- Assert that with the peer still `Connected`, `worker_result` clears
`active[rid]` as expected
- Confirm the disconnect test is exercising the discard path rather than
passing vacuously
- Drive `worker_result` through the real wire layer with the peer in
`Disconnecting` state, mirroring a fetch that ends mid-disconnect
- Assert `active[rid]` survives, proving the result is discarded without
calling `service.fetched`, which is what orphans the entry
- Add helpers to wrap a service with an active fetch in a `Wire`
- Drive the real cancel-skip path: a connection conflict makes
`disconnected` early-return on a link mismatch, leaving the fetch
uncancelled
- Combined with an undelivered fetch result, the `active[rid]` entry is
orphaned and the repo can no longer be fetched from any node
- Confirm delivering the missing result clears the entry and unblocks
the queued fetch, showing what the fix must guarantee
A fetch result was discarded whenever the peer was no longer `Connected`
by the time the worker reported, and queued fetches were dropped silently
when the peer disconnected before the `Io::Fetch` was processed. In both
cases the fetcher's `active` entry for the repo was never cleared. Since
that entry is keyed by repo, it blocked the repository from being fetched
from any node until the process restarted.
- Report the result in `worker_result` even when the peer is no longer
connected, instead of returning early
- Report a failed fetch when an `Io::Fetch` is dropped for a disconnected
peer, so the active entry is cleared
- Flip the wire regression test to assert the entry is now cleared on
disconnect
When a connection drops and the node attempts to reconnect within the
same second, the `last_attempt` timestamp written to the database is
identitcal to the `last_success` timestamp due to the Unix time (seconds)
precision loss.
The `maintain_connections` logic filters eligible peers using the
condition `last_success >= last_attempt`. Because the truncated
timestamps are equal, this evaluates to true, causing the node to
bypass the `CONNECTION_RETRY_DELTA` backoff.
If the OS instantly rejects the dial attempt (e.g. `EHOSTUNREACH`),
the node synchronously queues another connection attempt. This traps
the reactor in an infinite `while` loop.
Changing the condition to `last_success > last_attempt` ensures that a
failed attempt in the same second correctly evaluates to false, forcing
the node to respect the backoff timer and breaking the loop.
Add custom check for ellipses "...", asking for replacement of "…".
Git ranges and the CLI wildcard matches are ignored.
This change includes fixes to all sites that did not pass the check.
The invalid timestamp error will only tell the timestamp of the
offending node's message.
Currently, there are issues where that error is occurring but the
timestamp seems legitimate. This patch adds the running node's
timestamp to help debug the issue.
Move logging-related module `radicle::logging` into its own crate.
While at it, remove the "logger" feature flag from `radicle`.
Co-authored-by: Lorenz Leutgeb <lorenz.leutgeb@radicle.dev>
`cargo check` fails because `human-panic` and `sysinfo` require at least
1.88.0.
+++ command cargo check --release --locked --all-targets
error: rustc 1.85.0 is not supported by the following packages:
human-panic@2.0.6 requires rustc 1.88
sysinfo@0.37.2 requires rustc 1.88
Bump MSRV to fix this.
1.88.0 introduced [let chains], which in turn has `clippy` warn about
nested if statements. All of these sites are fixed in this change.
1.87.0 introduced [`is_multiple_of`], which is a more readable version
of `x % y == 0`.
[let chains]: https://blog.rust-lang.org/2025/06/26/Rust-1.88.0/#let-chains
[`is_multiple_of`]: https://doc.rust-lang.org/std/primitive.usize.html#method.is_multiple_of
In `cb57e6560fe7767e79b87473a11261fbb82ce84c` the loading of systemd
credentials was refactored. The idea was to only change the accepted IDs
slightly after switching from radicle.xyz to radicle.dev. Unfortunately,
this was done in a sloppy way. The two cases for passphrase and secret
actually differ in ways that was overlooked.
Rectify these mistakes and restore the old behavior:
1. The logic to read the passphrase from the file that contains the
credential is moved out of `load_credential` since it only applies
to the passphrase, not the secret.
2. The warnings printed are now generic, and will not mention the
passphrase in case reading the path for the secret fails.
- Fixes a bug where the log level set in the config file was
ignored: `Logger` and `StderrLogger` captured the level in a
`self.level` field at construction time and checked it in
`Log::enabled`. After config was loaded, the global
`log::set_max_level` was updated but `self.level` was not,
so verbose messages were dropped by the per-instance filter
even when the global filter allowed them.
