Guillaume Bury
63d8dc1dc2
[bugfix/major] Clauses can be added at level > 0
2016-04-15 12:49:38 +02:00
Guillaume Bury
ec7c6602e9
Merge branch 'master' of github.com:Gbury/mSAT
2016-04-15 12:11:38 +02:00
Guillaume Bury
815098dde4
Propagation reasons are now far more explicit
2016-04-15 12:09:23 +02:00
Simon Cruanes
479f4c1b79
small checks
2016-04-15 11:54:00 +02:00
Guillaume Bury
9a5c23d9c5
[bugfix] termination check after full slice was wrong
...
When the solver finds a SAT result, it sends the whole
model to the theory, because maybe it can do something
interesting/costly to expand the proof search. After
that there must be a check to see if the theory has effectively done
something, in which case we should resume proof search, or if nothing
has been done, in which case the solver should return that the problem
is satisfiable. That check was incorrect before (checking number of
assumptions, and if the queue is all caught up), because new learnt
clauses (i.e tautologies, which are *not* assumptions) can be added that
do not immediately causes propagation, so that the number of assumptions
and the element queue is constant, but we should still resume the
search.
2016-03-04 16:30:51 +01:00
Guillaume Bury
0883615b99
Replaced Either.destruct by explicit matching
2016-03-04 15:50:13 +01:00
Guillaume Bury
ea518c6ab3
Update for compatibility with ocaml 4.00.1
2016-02-29 13:43:46 +01:00
Guillaume Bury
b6effe691c
Added dummy arguments to some functors
...
Some functors lacked a dummy argument, and thus
could introduce some problems if duplicated (because
they would share the same internal state).
2016-02-29 10:58:22 +01:00
Guillaume Bury
94ce8dbd25
Fixed push/pop base level and reset function to do what is said in doc
2016-02-05 14:32:41 +01:00
Guillaume Bury
beff9ee7c1
Merge branch 'master' of github.com:Gbury/mSAT
2016-01-30 17:04:23 +01:00
Guillaume Bury
30842da947
Fixed refreshing of clauses with push/pop
2016-01-30 17:03:14 +01:00
Simon Cruanes
60250f2611
add UndecidedLit for eval, eval_level
2016-01-29 17:00:05 +01:00
Guillaume Bury
ea1a360148
Merge branch 'master' of github.com:Gbury/mSAT
2016-01-21 03:52:53 +01:00
Guillaume Bury
9a481f6450
Better proofs
...
Proof graphs are now entirely stored in
the cpremise field of clauses
2016-01-21 03:34:18 +01:00
Guillaume Bury
2613926ab1
First changes for better persistent proofs
...
This commit ensures that clauses now contain
all necessary information to construct the proof
graph (without relying on propagation reasons).
2016-01-21 00:06:41 +01:00
Simon Cruanes
756363ffd6
everwhere, use new Log interface and remove the functor on Log_intf
2016-01-20 21:05:22 +01:00
Simon Cruanes
facfe336a1
add eval_level in the API of the SAT solver
2016-01-20 20:06:56 +01:00
Guillaume Bury
a21807c624
Fix for semantic conflict and decision levels
...
When faced with a semantic conflict, a new decision
must be made to solve the conflict, however, because
we just backtracked, it is not ensured that the solver
is in a correct state to add a new level, so the decision
is now delayed until propagation has been done.
2015-12-11 16:26:51 +01:00
Guillaume Bury
df1f28ccb1
Fix for when the solver becomes unsat during if_sat
2015-12-11 09:08:10 +01:00
Guillaume Bury
1d1ba51329
Fixed some bugs related to push/pop and propagation
2015-12-08 15:58:49 +01:00
Guillaume Bury
3168d4ae2b
Fixed bug in if_sat
...
The trick using a bool ref to see if a new clause has been pushed
did not detect cases where new clauses were added using assume.
2015-12-08 13:11:05 +01:00
Simon Cruanes
19e7179334
wip: remove Log in critical path (improves perf)
2015-11-27 14:59:33 +01:00
Guillaume Bury
3f6de07440
Merge branch 'master' into push_pop
2015-11-27 14:53:41 +01:00
Simon Cruanes
99b1f25e4f
style
2015-11-25 10:28:21 +01:00
Simon Cruanes
15e5a4224d
ocp-indent all the files, for the greater good!
2015-11-25 10:04:01 +01:00
Guillaume Bury
a32d443b96
Cleaned unused vec
2015-11-24 16:09:59 +01:00
Guillaume Bury
3fd91d9751
Fixed push in case of unsat env
...
Renamed some field names of env in solver/internal
2015-11-24 14:33:02 +01:00
Guillaume Bury
5911f18cb4
Res module adapted to accomodate puush/pop
2015-11-19 14:59:54 +01:00
Guillaume Bury
28f32de24c
Removed assertion
...
Long explanation: when backtracking to level 0,
while already being at level 0, very strange
things might happen, most notably there might still
be facts left to propagate...
2015-11-18 17:43:26 +01:00
Guillaume Bury
6567d32900
Fixed bad decision level updating during pop
2015-11-18 17:37:36 +01:00
Guillaume Bury
763d23146f
A *lot* of fixes for push/pop
2015-11-17 16:17:14 +01:00
Guillaume Bury
e2cac78d39
Fixed typos in clause simplification
2015-10-20 16:48:43 +02:00
Guillaume Bury
ac5e8a9766
First test (probably unsound)
2015-10-19 22:04:15 +02:00
Guillaume Bury
aed3aeb17c
A bit of restructuring to have cleaner dependencies between fonctors
2015-07-21 19:20:40 +02:00
Guillaume Bury
9c1ca06aea
Dot output is now available through independent backend
2015-07-09 19:03:44 +02:00
Guillaume Bury
4b51f22464
Changed internal representation of proofs
2015-07-09 16:29:57 +02:00
Guillaume Bury
e7140d6897
Added some abstraction to allow for more direct types int the pure SAT
...
solver
2015-06-26 14:12:47 +02:00
Guillaume Bury
ce05d8fe62
Simpler representation of solver types
2015-06-26 12:58:00 +02:00
Guillaume Bury
6f384fb80b
Big refactoring of code. Some performances were lost on pure SAT Solving.
2015-06-25 15:37:29 +02:00
