Ben Darnell

OK, I think I've got a minimal reproduction. Try to import class Bar from FooApp.java in https://www.dropbox.com/s/0ik6plgndqgvi38/malabar-report.tar.gz

Yes, of course I'd be willing to donate this test case to the project. I hadn't noticed that the individual pom files weren't enough to compile on their own (or...

Fixed by #3471

First, thank you for the thoughtful contribution. I'm not opposed to making it easier to work with synchronous code and threads from Tornado, but it's a very tricky area because...

For the record, this is still occurring with language server version 2027.7.1 (`Pylance language server 2024.7.1 (pyright version 1.1.369, commit 1168be71) starting`). However, it's moved around a bit. I no...

What exactly are you trying to do here? This looks the same as the chunking in `_read_fixed_body`. (and remember that all new functionality needs tests, and comments and commit messages...

Without a Content-Length or Transfer-Encoding header, this is an HTTP/1.0-style read-until-close response, so it is handled by `_read_body_until_close`. This is discouraged in HTTP/1.1 but allowed for backwards compatibility. Special-casing the...

This change looks good, but it needs a test and should be rebased into a single commit with comments in english.

There are a few likely possibilities here: * Some long-running operation is blocking the event loop thread. Use [asyncio debug mode](https://docs.python.org/3/library/asyncio-dev.html) to investigate this possibility * Poor load balancing. I've...

> During the stress test, I observed occasional callbacks in the DEBUG log that took more than 100 milliseconds. These messages that say "poll %.3f ms took %.3f ms" are...