Kyle Cripps
Kyle Cripps
https://github.com/p4lang/p4-spec/pull/1363
> So this does work in the P4C, right? @vlstill For backends that don't run the `EliminateSwitch` midend pass, yes, it already works. The `p4c` frontend passes produce valid IR...
I can't merge, so feel free to merge if everyone is happy with the changes.
@jonathan-dilorenzo I realized that I did not add a bullet to this section in https://github.com/p4lang/p4-spec/pull/1358, so I also added a bullet summarizing the changes made by https://github.com/p4lang/p4-spec/pull/1358.
> Approving this, but maybe clearing up the renaming behavior of inlining is the better long-term solution. Personally, I do not see why inlining should use the control-plane name. @oleg-ran-amd...
> > Approving this, but maybe clearing up the renaming behavior of inlining is the better long-term solution. Personally, I do not see why inlining should use the control-plane name....
@oleg-ran-amd From @asl: > for 4900 – I’d rather implement proper approach with renaming rather than doing band-aid later
Per DM with @oleg-ran-amd, I will merge this now and leave #4883 open for Oleg to look into the suggested better long-term solution later.
> Should there be a test-case for this? > > > > (From what I can see, the code is OK.) I do not know how to test command-line options...
Apologies, I forgot about diagnostics that are both errors and warnings. The original changes I proposed in this PR will not work. @ajwalton Please see the new changes.