Wilson Snyder

>checking for g++ option to enable C++11 features... -std=gnu++11 This is done by configure itself under the AC_PROG_CXX macro call Verilator uses. That it is needed suggests something is broken...

Thanks for the example - making some intermediate flag variables (COROUTINES_FLAGS etc) is a good idea. For the final fix we'll need something that doesn't add any of them if...

@thecaralice does the above patch work?

>is there a way to dynamically allocate std= and fcoroutines flags depending on checks in runtime or some option to reconfigure it for prebuilt binaries? All compile flags are overwritable...

>prebuilt verilator binaries from nixpkgs returns unsupported argument error for -fcoroutines-ts The best route is to build verilator yourself then. Also, likely the version they have is old, so you...

#6352 will add MacOS builds to the CI, so that will become the "official" osx/macos recipe. Not sure if this is otherwise still relevant, but assuming it got worked out,...

The error message you posted is indicating what signals/blocks that Verilator thinks is changing. These don't seem to correlate with what you posted exactly, so I can't offer exact advise,...

You'll need to debug further, there's nothing indicated that helps to give you a direct suggestion to solve the issue.

As we don't see this error can you try to make a pull request to fix it? Maybe: ```return wrapIfRandMode(VN_AS(m_modp, Class), varp, assignp); ``` Try making a temporary first ```...

Looked at this a bit. Basically the current tristate resolution logic assumes that all tristates are generated in wire/assign statements. Either this needs to be fixed - I fixed one...