Arkadiusz Kozdra

I started some work on this, you can see the changes on the branch for now: https://github.com/antmicro/verilator-1/tree/standalone-this

I can think of it as similar to SYMRSVDWORD — a rare coincidence that can be easily fixed by the user, but is relatively time-consuming to track down. This actually...

The reproducer code works correctly with `-fno-inline` as a workaround. The source can also be made to work by adding a `/* verilator public */` attribute to the output variable...

I managed to reduce the code down to ~50 lines. Running the test fails by itself, but adding either `-fno-inline` or `-fno-dfg` fixes it, and so does annotating the problematic...

@wsnyder Would you mind taking a look again? I think this code is on its way to being useful.

I believe I don't have anything to add here currently.

Sorry, @YilouWang I cannot see any other message from you, except for the commits. Maybe the GitHub UI set your comments as pending, if so you need to submit them....

@YilouWang This probably means that multi-dimensional arrays do not work properly. You can stick to one-dimensional arrays for the time being, and leave the multiple dimensions for some next PR...

Use of an actual interface as a value of virtual interface dtype is required in V3LinkDot, because unused interfaces are marked dead and not resolved correctly afterwards. We could try...

Unfortunately std::randomize is not yet supported (one previous feature request being #4706), and the current error message on master (#5354) is: ``` %Warning-CONSTRAINTIGN: issue5438.sv:4:14: std::randomize ignored (unsupported) : ... note:...