Nick Gasson

> It would be better to just make the counter saturate at 2^32 - 1 instead. I did this in 1745c08.

Generic types/packages with components and configurations don't work very well at the moment. Here's another example: ``` entity sub is generic ( type t; function "+" (a, b : t)...

This no longer crashes, but produces this error when run: ``` ** Fatal: 0ms+0: actual length 54 does not match port B length 27 > /home/nick/nvc/test/regress/issue883.vhd:141 | 141 | b...

I've improved the error message so it points at the `complex.pack` call. I think the problem is that function returns a value 2x the width of the port, perhaps it...

No good ideas unfortunately. But the other programs you mention above are all using X11 directly whereas xcowsay is actually a Gtk app. Do you get the same error with...

Thanks for reporting the bug. However I cannot reproduce on the current master branch. Which version are you using? It's possible it was fixed by 666ad997d (which isn't in any...

Yes the problem is with LLVM 18. I managed to reproduce it myself now.

This was a compatibility issue with LLVM 18 and is fixed in the latest 1.12.2 release.

Doesn't seem to work with any of the commercial simulators but I don't have a license at the moment to debug.

For older standards we have the "Billowitch" tests from VESTs. The old CVS repository is down but I have a mirror [here](https://github.com/nickg/vests/tree/master/vhdl-93/billowitch) and I believe there's a copy in the...