- Feb 18, 2021
-
-
Loïc Correnson authored
-
- Feb 16, 2021
-
-
Virgile Prevosto authored
-
- Feb 12, 2021
-
-
Virgile Prevosto authored
-
Virgile Prevosto authored
-
- Feb 08, 2021
-
-
Virgile Prevosto authored
-
Virgile Prevosto authored
-
Virgile Prevosto authored
-
- Remove unused Output_Spec - Add GenerateAnnotations and GenerateDeterministicLemmas true by default to preserve default behavior - Move change_formals and change_result classes higher in the file to apply them as soon as possible in the pre/post generation phase - This last step ensures that add_pre_post_from_buch actually only deals with specification, it is not responsible for remapping formals and return value anymore - Now add_pre_post_from_buch can safely be disabled when GenerateAnnotations is false
-
-
- use more Cil smart constructors - use an optional parameter for global variable initial values - factorisation of common code
-
-
- add an "observables" option at the begining of ya files - multiple "observables" options merges together (union of symbols) - check that every function referred to is inside the observable list - remove the initial curOperation and curOpStatus : the main function is not necessarily observable and the pre/post function will set those variables anyway - ignore unobservables functions inside the relevant visitors
-
- Feb 03, 2021
-
-
David Bühler authored
-
- Jan 21, 2021
-
-
Andre Maroneze authored
-
- Oct 12, 2020
-
-
Virgile Prevosto authored
Actions are sequential, which means that `$x=foo; $y = $x;` must be specified as `ensures aorai_y == aorai_x;`, not `\old(aorai_x)`.
-
Virgile Prevosto authored
it turns out that meta-variables can be modified through several transitions, possibly with distinct end states. It is thus not possible to mention that they are unchanged in behaviors that are grouped by the state at the end of the call to the transition function. This commit creates specific behaviors for indicating in which case an auxiliary variable is unchanged.
-
Virgile Prevosto authored
-
Virgile Prevosto authored
Determistic C translation is now a series of `if ... else if ... else ...` This avoids the need for auxiliary variables and inadvertently updating an auxiliary variable that is used in a subsequent test. Additionally, we have a marginally better ordering of tests wrt state numbering
-
- Use the same field "actions" for parsed and typed transitions - Rename confusing types in Promelaast - Reorganized Promelaoutput to mirror the naming convention
-
-
-
- Oct 06, 2020
-
-
Virgile Prevosto authored
-
Virgile Prevosto authored
-
Virgile Prevosto authored
when an action is required for a transition in non-deterministic mode, it was done before the test for updating the boolean variables representing active state, leading to a possibly incorrect decision when said test depends on the value of the variable modified by the action
-
Virgile Prevosto authored
-
Virgile Prevosto authored
-
- Sep 10, 2020
-
-
Virgile Prevosto authored
this is a generalization of `check` vs. `assert` to other annotations. Not all annotations are relevant though. Currently, there are more annotation nodes in the AST that can have a flag `{ tp_only_check }` than was deemed useful in pub/frama-c#25, but said flag can in fact safely be ignored. Moreover, this commit only add this flag in the AST, but provides no further mean to set it to true except for the original `check` keyword (i.e. on `AAssert`). The parser and the behavior of the plugins that can handle the flag will be updated in subsequent commits
-
- Mar 31, 2020
-
-
Loïc Correnson authored
-
- Mar 06, 2020
-
-
- Oct 31, 2019
-
-
- Feb 05, 2019
-
-
Loïc Correnson authored
-
- Jan 21, 2019
-
-
Loïc Correnson authored
(blind make headers from specifications)
-
- Jan 14, 2019
-
-
Loïc Correnson authored
-
- Dec 12, 2018
-
-
Andre Maroneze authored
-
- Dec 04, 2018
-
-
Andre Maroneze authored
-
- Dec 03, 2018
-
-
Valentin Perrelle authored
- Fix Value/Value#5 - Fix Value/Value#14
-
- Nov 28, 2018
-
-
David Bühler authored
-
Andre Maroneze authored
Some case studies (e.g. dyad) use some ugly casts from fd_set_t which lead to the analysis stopping too early. Changing the representation of fd_set_t should also help it better conform to the standard (since a fd_set_t should be able to hold FD_SETSIZE elements).
-
Andre Maroneze authored
-