This PR resurrects the tests about cancellation of `try?` tactics. They were flaky before. By introducing an option (`debug.tactic.try.onlyUserSuggestions`) to make these tests not run `exact?` as part of `try?`, and thus be generally faster and use less memory, the flakiness seems to have disappeared, at least in non-`master` testing. We’ll see. --------- Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
3 lines
59 B
Text
3 lines
59 B
Text
first: blocked
|
|
attempt_all_par: blocked
|
|
first_par: blocked
|