Will Dietz

Looks like we call success when parsing statements once the indent level drops too low (below level started at), and since we mostly-blindly chunk the input by "module"/EOF (or so,...

Previously: #2787.

I am not certain this is the way to go-- just as there was pain caused by our behavior being different in this way previously, there may now be folks...

Updated to detect + generate alternate name for port when wiring through modules that already have a port of that name. Was surprised to find we don't already verify/reject duplicate...

Any idea when this was better? Looking through action logs it seems they've taken 1-2 hours for few months at least. Anyway, looks like we're getting cache "misses" when restoring...

Looking at the [latest run](https://github.com/llvm/circt/actions/runs/2942932294) all ccache caches were restored (!), causing 8/16 configurations to complete in about or less than 5 minutes. These were the Release configurations, and showed...

cc #3003 (as pointed out elsewhere).

Is this still an issue? Drop names is enabled by default now, but I can't reproduce with the linked `FPU.fir`.

## Observations Debugged this a little, and tried a few alternate approaches just to see what's going on. I need to step away from this for now, but in case...

In this post[0] (can't wait, love the pipe syntax and views re:c++20/rangev3) it mentions a known problem with some LLVM bits, particularly mapped_iterator. This is explained by a linked post[1]...