Kaleb Barrett

Can you provide an example?

I mean a *whole* example: HDL code, cocotb test, Makefile/Python runner, etc. Everything we need to reproduce the error.

Based on what you posted before, that's expected. ``` # COUT: 1000.00ns DEBUG gpi VHPI: vhpiVerilog returned from vhpi_get(vhpiType, ...) # COUT: 1000.00ns DEBUG gpi VHPI: Unable to fetch object...

Yes, that's how mixed language discovery currently works in cocotb: 1. Use the interface of the parent object to attempt to discover the child object 2. If not fall back...

Yeah it looks like Riviera's VPI isn't finding the object. More debugging will be needed.

We are not setting the compiler to that. It is being done by the configure script automatically ([see here](https://github.com/cocotb/cocotb/actions/runs/10543210110/job/29210898523?pr=3861#step:17:116)). This line reads `none needed` instead of `-std=gnu++11` when running on...

So I checked the autoconf Changelog and 2.70 introduced checking for C++11, and that change was reported as causing issues and reverted 3 months ago. It should be a part...

@gezalore I tried modifying `t_trace_no_top_name.v` to make the `a` variable an output since that test uses `--main-top-name '-'`, but it seems to pass fine.

Is the `module t` not the top, or is `output logic a` not a port?

Isn't that check assuming the output is produced by Verilator's default C model generation you linked [here](https://github.com/verilator/verilator/pull/5313#issuecomment-2261161597)? We aren't using this code. cocotb (and probably other users) have their own...