Roman Bogorodskiy

This is still actual. CDATA section might contain invalid XML that results in issues with Jenkins for example: ``` org.dom4j.DocumentException: Error on line 2387 of document : An invalid XML...

I have workarounds for the build time issues, but probably it doesn't make a lot of sense to send patches before I can get it working. Right now I'm not...

I noticed that trio fails to poll even on files like `/dev/random` or `/dev/null`, while kqueue generally works fine with that: https://github.com/python-trio/trio/issues/1050 Also, when I try to run some pyfuse...

Yeah, I've changed port from 4444 to 4445 and it doesn't fail.

Funny thing: when I was rechecking this today, I tried the unchanged version first and got some failures. Now I'm getting back to it and I see that each failure...

I decided to check if I could see something similar if I libzmq tests. There's one test that always fails: ``` (lldb) target create "/usr/ports/net/libzmq4/work/zeromq-4.2.2/tests/.libs/test_unbind_wildcard" --core "/tmp/test_unbind_wildcar_37479_0.core" Core file '/tmp/test_unbind_wildcar_37479_0.core'...

I've updated libzmq to 4.2.3 and test_cffi_backend.py is not failing anymore. However, there's a new failing test: `zmq/tests/test_auth.py::TestThreadAuthentication::test_null`. Failure rate is if not 100% then very close to it. ```...

Hm, it doesn't seem the problem is related to the Python version, it looks like formatting was changed again.

Is there any way to obtain information from the driver on what it fails on specifically to understand what's wrong on the hypervisor side.

Thanks for the pointer, I was able to get tracing messages: ``` [0]0004.018C::10/22/2025-10:38:03.405 [vioinput.sys]Virtio-input driver started...built on Jul 9 2025 07:24:09 [1]0004.018C::10/22/2025-10:38:03.405 [vioinput.sys] VIOInputEvtDeviceAdd [1]0004.018C::10/22/2025-10:38:03.405 [vioinput.sys]--> VIOInputInitInterruptHandling [1]0004.018C::10/22/2025-10:38:03.405 [vioinput.sys] VIOInputEvtDeviceReleaseHardware...