/
|
|
|
|
| |
We will construct this object based on the circuit parameters _and_
on the target's supported protocol versions, so we need to do so
when we have both pieces of info.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously we logged a warning whenever `choose_multiple_weighted`
returns `Ok(iter)`, whenever `iter` contains fewer relays than
requested.
However, this is expected behavior of `choose_multiple_weighted` if the
filtered list of relays doesn't have the requested number of relays, in
which case it should return all of the relays in the filtered list.
In particular, this condition occurs regularly in test networks when
looking for candidate relays to add to the potential guard set while
excluding those already in the potential guard set.
We still warn if the returned list is fewer than requested, *and* fewer
than available in the filtered list.
|
| |
|
|
|
|
| |
Since we don't know what kind of traffic we'll use a rendezvous
point for, we don't want to use it if it isn't "Fast" (reasonably
high bw) and "Stable" (unlikely to crash soon).
|
| |
|
|
|
|
| |
> The actual impact of this patch is to prevent usage of MiddleOnly
> relays as Introduction Points. The Rendezvous Point logic isn't
> hooked up yet - nick
|
| | |
|
| |
|
|
|
|
| |
Now that choose_multiple_weighted can once more return
fewer items than requested, we should warn when it does so,
to be consistent with our other cases of warning about #1907.
|
| |
|
|
|
|
|
|
| |
Rand 0.9.1 reverted the changes that caused #1903,
in https://github.com/rust-random/rand/pull/1623.
This broke our test that made sure that rand wasn't going to change
their behavior _again_ :)
|
| |\
| |
| |
| |
| | |
various crates: MSRV TODO standardization and cleanup of an old TODO
See merge request tpo/core/arti!2945
|
| | |
| |
| |
| |
| | |
- https://github.com/rust-lang/rust-clippy/issues/11764 was fixed upstream, so
this is no longer needed.
|
| | |
| |
| |
| |
| | |
This name reflects its purpose better than the original one,
since it includes required protocols as well as recommended ones.
|
| | | |
|
| | | |
|
| |/ |
|
| |
|
|
|
| |
Instead of doing the more sensible thing when every relay has weight
0, give a warning message.
|
| | |
|
| |
|
|
|
|
| |
Here we test the (arguably inconsistent behavior) where can we
return fewer items than requested if no item had zero weight.
(If this ever becomes an error, we need to modify the calling code.)
|
| |
|
|
|
|
| |
WeightedError::InsufficientNonzero is expected under some
circumstances, but the other instances would mean that we have a
bug.
|
| |
|
|
|
| |
We want to make sure that behavior for 0-weighted elements is
consistent for choose_multiple_weighted and choose_weighted.
|
| |
|
|
| |
Possible fix for #1902.
|
| |\
| |
| |
| |
| |
| |
| | |
protover: Add support for subprotocol version mnemonics.
Closes #1891
See merge request tpo/core/arti!2854
|
| | | |
|
| | |
| |
| |
| | |
- The Rng::gen() functions have been renamed to Rng::random().
|
| | |
| |
| |
| | |
- Several methods have been moved out of SliceRandom.
|
| |/
|
|
| |
- The rand::distributions module has been renamed to rand::distr
|
| |
|
|
|
|
|
| |
This implements the client side of proposal 321.
It's a breaking change in netdir, since in_same_family now takes an
extra argument.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
When we're trying to exclude relays by family,
we need to know which lists to look at.
This information ultimately comes from the network parameters.
We could avoid this change if we just told clients
"look at all family information all the time",
but that's not what the proposal says.
This is a breaking change.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Related #534
Signed-off-by: David Goulet <[email protected]>
|
| |
|
|
|
|
| |
These are tentative, so I haven't added them to param-spec yet.
Part of #1729
|
| |
|
|
|
|
| |
Denies 'mod.rs' files for consistency.
https://rust-lang.github.io/rust-clippy/master/index.html#mod_module_files
|
| |
|
|
| |
These are new in Rust 1.83.
|
| |
|
|
|
|
|
|
| |
In 1.83, this warning triggers on many of our crates.
We're thinking of fixing them all, but for now,
we're going to disable the warning.
This is part of #1765.
|
| | |
|
| |
|
|
|
| |
These helpers seem potentially broadly useful, and only really
discoverable if they're here.
|
| |
|
|
| |
netdirs (fmt).
|
| |
|
|
|
| |
This allows us to set SRVs for example (needed because by default, the
test `NetDir` is built from a consensus that doesn't contain any SRVs).
|
| |
|
|
|
| |
This warning complains when we say `where T: SomeTrait + ?Sized`
when `SomeTrait` is inherently Sized.
|
| |
|
|
| |
This commit is automatically generated.
|
| |
|
|
|
|
| |
The old code produced a warning from clippy nightly; we may as well
update to use the new associated consts. (They've been there since
Rust 1.4x.)
|
| |\
| |
| |
| |
| |
| |
| | |
tor-guardmgr: Make lite vanguards work
Closes #1275 and #1340
See merge request tpo/core/arti!2075
|