Tynan McAuley

Thanks @hzeller! That resolved the `install: illegal option -- T` error. At this point, we still have to deal with the `source` issue in `bazel-bin/install`. This comment shows how I...

> In the PR comment, I don't understand `@mem0` instantiating mem0_ext_0 and mem0_ext. Shouldn't `@Dedup` simply be instantiating mem0 twice? The change to the test in the PR looks more...

Alright, I've made a few updates: - `MemOps` is a `struct` instead of a `pair` - We create the external memory instance when we create the wrapper module definition, to...

Side question: in the `NonLocalAnnotation` circuit in `lower-memory.mlir`, should the `@nla0` and `@nla1` non-local annotations be deleted during this pass (since they refer to the original `firrtl.mem` instances), or are...

> The changes make sense. Dead `HierPathOp` can be left for cleanup by later passes, but if symbol names are updated the `HierPathOp` needs to be updated and cannot be...

> > > The changes make sense. Dead `HierPathOp` can be left for cleanup by later passes, but if symbol names are updated the `HierPathOp` needs to be updated and...

@prithayan just rebased again, can you merge this please? Thanks!

> Sorry, missed to merge it last week. Thanks for fixing the issue :) Looking forward to more future contribution @tymcauley . No problem @prithayan, thanks so much for taking...

Just checking in, I see that https://github.com/chipsalliance/firrtl-spec/issues/189 was resolved, with the conclusion that manifests aren't the path forward. Would the idea be to place all outputs in a single directory...

I realize this might be a bit easier to reproduce by just using the `.fir` output. Here's the FIRRTL from Chisel 5.1.0: ``` FIRRTL version 2.0.0 circuit Foo : module...