module RegExpOps where import RegExp import List {-+ "" =~ r <==> nullable r -} nullable r = case r of Zero -> False One -> True S _ -> False r1 :& r2 -> nullable r1 && nullable r2 r1 :! r2 -> nullable r1 || nullable r2 r1 :-! r2 -> nullable r1 && not (nullable r2) Many re -> True Some re -> nullable re --nullR r = if nullable r then e else z {-+ For all i in first r, there exists an is such that i:is =~ r -} first r = case r of Zero -> [] One -> [] S i -> [i] r1 :& r2 -> first r1 `union` (if nullable r1 then first r2 else []) r1 :! r2 -> first r1 `union` first r2 -- r1 :-! r2 -> first r1 \\ first r2 -- this is wrong! r1 :-! r2 -> first r1 -- `union` first r2 Many re -> first re Some re -> first re {-+ i:is =~ r <==> is =~ follow i r -} follow i r = case r of Zero -> Zero One -> Zero S i' -> if i'==i then One else Zero r1 :& r2 -> follow i r1 & r2 ! (if nullable r1 then follow i r2 else Zero) r1 :! r2 -> follow i r1 ! follow i r2 r1 :-! r2 -> follow i r1 -! follow i r2 Many re -> follow i re & Many re Some re -> follow i re & Many re factors r = (nullable r,[(i,r')|i<-sort (first r),let r'=follow i r,r'/=Zero]) factorsR (n,fs) = (if n then e else z) ! alts [S i & r|(i,r)<-fs] {-+ Inspired by: comp.compilers, comp.theory http://compilers.iecc.com/comparch/article/93-10-022 The Equational Approach: (1) Regular Expressions -> DFA's. markh@csd4.csd.uwm.edu (Mark) Computing Services Division, University of Wisconsin - Milwaukee Sat, 2 Oct 1993 17:47:13 GMT -}