John Jannotti

I would be surprised if there's much to be optimized on EvalParams allocation. I converted to creating a single EvalParams per group a while back. Further, it creates a small,...

I don't think we going to "see" any of the copying overhead (of passing Transaction, ConsensusParams, BlockHdr) in allocation or GC costs. It's just an execution tax on all of...

Worth connecting with this PR: #3655

This is not completed. On Thu, Sep 29, 2022, 10:43 AM Will Winder ***@***.***> wrote: > Closed #3217 as > completed. > > — > Reply to this email directly,...

During an in-person discussion, we concluded that a better form for this opcode would be: ``` int N0 int N1 int N2 . . int Nn match label0 label1 label2...

In your minimal code size, you show one byte for each value. Are you presuming that the values must be in the `bytecblock`? An advantage of the proposed "stack" scheme...

Ok, I think we have enough context now to understand what is being proposed. I'll call the two proposals "stack" (the one that places match values on the stack) and...

Would a `pushints` (and `pushbytess` or `pushconsts`) opcode satisfy everyone? Then the stack encoding can be used, but Reach can do: ``` pushints 0x55636708 0x90208995 ... match buy sell ......

Don't we need 1001 now to support `block Timestamp` opcode? On Fri, Sep 30, 2022, 3:33 PM AlgoAxel ***@***.***> wrote: > Summary What > > This change modifies the loop...

> > Don't we need 1001 now to support `block Timestamp` opcode? > > ( @jannotti ) > > There should be no impact to opcodes, as `txTail.recent` [is a...