Closed shingarov closed 9 months ago
This is happening because #reduceEnvironments
is unimplemented. It is somewhat annoying (we have to read longer constraints) but is not a real problem because there is the --no-environment-reduction
flag on the upstream Fixpoint side.
Consider the following Horn query:
During the initial
solve
, SimpC₁'s IBindEnv is {0,1}. These ids point to the following entries in the BE:This in itself does not lead to incorrect answers, but blocks working on PLE, because from such IBindEnv, mkCTrie produces the wrong Trie.
This Issue calls to investigate how this "0" entry creeps into the SInfo.