Make everything async
#26
Conversation
This reverts commit fdef833.
Here we migrate from wrapping all upstream errors to exposing only select error variants to the downstream user. Also, we clean up and unify logging outputs.
| } | ||
| } | ||
|
|
||
| pub(crate) fn set_runtime(&self, tokio_runtime: Arc<tokio::runtime::Runtime>) { |
There was a problem hiding this comment.
I now have taken the approach of setting and dropping the runtime on start()/stop(). However, a probably cleaner alternative approach would be to use tokio::Runtime::enter upon setup and just storing the EnterGuard somewhere. However, this may result in panicking if we'd end up calling tokio::spawn out-of-context. As I'm currently not 100% sure that this could never happen, I now went with this uglier approach for the time being, but will look into it again.
tnull
force-pushed
the
2022-10-make-everything-async
branch
from
623116a
to
bac34b5
Compare
src/access.rs
Outdated
| return fallback_fee; | ||
| } | ||
|
|
||
| locked_runtime.as_ref().unwrap().block_on(async { |
There was a problem hiding this comment.
As block_on seems to have quite a few drawbacks even in the multithreaded runtime case I'll switch all of these to using spawn+mspc channels.
There was a problem hiding this comment.
Got rid of most of them, one non-trivial left for the moment.
There was a problem hiding this comment.
Might be blocked on bitcoindevkit/bdk#795, otherwise we need to run all wallet operations in it's own thread and keep it block_on.
tnull
force-pushed
the
2022-10-make-everything-async
branch
from
bac34b5
to
ed49984
Compare
We migrate BDK and other remaining blocking parts to async.
tnull
force-pushed
the
2022-10-make-everything-async
branch
from
ed49984
to
a6174dc
Compare
|
Closing as it became a requirement to already include this in #11. |
Based on #9, #10, #11.
We migrate BDK and other remaining blocking parts to async.