1. 11 Dec, 2015 1 commit
  2. 12 Jun, 2015 2 commits
  3. 23 May, 2015 1 commit
  4. 23 Mar, 2015 1 commit
  5. 20 Mar, 2015 1 commit
  6. 19 Mar, 2015 1 commit
  7. 18 Feb, 2015 1 commit
  8. 17 Feb, 2015 1 commit
  9. 16 Feb, 2015 1 commit
  10. 02 Sep, 2014 1 commit
  11. 14 Mar, 2014 1 commit
  12. 14 Jan, 2014 1 commit
  13. 28 Oct, 2013 1 commit
  14. 22 Oct, 2013 1 commit
  15. 16 Apr, 2013 1 commit
  16. 09 Apr, 2013 2 commits
  17. 06 Mar, 2013 1 commit
  18. 04 Mar, 2013 1 commit
    • Andrei Paskevich's avatar
      whyml: match expression is ghost if we look inside ghost fields · 43b684d0
      Andrei Paskevich authored
      We store in every lsymbol a new integer field ls_constr,
      equal to zero if the lsymbol is not a constructor, and equal
      to the number of constructors of the lsymbol's type otherwise.
      It is allowed to declare or define an lsymbol with ls_constr > 0
      as an ordinary function (otherwise algebraic type elimination
      wouldn't work - though we might still check this in theories),
      but it is forbidden to use a wrong ls_constr in algebraic type
      definitions.
      
      The ghostness of a match expression is now determined as follows:
      
      If at least one branch expression is ghost,
        then the match is ghost;
      else if there is only one branch,
        then the match is not ghost;
      else if the matched expression is ghost,
        then the match is ghost;
      else if at least one pattern matches a ghost field
              against a constructor with ls_constr > 1
        then the match is ghost;
      else
        the match is not ghost.
      
      We do just enough to recognize obvious non-ghost cases, and
      make no attempt to handle redundant matches or to detect
      exhaustive or-patterns in subpatterns.
      43b684d0
  19. 08 Jan, 2013 1 commit
  20. 07 Jan, 2013 1 commit
  21. 21 Nov, 2012 1 commit
  22. 06 Nov, 2012 1 commit
  23. 05 Nov, 2012 1 commit
  24. 04 Nov, 2012 1 commit
  25. 30 Oct, 2012 1 commit
  26. 21 Oct, 2012 2 commits
  27. 20 Oct, 2012 1 commit
    • Andrei Paskevich's avatar
      simplify copyright headers · 11598d2b
      Andrei Paskevich authored
      + create AUTHORS file
      + fix the linking exception in LICENSE
      + update the "About" in IDE
      + remove the trailing whitespace
      + inflate my scores at Ohloh
      11598d2b
  28. 18 Sep, 2012 1 commit
  29. 11 Sep, 2012 3 commits
  30. 30 Aug, 2012 1 commit
  31. 23 Jul, 2012 2 commits
    • François Bobot's avatar
      prover identification: use shortcuts · 92165a83
      François Bobot authored
         shortcuts are defined in why3.conf. They are automatically
         generated using two mechanism:
         - a shortcut section in prover-detection-data.conf creates a shortcut
         for the first prover that match the regexp
      
         - the identifier used as family argument for the prover section in
         prover-detection-data.conf is used as shortcut for the prover. If
         different sections use the same argument the first one that match an
         existing prover is used for the shortcut.
      92165a83
    • François Bobot's avatar
      new prover identification: remove id · 5a3641ec
      François Bobot authored
             Remove the id in prover that is used only for command-line, use
             instead the name,version,alternative of the prover. One can
             also use regular expression (start with ^).
      
             "Alt-Ergo,0.92,with arrays" corresponds only to one prover
             "Alt-Ergo,^0\.9.*,with arrays" correspond to all the Alt-Ergo prover with arrays which version match "0\.9.*"
             "Alt-Ergo" is the same thing than "Alt-Ergo,^,^"
             "Alt-Ergo,0.92," corresponds only to one prover with the alternate fields empty
             "Alt-Ergo,,with arrays" corresponds to "Alt-Ergo,^,with arrays" since the version is never empty.
      
             Provers identification are case sensitive even if it is
             currently more complicated for the user because
             case-insensitiveness is not sufficient. Specifiying "alt-ergo"
             for "Alt-Ergo,^,^" is great, but not if there is more than one
             match. A more general system of shortcut would be more
             appropriate.
      5a3641ec
  32. 01 Jun, 2012 1 commit
  33. 22 May, 2012 1 commit
  34. 18 Apr, 2012 1 commit