Pierre Krieger
Pierre Krieger
Tweaks the trie cache test in order to test prefix removal operations. As I'm opening this PR, the test is failing, and it clearly indicates a bug in the trie...
It is legal to start nested transactions, in other words for example call `start_transaction` twice in a row then `commit_transaction` twice in a row. While the spec doesn't say that...
This concerns section C.5. The title is a bit weird, but when the `Metadata_metadata` runtime entry point is called, Substrate encodes the metadata, producing a `Vec`, then encodes that `Vec`...
I would suggest to remove child tries before they make it in production. Conceptually this feature doesn't bring any advantage over a single trie. It is overly complicated, with many...
When connecting a local Substrate node to a bootnode, the local Substrate node thinks that it has successfully peered with the bootnode, while in reality the bootnode has kicked out...
- [ ] ~~Make it possible for node operators to attach a TLS certificate to their libp2p WebSocket server. I believe this can be done by adding a CLI option...
To my surprise, this trait isn't implemented on channels, even though channels are kind of a critical component to unwind safety. See the code comment for an explanation of the...
Since Unix domain sockets are available in async-std, I think it is also in scope to add support for Windows named pipes. There already exists a crate on top on...
...and not the other way around.
This issue is mostly notes. Some IRC logs: https://botbot.me/mozilla/rustc/2017-01-11/?msg=79199876&page=1 Summary: - The GPU code would be written in Rust and embedded inside the regular CPU code. - There would be...