IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
281: Some Kademlia debugging r=koivunej a=ljedrz
These changes were sparked by the following 2 observations:
- the logs sometimes indicate that a Kademlia query was executed twice
- `finish_subscription` didn't always result in futures being awoken
While the former remains a mystery (its occurrence is not correlated with subscriptions, meaning it's either some polling issue that eludes me or a bug in `libp2p`), investigating the `kad`<>`Subscription` route yielded a few improvements, namely:
- tweaked `kad` log levels
- proper `SubscriptionFuture` handling in `put_block`
- improved `SubscriptionRegistry` logs and a small fix
- simpler `SubscriptionFuture` type handling (it always returned a `Result`, so now it's a default)
- a `debug_assert` checking that we don't trigger zero-wake cases in `finish_subscription` during tests
- improved `swarm_api` test (that sometimes caused issues with the new `debug_assert`, but could be improved regardless)
Co-authored-by: ljedrz <ljedrz@gmail.com>
290: Some maintenance r=koivunej a=ljedrz
Some maintenance work so as not to conflict too much with the pending PRs:
- depend on cid instead of libipld where only Cid is used
- update the `domain` dep (big `Cargo.lock` wins)
cc #75
Co-authored-by: ljedrz <ljedrz@gmail.com>
288: fix: update async-stream to 0.3 r=ljedrz a=koivunej
this uses the rust 1.45 features to remove recursion inside the macro. great for #284 which adds even more complicated async-stream(s).
good to go if CI passes.
Co-authored-by: Joonas Koivunen <joonas@equilibrium.co>
289: Ignore a flaky Kademlia test r=koivunej a=ljedrz
`#[ignore]` that Kademlia test that likes to randomly fail due to being dependent on actual network.
Co-authored-by: ljedrz <ljedrz@gmail.com>
283: Remove the rocksdb feature r=koivunej a=ljedrz
We no longer intend to pursue that feature; remove the related code.
Co-authored-by: ljedrz <ljedrz@gmail.com>
277: Add the means to connect by PeerId r=ljedrz a=ljedrz
I've encountered this while working on improved bitswap <> kademlia interop - we still can't easily connect using `PeerId`. Improve this by introducing an `enum ConnectionTarget` that is either a `Multiaddr` or a `PeerId` and adjusting the `SwarmApi.connect_registry` accordingly. There's also a new related test.
As a drive-by, adjust some `tracing` features that are among the defaults but we'd still enjoy having 😃.
Co-authored-by: ljedrz <ljedrz@gmail.com>
276: Remove unused dependencies r=koivunej a=c410-f3r
First commit only adds `default-features = false` where applicable while the second commit future-proofs possible upstream modifications for every single crate.
```
* bitswap
cargo build: 199 -> 179
cargo check: 212 -> 191
cargo test: 201 -> 181
* ipfs
cargo build: 419 -> 341
cargo check: 433 -> 354
cargo test: 445 -> 359
* ipfs-http
cargo build: 518 -> 408
cargo check: 556 -> 429
cargo test: 522 -> 412
```
By looking into the differences, it is possible to see astonishing results across all operations.
Co-authored-by: Caio <c410.f3r@gmail.com>
274: fix bogus empty bitswap messages r=ljedrz a=ljedrz
We've been seeing `(empty message)` bitswap for a while, but these logs were bogus - this PR fixes those, and adds a small bitswap stress test that can be used for future bitswap debugging.
Also removes some noisy logging and, as a drive-by, fixes a subscription-related log.
Co-authored-by: ljedrz <ljedrz@gmail.com>
Co-authored-by: ljedrz <ljedrz@users.noreply.github.com>
268: ci: caching updates r=aphelionz a=koivunej
node_modules: It's quite slow on windows (about 3min), linux (50s) and mac is in the between. This is a bit hacky given the patching and whatnot, but I think the hashFiles combination is good.
~/.cargo: this was never cached, but now it should be cached correctly `hashFiles(Cargo.lock)`.
Co-authored-by: Joonas Koivunen <joonas@equilibrium.co>
272: Split Entry into two different enums to remove unnecessary expects r=koivunej a=c410-f3r
Second take on #200
`Entry` is now composed of `Bucket(...)` and `Metadata(MetadataEntry { ... })` to separate things that have and doesn't have metadata, which avoids returning optional values derived from a single entry-point.
Next PR will address the newly introduced double matching while iterating over `continue_walk` but it will require deeper logical changes.
Co-authored-by: Caio <c410.f3r@gmail.com>
263: move bitswap Stats directly under Bitswap r=koivunej a=ljedrz
This makes the bitswap `Stats` persistent between peer disconnects.
In addition, remove the unused and no longer compatible `Ledger` tests.
Co-authored-by: ljedrz <ljedrz@gmail.com>
