Skip to content
Snippets Groups Projects
  1. Feb 25, 2019
  2. Jun 20, 2018
  3. Jun 11, 2018
  4. Jan 15, 2018
  5. Jan 11, 2018
    • Boris Yakobowski's avatar
      [Tests] do not use Frama_C_foo directly · 7ce18d3c
      Boris Yakobowski authored and David Bühler's avatar David Bühler committed
      Instead, include the proper libc headers, with the specifications.
      Otherwise, in the next commits(s), we won't have preconditions to evaluate, and the tests
      will be come unsound
      
      Changes:
      
      - share/libc/stdlib.c: previous version of calloc in stdlib. used Frama_C_memset, which is idiotic since
        - for non-free we now *have* a builtin for calloc
        - stdlib.c is meant for people with non-free builtins
      
      - same thing for tests/libs/string_c.c: the builtin was used
      
      - nonfree/imprecise.c:
        sizeof(v3), where v3 is an abstract struct, is evaluated slightly less precisely now.
        We don't care as long as it does not crash.
      
      - nonfree/strchr.c:
        the test is buggy, the behavior was undefined
      7ce18d3c
  6. Oct 27, 2017
  7. Sep 14, 2017
  8. May 24, 2017
    • Andre Maroneze's avatar
      [Eva] fix str builtins for negative offsets · d2159ad0
      Andre Maroneze authored
      Each call to fold_itv must be preceded by a check of the actual validity of
      the offsets (since fold_itv will take the intersection). In the case of an
      exact offset, the right side is already taken into account, so we only have
      to check the left (negative) bounds.
      d2159ad0
  9. Oct 02, 2015
Loading