Matthias Urlichs

Well, questions along the lines of "I've started that other thread from inside `trio.run` so why isn't this working" come to mind.

Well, the connection *may* be usable when we get it out of the pool and the socket is not readable, but that's not a guarantee. I once had a cute...

Quick fix for this: ```diff diff --git cq_editor/widgets/traceback_viewer.py cq_editor/widgets/traceback_viewer.py index d5e0baa..4b1043d 100644 --- cq_editor/widgets/traceback_viewer.py +++ cq_editor/widgets/traceback_viewer.py @@ -56,6 +56,6 @@ class TracebackPane(QWidget,ComponentMixin): root = self.tree.root code = code.splitlines() - tb...

I opened PR #431 with a better fix that doesn't suppress tracebacks in called modules. They're too helpful for finding the actual problem …

Hmm. Flask used to include code so that flask.ext.* is actually loaded from `flask_*`or `flaskext.*`. Presumably you could do something similar. See https://github.com/mitsuhiko/flask/commit/715a9a3e7 where this code was removed (because it...

GOAWAY helps clients when the server decides it needs to gracefully shut down the connection (server restart, load balancing). The client can then transparently open another socket and use that...

> on it's own Might you be persuaded to fix this issue's title? The non-grammatical apostrophe is somewhat grating IMO.

While the `StackSummary` contains a `locals` attribute, it defaults to being empty unless you specifically tell the `TracebackException` that `capture_locals=True`. That being said, building and storing a stack traceback in...

> AnyIO uses the stdout/stderr of the worker processes for internal purposes to communicate with the parent process IMHO it does no such thing, and frankly I don't know why...

> if it's common enough to be a recipe, then it's common enough to implement once and get right The problem is: What is "get right"? In many cases you...