jwaldmann

Depends on what behaviour you expect. Going by the definition https://hackage.haskell.org/package/pretty-1.1.3.5/docs/Text-PrettyPrint-HughesPJClass.html "simlar to Show but nicer looking" I'd expect the contract for Show, in particular, sth. like "read . show...

see also #37 for other possible instances of Pretty

Some improvement from ``` parseSolution' :: String -> IntMap Bool parseSolution' s = case words s of "SAT" : ws -> IntMap.fromList $ map ( \ v -> (abs v,...

For reference, profiling the above gives this: ``` Speed +RTS -P -RTS 50000 total time = 2.12 secs (2115 ticks @ 1000 us, 1 processor) total alloc = 1,660,659,976 bytes...

with attoparsec: ``` parseSolution :: B.ByteString -> IntMap Bool parseSolution s = case P.parseOnly ( assignment IntMap.empty -- WRONG Right ps -> IntMap.fromList ps assignment :: P.Parser [(Int,Bool)] assignment =...

I will re-check. These days, I am mostly using https://github.com/jwaldmann/ersatz-minisatapi

should apply this patch to cryptominisat5 (and perhaps others) as well, see https://github.com/ekmett/ersatz/issues/16#issuecomment-754148295

ping

I'm on it (just did a rebase). I wanted to take up #787 again. I was benchmarking NFA-DFA conversion with several `Map IntSet foo` implementations (hint: HashMap wins, Data.Map seems...

Yes. Any thoughts on the name (solverPathOptions) and the module (currently, Ersatz.Solver.Minisat)? The semantics is "a solver that respects SAT-Comp specification"? (cf. http://www.satcompetition.org/2009/format-solvers2009.html) But strictly speaking, "glucose" does not, only...