The previous banning implementation didn't work, since it included port
numbers in the address value.
Here, we keep a separate table with (node, ip) pairs and update this
table when a node connects or is banned.
The timeout for a `UnixStream` will interrupt a `read` even when more data is
available. The preference would be that the timeout applies to each read, where
the timeout is reset again once a read is successful. This better supports
streaming semantics expected for the `Handle::subscribe` method.
To achieve this, the `LineIter` type holds the `UnixStream` and timeout
`Duration`. The `Iterator` implementation can then use `read_line` to progress
the stream one line at a time, and the timeout only applies to each read.
Co-authored-by: Alexis Sellier <alexis@radicle.xyz>
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Add a variant for upload-pack events to allow reporting from the upload-pack
process.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Move the events Emitter to the `radicle::node::events` module to allow it to be
used outside of `radicle-node`.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Instead of resetting the number of attempts on connection, reset it when
the connection is stable, ie. after some time has passed.
This prevents connection retries from happening in succession when a
peer is disconnected during the initial handshake, as is the case with
invalid subscribe timestamps.
If our inventory hasn't updated, don't issue a new inventory
announcement, simply use the previous message.
This avoids identical inventories propagating over the network and using
up resources.
Allow unknown fields to pass through when deserializing the `Project` payload.
This allows the deserialization to be more resilient when unknown fields are
added, e.g. from the CLI, and improves backwards-compatability when new fields
may be added in the future.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
The rate-limitter gets hit on connect, due to historical gossip messages
being exchanged. This results in a high initial burst that needs to be
accounted for.
Relay now has three values: "always", "never" and "auto".
The first two operate like the previous `true`/`false`. The
new (default) value turns relaying on for nodes with public (external)
addresses and off for nodes without. This should reduce redundant traffic on
the network, that was causing the rate limiters to trigger.
With the increase in traffic on the network after launch, the old
defaults don't make sense anymore.
Longer term, we should look into throttling outgoing messages as well,
prioritizing original messages over relayed messages.
* Try to use constants instead of strings
* Move crypto seed override out of radicle-crypto
* Use GIT_COMMITTER_DATE for commit overrides
* Use RAD_LOCAL_TIME as general time override
* Allow variables to be used in release mode
In radicle/src/test.rs, we have a helper type for a Node, which
creates a temporary directory using tempfile::tempdir, which creates a
TempDir. When the TempDir is dropped, it deletes its associated
directory. Before that happens, the Node has already stored the path.
Right after that the TempDir is dropped, and the directory deleted.
When the test code later opens the repository, using Storage::open,
it recreates the directory. As a result, there is nothing that would
delete the directory when the test finishes.
This means tests that use the helper Node type leave temporary files
behind.
Fix this by storing the TempDir in the helper Node, not just its path.
This means the TempDir is only dropped at the end of the test its used
in, and no temporary files are left behind.
Signed-off-by: Lars Wirzenius <liw@liw.fi>
Implement a new build pipeline using `podman` and `zig` that is
reproducible and can be run entirely on linux.
We also simplify the versioning system, defaulting to the output of `git
describe` when there are no exact tag matches.
If multiple ids are specified for `rad inbox clear`, it would report
that only 1 item was cleared, when in fact multiple were.
This was found to be due to the SQL code returning a change count of
`1`. This is likely due to the way the statement is being reset each
time. The `sqlite` library does not seem to easily support the usage
of `IN` in a `WHERE` clause. So instead, the `count` is aggregated in
the loop and returned instead.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
Implement this function on `RemoteRepository`, and move `repositories`
to the trait level.
This function will be used to quickly get the `RefsAt` of all repository
remotes.
There are two areas where we can be more tolerant of delegate namespaces
being missing or not validating:
1. Calculating the canonical HEAD
2. Fetching from a remote
In 1. the protocol is tolerant in that if the local node does not have
the default branch for a delegate, it will still attempt to use any of
the delegates it does have to reach the threshold.
This is made safe by ensuring that if the threshold is being updated
then the node performing the update must have a threshold of delegates
locally in their storage. It also made safe by 2.
In 2. the protocol is tolerant by allowing delegates to be missing
from the serving side, as long as they can still meet a threshold of
delegates. This is further tolerant, when validating the received
data, a threshold of delegates are valid to consider the fetch
successful -- otherwise it will fail.
Signed-off-by: Fintan Halpenny <fintan.halpenny@gmail.com>
X-Clacks-Overhead: GNU Terry Pratchett
When deciding whether or not to fetch a repo based on a refs
announcement, we need to know whether our signed refs are stale or not,
compared to the ones in the announcement. This can be quite expensive as
it requires multiple reads from Git's ODB, which often results in
packfile loading.
To remedy this, we introduce a refs cache in our node database that is
consulted when an announcement is received.
We also remove the logic that generates `info` messages, since they are
rarely needed and introduced more complexity.
Additionally, we revert commit 3ad2b4431f,
which introduced logic to minimze wants and haves with additional logic,
since it potentially also increases the load on the ODB.
Instead of dropping the existing connection(s), we drop the same
connection on both sides, by using a simple rule that yields the same
result whether the node sees the connection as inbound or outbound.
The trick is to use something that both nodes agree on, for instance
whos public key is greater than the other, and use that information
to close the same connection on both sides.
cloudhead
Fintan Halpenny
Dr Maxim Orlovsky
Lars Wirzenius