l

`futures` should probably be replaced with `futures-lite` if possible, because `futures-lite` is a dependency of `async-std` anyway and `futures-lite` is much smaller and easier to review. But it's a different...

After move from async-std to tokio we only depend on `futures-lite` directly, so closing this. `futures-lite` cannot be replaced with `futures` easily because it has some nice APIs like `.race()`...

node bindings are integrated into core repo. It's easier for me to clone deltachat-desktop repo, point deltachat core dependency to deltachat-core-rust worktree and build it once, then to test a...

It's also not clear what to do about the nonstandard footer appended by the mailing list to the chat message: it is currently displayed as `[...]`.

cc @adbenitez

> the context is a very basic object, more regarded as "the app" (logging, stockstrings, database) and not only a database Logging is an internal thing, all the UI sees...

> So, I tentatively suggest we attempt to adopt this? Could we convince a UI to try this out before merging? cc @Hocuri @Jikstra @Simon-Laux for someone who can try...

Another way to fix this in backward-compatible way is to guarantee that context is never actually deallocated until you drain the queue and finish processing the last event. So `Event`,...

A simple way to fix this: add thread-safe `dc_event_emitter_close(dc_event_emitter_t* emitter)`. This will make event emitter to start returning `NULL` afterwards. To terminate event loop from any thread, you close the...

I opened a draft PR at #3285