Molly Pan

I have encountered the same issue on coqide 8.13.2 and MacOS Monterey 12.0.1 with Intel chips. The -debug log: ``` coqide -debug (Spawn ,90557) EXEC: coqidetop.opt -main-channel stdfds --xml_format=Ppcmds -async-proofs...

@mbudiu-vmw Do you have any comments on this? Based on the post, I think the spec should Clarify the definition of name duplication and name shadowing State clearly the situations...

To summarize what we know at the current status, I am copying over the conclusion from [here](https://github.com/p4lang/p4c/issues/3057#issue-1116122059) and [here](https://github.com/p4lang/p4-spec/issues/960#issuecomment-984005608): > it seems tempting to summarize the allowed comparisons as: >...

Hi @jafingerhut Sorry about the confusion. By discrepancy, I actually meant "difference". (I just realized that "discrepancy" implies "a difference **especially between things that should be the same**.")

Maybe @vgurevich will have a better idea about Andy's suspicion. Is header union implemented/allowed in the Tofino architecture?

I also like Andy's new proposal better. It is more clear from the operation itself (`u.setInvalid()`) that it sets all members invalid, while looking at `u.h1.setInvalid()`, it seems reasonble to...

I have tested set operations in BMV2, and there are many discrepancies between the compiler behavior and the spec. - The spec restrains the operands of `&&&` to only `bit`...

Hi @vgurevich Thank you for your reply. I don't think that implicit casts should be applied recursively into the operations based on what the spec said in 8.9.2. For example,...

When merging, the description of bit-slicing should be shortened by referring to the "numeric types".

As @QinshiWang suggests, maybe we want to rename the "numeric types" to "integral types".