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.
391: Move the README example, remove the nightly feature and CI run r=koivunej a=ljedrz
Moving the code example from the README to the `examples` simplifies our CI setup and makes one of the runs redundant.
Co-authored-by: ljedrz <ljedrz@gmail.com>
382: return errors for some extra DHT issues r=ljedrz a=ljedrz
Return explicit subscription errors on some of the more user-facing DHT requests, namely `Bootstrap`, `StartProviding` and `PutRecord`; `Kademlia` errors for those are currently only logged and `KadResult::Complete` is returned, potentially concealing issues.
Co-authored-by: ljedrz <ljedrz@gmail.com>
390: Example tuneup r=koivunej a=koivunej
Removes the examples:
- client1 and client2
- ipfs_bitswap_test
Also removes the stray file `examples/block.data`. It used to be used by something.
I wrote a longer one but the short gist is that fetch_and_cat has superceded both, except for the ipld stuff, which will now have zero examples but there are tests. It's better to have one clear example than to have multiple confusing and out of date ones.
Co-authored-by: Joonas Koivunen <joonas@equilibrium.co>
Co-authored-by: Joonas Koivunen <joonas.koivunen@gmail.com>
client1 and client2 were a pair of two peers working together to form up
an example where one client creates an ipld document and the other one
downloads it, reading IPLD projections. the bitswap side is already done
elsewhere and the nature of use having to run two clients, there just
wasn't an easy way around without running the two nodes in the same
process. if that was done, then the example would be the same as
tests.
383: remove the default random listening address r=ljedrz a=ljedrz
Remove the default random listening address, listen to no addresses by default.
Cc https://github.com/rs-ipfs/rust-ipfs/issues/356
Co-authored-by: ljedrz <ljedrz@gmail.com>
387: fix: restore bootstrappers from static list, update the bootstrapper list r=koivunej a=ljedrz
Alters the `BOOTSTRAP_NODES` list to contain only the currently working addresses and changes the behavior of `Swarm::remove_bootstrappers` to not use the config file, but rather the aforementioned in-memory list.
Cc https://github.com/rs-ipfs/rust-ipfs/issues/356
Co-authored-by: ljedrz <ljedrz@gmail.com>
388: fix: unhandled panics in the fs blockstore r=koivunej a=koivunej
Fixes#386 -- at least there are no more unhandled panics, seems that the blockstore is still racy but I am not sure how to fix it right now. I am not 100% what caused those panics but I suspect it was related to having multiple broadcast::Sender's alive in the blocking task queue trying to send to receivers which were already dropped because the outer async context was dropped.
In addition there are even more comments, minor comment adjustments, minor splitting to comment less but more clear code.
First commit is the fix.
Co-authored-by: Joonas Koivunen <joonas@equilibrium.co>
Co-authored-by: Joonas Koivunen <joonas.koivunen@gmail.com>
384: remove unimplemented ipns functionalities from Ipfs r=ljedrz a=ljedrz
Remove ipns insertion and cancellation features from `ipfs::Ipfs` (they are not implemented).
Cc https://github.com/rs-ipfs/rust-ipfs/issues/356
Co-authored-by: ljedrz <ljedrz@gmail.com>
See related issue #386, the panic reports were popping up on GHA builds.
The reason why this fixes the issue is not entirely clear, even in
hindsight moving the tx over to the blocking task did seem like the
right thing to do, and it could had enabled other ways to cleanup the
synchronization as well.
However perhaps it is not more explcit that the synchronization key is
created and cleared in the async context.
379: A regular maintenance pass r=aphelionz a=ljedrz
The usual `cargo update` run plus the removal of an unneeded direct dep and a dead test.
Co-authored-by: ljedrz <ljedrz@gmail.com>
377: Enabling pinned blocks test r=ljedrz a=aphelionz
This PR un-skips one conformance test that works now that we have pinning. This brings our number up by one! 🧨
Co-authored-by: Mark Henderson <mark@mrh.io>
375: A fistful of tweaks r=aphelionz a=ljedrz
A pretty wide range of test improvements, small fixes, refactorings and renames. Each commit describes the specific change, but I'll list the more noteworthy ones:
- automatically remove `Protocol::P2p(_)` from the `Multiaddr`s passed to `Ipfs::add_peer`
- only produce a `RepoEvent::RemovedBlock` when a block was actually removed
- add unit tests for the `Node` topologies
Co-authored-by: ljedrz <ljedrz@gmail.com>
376: update README, remove old links r=aphelionz a=ljedrz
Self-explanatory: just a few updates to the README, plus the removal of old issue links.
Co-authored-by: ljedrz <ljedrz@gmail.com>