- Make `log::set_max_level` the single source of truth: remove
the `level` field and have `Log::enabled` defer to
`log::max_level()` so the two filters can no longer drift.
- Update call sites to construct loggers without a level.
- Disable the structured logger's internal filter (set to "trace") so
that it also falls back to `log::set_max_level`.
Canonical references can not be used to model symbolic references.
Relax this restriction by adding another member to the payload, named
"symbolic". Its key-value/name-target pairs then translate directly to
canonical symbolic references.
Care is taken to not allow circular references, and that there always
is a rule that could generate the target of a symbolic reference
(or a chain of symbolic references). Still, a symbolic reference may
dangle (for example when its target reference cannot be computed
because of divergence), but at least it can be prevented that a
symref *always* dangles, because there is no rule that would produce
its target.
Co-authored-by: Fintan Halpenny <fintan.halpenny@gmail.com>
Previously `test_fetch_emits_canonical_ref_update` was passing by
catching a leftover `CanonicalRefUpdated` event from the initial
repository fetch. Furthermore, the new commit was written to the
un-namespaced root. It can be proven to be the case by commenting out
the `alice.commit_to(...)` line and watching the test still pass.
Fix the test by committing to the namespaced default branch which will
trigger the `CanonicalRefUpdated` that is expected by Bob. Drains events
so we don't catch the wrong event, and explicitly announces.
The `commit_to` method should namespace the reference, so that node is
only ever committing the namespace of that particular node.
This prevents the node from committing to another node's namespace or
the canonical namespace.
Prevent `rad` from panicking when its stdout is a closed pipe, e.g.
when running `rad config | head`.
The broken pipes are handled at the output boundaries using three
layers of defence:
1. Inner: Route all stdout output through `term::print()` and
`term::print_inline()`, which use explicit `writeln!`/`write!` to
a locked stdout and silently ignore write errors. All `println!`
calls in `radicle-cli` and `radicle-term` are converted.
2. Middle: Catch `BrokenPipe` errors that propagate up via `anyhow`
from subcommands in `main::run()`, and exit cleanly with code 0.
3. Outer: Install a panic hook that intercepts `println!` panics
containing "Broken pipe" (from dependencies like clap) and exits
cleanly, chained in front of human_panic's hook.
To prevent regressions, `#![deny(clippy::print_stdout)]` is added to
`radicle-cli`, requiring all future stdout output to go through the
safe `term::` functions.
Note: `eprintln!` calls (stderr) are not handled, as broken stderr
pipes are extremely rare in practice.
`fn sha1_zero` can be `const fn`, and since it does not take arguments,
it can just be a `const`. There is no good reason to not have a constant.
Justification for changing the order to first specify "zero" and then
"SHA1": In the future we will support multiple object formats. A
function that produces a zero OID would be parameterized by the object
format then. Imagine `fn zero(format: ObjectFormat)` which would be
called as `Oid::zero(ObjectForma::Sha1)`. In anticipation, we change the
ordering to match.
Because we changed our main domain from radicle.xyz to radicle.dev, we
should also change our systemd credential identifiers.
Handling is move into a new function, which also checks the old names
for backwards compatibility.
With `connect: std::collections::HashSet`, we suffer unpredictable reordering
of values in our test cases.
Use `connect: indexmap::IndexSet` instead, which preserves insertion
order for iteration.
Fixes an intermittent failure caused by a race condition with background
gossip messages.
Previously Eve's `RefsAnnounced` message could be processed by Bob
after the test updated the identity document to include Alice-Laptop
as a delegate. This caused Bob to unexpectedly fetch the issue in the
background, failing the `!has_issue` assertion.
We now explicitly wait for Bob to receive Eve's announcement before
updating the identity document.
Avoid relying on a thread sleep to wait for the nodes to connect.
Instead rely on the connection condition being established by using a polling loop.
The loop is controlled by a number of iterations, panicking if the
number of iterations is exceeded.
Addresses intermittent failures in the `test_connection_crossing` e2e test,
which were particularly prevalent on slower CI environments (such as
`rust:trixie`).
Previously, the test spawned two threads to make Alice and Bob dial each
other concurrently and strictly asserted that the "preferred" peer (the
one with the higher Node ID) would always win the `Outbound` link direction.
However, this assumption is flawed in real-world, OS-level network execution
due to two race conditions:
1. Thread Scheduling: One thread could execute and fully establish a
connection before the other thread even began processing its dial command.
2. Reactor Event Ordering: A node's reactor might wake up and process an
incoming TCP connection from its peer *before* it processes the `Connect`
command sent by the test. When it finally processes the `Connect` command,
it sees a session already exists and skips dialing entirely.
