- 11 Oct, 2018 4 commits
-
-
Andrei Paskevich authored
"at/old operator unused" is now a warning not an error. See merge request !42
-
Guillaume Melquiond authored
-
Sylvain Dailler authored
The warning can be disabled by a debug flag. The bench was changed so as to accept examples that **should** output warning but not fail (this category did not exist, it does now) and label_scope has been switched to it.
-
Guillaume Melquiond authored
-
- 10 Oct, 2018 6 commits
-
-
DAILLER Sylvain authored
Ce bench update See merge request !39
-
Sylvain Dailler authored
-
Sylvain Dailler authored
-
MARCHE Claude authored
Added debugging for driver's transformations See merge request !38
-
Sylvain Dailler authored
-
Mário Pereira authored
This closes isse #179
-
- 09 Oct, 2018 1 commit
-
-
Sylvain Dailler authored
-
- 08 Oct, 2018 8 commits
-
-
MARCHE Claude authored
-
MARCHE Claude authored
-
Sylvain Dailler authored
-
Andrei Paskevich authored
This does not fix #183, just prevents the failure. TBC.
-
MARCHE Claude authored
Resolve "Reloading should interrupt running provers" Closes #174 See merge request !36
-
MARCHE Claude authored
interrupt is now properly handled by why3server
-
-
DAILLER Sylvain authored
Fix ce-bench for diffs between platforms See merge request !34
-
- 05 Oct, 2018 3 commits
-
-
Sylvain Dailler authored
This makes Timeout and Unknown equivalent and remove the diff on floats that differs on moloch and CI.
-
Guillaume Melquiond authored
-
Andrei Paskevich authored
-
- 04 Oct, 2018 7 commits
-
-
Andrei Paskevich authored
subst can make constant definitions recursive, making them fail termination check. Convert such definitions into axioms on the fly. Tentative fix for #195.
-
Andrei Paskevich authored
use variant inference from Decl to prove termination for pure recursive functions without variant.
-
Jean-Christophe Filliâtre authored
-
Guillaume Melquiond authored
For a measured time of 0.49 second, a prover would only get 1 second to replay the proof, i.e., half of the expected 1.98 seconds. The time limit is now rounded to nearest to prevent this underestimation on fast proofs.
-
Guillaume Melquiond authored
-
Andrei Paskevich authored
-
Andrei Paskevich authored
-
- 03 Oct, 2018 4 commits
-
-
Raphaël Rieu-Helft authored
C extraction cosmetics See merge request !32
-
Raphael Rieu-Helft authored
-
DAILLER Sylvain authored
Naming proposition for expl of transformations with arguments #191 See merge request !28
-
Raphael Rieu-Helft authored
-
- 02 Oct, 2018 7 commits
-
-
Raphael Rieu-Helft authored
-
MARCHE Claude authored
the beginner transformation `split_vc` is now using `introduce_premises` followed by `subst_all`, instead of `simplify_trivial_quantification` followed by `introduce_premises`. following Andrei's suggestion, instead of `subst_all` we instead substitute only the symbols that (1) were introduced earlier and (2) do not have any attributes `[@model...]` so as to keep symbols present in the initial code.
-
DAILLER Sylvain authored
Fix issue #190 Closes #190 See merge request !27
-
Sylvain Dailler authored
This puts explanations on goals generated by some transformations with arguments. Ideally, no goals should be without explanations.
-
Raphael Rieu-Helft authored
-
Mário Pereira authored
-
Sylvain Dailler authored
Exceptions from transformations are of two kinds: - fatal exception which are then raised into a popup in the ide - normal exception which appears in the message view
-