Eric Astor
For reference, I sent https://github.com/antmicro/xls/pull/1 as an example of the sort of cleaned-up logic that's already possible. It's nice to see that it's possible to write a reasonable clean controlled...
FWIW, I like the second option suggested in @grebe's comment; I think it's nice if a `chan` is the thing that can be passed to send/receive, where only the interface...
Mutually exclusive side effects with a token relationship can cause (avoidable?) scheduling problems
This seems more like a case where the token relationship shouldn't be in place? By definition, the optimizer is expected to respect token relationships. Out of curiosity, can you give...
I've hit this again when adding functionality to an internal library (🧦) - adding a function in one place caused another unrelated function to start reporting TypeInferenceErrors.
~Sorry if I'm missing something, but what does issue #1000 have to do with this?~ (Fixed in original)
In DSLX, I'd propose an optional keyword "template parameter" for `chan`. I think we could go with `chan`, etc.
On more consideration, I'd propose the following syntax, using an attribute on `chan` members of a proc: ``` #[strictness(total_order)] ch: chan in; ``` or possibly: `ch: chan in #![strictness(total_order)];`, our...
I believe this is a bug introduced with the new token semantics; the inliner wasn't handling that appropriately. I think I have a fix forthcoming!
@meheff - this is the issue we discussed today!
> For example: > > ``` > (token, data, valid) = peek(token, ch0); > (token, data0) = recv(token, ch0); > ``` > > In hardware, the behavior depends on how...