In both scenarios, a true "simultaneous crossing" never occurs. Instead, a
standard sequential connection happens, meaning the link direction is dictated
by whoever dialed first, not by the "preferred" peer logic. This caused the
strict `left: Outbound, right: Inbound` assertions to panic.
To fix this, two changes were made:
- Introduced a `std::sync::Barrier` to synchronize the two test threads.
This forces both threads to wait for each other before calling `.connect()`,
maximizing the probability of a true simultaneous dial.
- Relaxed the final assertions. Because OS-level TCP handshakes and reactor
polling can never guarantee perfect simultaneity, we no longer assert
*which* peer gets the `Outbound` link. Instead, we assert the core invariant:
that exactly one connection is established between the nodes, and that their
link directions are opposite (`s1.link != s2.link`).
The main change that needs adjustment on our side is that
`impl EcPk for ec25519::PublicKey` changed from `Compressed = [u8; 32]`
to `Compressed = amplify::Bytes32`.
Also remove `impl Deref for PublicKey` to avoid dependents of
`radicle-crypto` depend on a particular implementation, and make the
anonymous member private.
See
<f42396139a>.
The value of `RAD_SOCKET` influences the outcome of `cargo test`.
Users that set this environment variable to, e.g. point `rad` at
a socket at a custom location, were surprised that execution of
tests would interact with their node.
Split `Home::socket` into two functions. `Home::socket_default` is
"pure" and only dependent on the path represented by `Home`. On the
other hand, `Home::socket_from_env` interprets the environment.
Most users will want to use `Home::socket_from_env`.
There is only one type that can really construct canonical references,
which is the the identity document. Remove the corresponding method
from the trait and rename it to `GetRawCanonicalRefs` accordingly.
Separate module for the concern of protecting `refs/rad`.
Note that this also fixes a bug, as previously all refs
*starting with* `refs/rad`, such as `refs/radieschen` were protected.
The method `set_head` does two things:
1. Compute the canonical head and set the default branch to target.
2. Set the symbolic reference `HEAD` to target the default branch.
Split these two concerns into:
1. `set_default_branch_to_canonical_head`
2. `set_head_to_default_branch`
The main motivation behind this change is to get just a little
telemetry information from nodes on the network, namely the version
of Radicle they are running.
This is achieved by rewriting `impl Default for UserAgent` which now
uses the version information provided by the build script also used
in other crates.
Also, a new configuration option `node.userAgent` is added, which
allows users to override the user agent if they so please, or set
the value `null`, which will in turn send the user agent
`/radicle/`, which is not really helpful, and the default prior to
this commit.
Creations of `UserAgent` in the whole workspace is cleaned up. In
order to do that `UserAgent::test` is introduced.
Instead of passing the signer as an argument to many methods on `Store`,
scope the `Store` itself to a signer.
This further allows to differentiate two different access modes on the
store in `radicle::cob::store::access`: `WriteAs` (which requires signer)
and `ReadOnly` (which does not require a signer).
The caches for issues and patches in `radicle::cob::{issue,patch}::Cache`
are concretised by removing the first type parameter, since it was
specific to issues and patches anyway. This was done in this commit as
it touches very similar usage sites.
Make `Device` less prominent, and instead lean more heavily towards
traits from the `signature` crate, such as `Keypair` and `Verifier`
in addition to `Signer`. Trait bounds regarding `Signer` could be
simplified, but this is left for the future.
In `radicle-cli`, the function `term::cob::patches_mut`, which generates
errors with a hint is used instead of the lower-level `Profile::patches_mut`.
Commands `rad issue cache` and `rad patch cache` now construct a
writeable cache on top of a read-only store.
Many knock-on changes are handled as well, to arrive at a clean state.
The intent to drop outgoing connections is modeled as `Option::None`
which is brittle and easy to miss.
Extend `enum AddressConfig` with a variant that is more explicit.
Teach the `rad sync` and `rad clone` commands to accept the
`--signed-refs-feature-level` option. This option configures that
fetch to use the minimum feature level provided when fetching from
other nodes. This overrides the value of
`node.fetch.signedReferences.featureLevel.minimum`, and should only be
used in scenarios where it is necessary to downgrade the security of a
fetch for backwards compatibility.
To achieve this, the `Handle::fetch` method introduces an optional
parameter for threading through the `FeatureLevel`.
Daniel Norman
Adrian Duke
Fintan Halpenny
Lorenz Leutgeb
Lorenz Leutgeb
Josh Soref