Closed IvoWingelaar closed 4 years ago
The code's behavior is the correct one. (We don't usually have any guarantee that the hypotheses will be used, or used in the right order, so it is more convenient to have them in the heap instead so we can reference them 0 or more times.) I will update the paper.
The comment to mm0-c/verifier.c
for CMD_PROOF_HYP
was also in error; fixed in 0c556a2.
In mm0-c, the comments near the execution of
CMD_PROOF_HYP
claims it pushes|- e
to the main stack (as specified in the paper), but the code pushes it to the main heap instead.