This PR fixes a bug where definitions with nested proofs that contain `sorry` might not report "warning: declaration uses 'sorry'" if the proof has the same type as another nested proof from a previous declaration. The bug only affected log messages; `#print axioms` would still correctly report uses of `sorryAx`. The fix is that now the abstract nested proofs procedure does not consult the aux lemma cache if the proof contains a `sorry`. Closes #10196 |
||
|---|---|---|
| .. | ||
| bench | ||
| compiler | ||
| elabissues | ||
| ir | ||
| lean | ||
| pkg | ||
| playground | ||
| plugin | ||
| simpperf | ||
| .gitignore | ||
| common.sh | ||
| lakefile.toml | ||
| lean-toolchain | ||