Rockingcool/kleingrep
2
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Implement PCRE Matching (prefer left-branch) #2

Merged
Rockingcool merged 48 commits from implementPCREMatchingRules into master 2025-02-09 15:24:29 -06:00
Conversation 0 Commits 48 Files Changed 11 +388 -190
6 changed files with 388 additions and 190 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit 858e535fba - Show all commits

41
regex/matching.go
View File

@@ -153,7 +153,7 @@ func pruneIndices(indices []Match) []Match {
func copyThread(to *nfaState, from nfaState) { func copyThread(to *nfaState, from nfaState) {
to.threadSP = from.threadSP to.threadSP = from.threadSP
to.threadGroups = from.threadGroups to.threadGroups = append([]Group{}, from.threadGroups...)
} }
// Find returns the 0-group of the leftmost match of the regex in the given string. // Find returns the 0-group of the leftmost match of the regex in the given string.
@@ -325,10 +325,33 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
if currentState.groupBegin { if currentState.groupBegin {
currentState.threadGroups[currentState.groupNum].StartIdx = idx currentState.threadGroups[currentState.groupNum].StartIdx = idx
// allMatches := make([]nfaState, 0)
// for _, v := range currentState.transitions {
// dereferenced := funcMap(v, func(s *nfaState) nfaState {
// return *s
// })
// allMatches = append(allMatches, dereferenced...)
// }
// slices.Reverse(allMatches)
// for i := range allMatches {
// copyThread(&allMatches[i], currentState)
// }
// currentStates = append(currentStates, allMatches...)
} }
if currentState.groupEnd { if currentState.groupEnd {
currentState.threadGroups[currentState.groupNum].EndIdx = idx currentState.threadGroups[currentState.groupNum].EndIdx = idx
// allMatches := make([]nfaState, 0)
// for _, v := range currentState.transitions {
// dereferenced := funcMap(v, func(s *nfaState) nfaState {
// return *s
// })
// allMatches = append(allMatches, dereferenced...)
// }
// slices.Reverse(allMatches)
// for i := range allMatches {
// copyThread(&allMatches[i], currentState)
// }
// currentStates = append(currentStates, allMatches...)
} }
// if currentState.isKleene { // if currentState.isKleene {
@@ -363,7 +386,7 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
} }
// Empty state - enqueue next state, do _not_ increment the SP // Empty state - enqueue next state, do _not_ increment the SP
if currentState.isEmpty && currentState.assert == noneAssert { if currentState.isEmpty && currentState.assert == noneAssert { //&& currentState.groupBegin == false && currentState.groupEnd == false {
isEmptyAndNoAssertion = true isEmptyAndNoAssertion = true
} }
@@ -386,10 +409,16 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
allMatches[i].threadSP += 1 allMatches[i].threadSP += 1
} }
} }
if currentState.groupBegin {
currentStates = slices.Insert(currentStates, currentStateIdx+1, allMatches...)
} else if currentState.groupEnd {
currentStates = append(currentStates, allMatches...)
} else {
nextStates = append(nextStates, allMatches...) nextStates = append(nextStates, allMatches...)
} }
}
if currentState.isLast { // Last state reached if currentState.isLast && len(nextStates) == 0 { // Last state reached
if foundMatch { if foundMatch {
if currentState.assert != noneAssert { if currentState.assert != noneAssert {
currentState.threadGroups[0].EndIdx = idx currentState.threadGroups[0].EndIdx = idx
@@ -400,7 +429,7 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
idx += 1 idx += 1
} }
return true, currentState.threadGroups, idx return true, currentState.threadGroups, idx
} else if currentState.isEmpty && currentState.assert == noneAssert { } else if isEmptyAndNoAssertion {
currentState.threadGroups[0].EndIdx = idx currentState.threadGroups[0].EndIdx = idx
if idx == currentState.threadGroups[0].StartIdx { if idx == currentState.threadGroups[0].StartIdx {
idx++ idx++
@@ -410,7 +439,7 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
} }
} }
copy(currentStates, nextStates) currentStates = append([]nfaState{}, nextStates...)
nextStates = nil nextStates = nil
} }
return false, []Group{}, i + 1 return false, []Group{}, i + 1