Philipp Wagner

@cmarqu good catch. I think it's slightly wrong right now anyways, but let's at least not make the problem worse. Updated.

Can you add a bit of motivation for this change, and how it potentially impacts users to the news fragment?

@marlonjames is the following understanding right? * Users have Python somewhere where `dlopen()` doesn't find it. * When building from source, we're baking that non-standard library path into the cocotb...

I just refreshed my mind on what we actually do behind the scenes, and we actually do more magic than I remembered to find the right libpython, and set LIBPYTHON_LOC...

And to confirm, @jwprice100, you're using the cocotb Makefiles? Or something else?

Naming decisions: * [hdl_]toplevel * [hdl_]toplevel_lang * gpi_interface

Anthony, thanks for filing this issue. I don't think we have a much easier way to get the test name right now. Can you elaborate a bit what you're using...

Hi @DiegoRosales. AFAIK, there was no progress on this issue so far. We do not have VCS available for testing, and no cocotb user with VCS has helped out so...

@aubindetrez I know @ktbarrett has been thinking about BinaryValue a lot, but I think nobody ever found the time to push it through completely. Would you be interested in spending...

A first step towards this is in #2980. There's more to be done, including pinning the OS versions.