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 +557 -364
6 changed files with 557 additions and 364 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit fbc9bea9fb - Show all commits

145
regex/matching.go
View File

@@ -74,58 +74,58 @@ func getZeroGroup(m Match) Group {
// given slice. It returns the resulting states. If any of the resulting states is a 0-state, // given slice. It returns the resulting states. If any of the resulting states is a 0-state,
// the second ret val is true. // the second ret val is true.
// If a state begins or ends a capturing group, its 'thread' is updated to contain the correct index. // If a state begins or ends a capturing group, its 'thread' is updated to contain the correct index.
func takeZeroState(states []*nfaState, numGroups int, idx int) (rtv []*nfaState, isZero bool) { //func takeZeroState(states []*nfaState, numGroups int, idx int) (rtv []*nfaState, isZero bool) {
for _, state := range states { // for _, state := range states {
if len(state.transitions[epsilon]) > 0 { // if len(state.transitions[epsilon]) > 0 {
for _, s := range state.transitions[epsilon] { // for _, s := range state.transitions[epsilon] {
if s.threadGroups == nil { // if s.threadGroups == nil {
s.threadGroups = newMatch(numGroups + 1) // s.threadGroups = newMatch(numGroups + 1)
} // }
copy(s.threadGroups, state.threadGroups) // copy(s.threadGroups, state.threadGroups)
if s.groupBegin { // if s.groupBegin {
s.threadGroups[s.groupNum].StartIdx = idx // s.threadGroups[s.groupNum].StartIdx = idx
// openParenGroups = append(openParenGroups, s.groupNum) // // openParenGroups = append(openParenGroups, s.groupNum)
} // }
if s.groupEnd { // if s.groupEnd {
s.threadGroups[s.groupNum].EndIdx = idx // s.threadGroups[s.groupNum].EndIdx = idx
// closeParenGroups = append(closeParenGroups, s.groupNum) // // closeParenGroups = append(closeParenGroups, s.groupNum)
} // }
} // }
rtv = append(rtv, state.transitions[epsilon]...) // rtv = append(rtv, state.transitions[epsilon]...)
} // }
} // }
for _, state := range rtv { // for _, state := range rtv {
if len(state.transitions[epsilon]) > 0 { // if len(state.transitions[epsilon]) > 0 {
return rtv, true // return rtv, true
} // }
} // }
return rtv, false // return rtv, false
} //}
// zeroMatchPossible returns true if a zero-length match is possible // zeroMatchPossible returns true if a zero-length match is possible
// from any of the given states, given the string and our position in it. // from any of the given states, given the string and our position in it.
// It uses the same algorithm to find zero-states as the one inside the loop, // It uses the same algorithm to find zero-states as the one inside the loop,
// so I should probably put it in a function. // so I should probably put it in a function.
func zeroMatchPossible(str []rune, idx int, numGroups int, states ...*nfaState) bool { //func zeroMatchPossible(str []rune, idx int, numGroups int, states ...*nfaState) bool {
zeroStates, isZero := takeZeroState(states, numGroups, idx) // zeroStates, isZero := takeZeroState(states, numGroups, idx)
tempstates := make([]*nfaState, 0, len(zeroStates)+len(states)) // tempstates := make([]*nfaState, 0, len(zeroStates)+len(states))
tempstates = append(tempstates, states...) // tempstates = append(tempstates, states...)
tempstates = append(tempstates, zeroStates...) // tempstates = append(tempstates, zeroStates...)
num_appended := 0 // number of unique states addded to tempstates // num_appended := 0 // number of unique states addded to tempstates
for isZero == true { // for isZero == true {
zeroStates, isZero = takeZeroState(tempstates, numGroups, idx) // zeroStates, isZero = takeZeroState(tempstates, numGroups, idx)
tempstates, num_appended = uniqueAppend(tempstates, zeroStates...) // tempstates, num_appended = uniqueAppend(tempstates, zeroStates...)
if num_appended == 0 { // break if we haven't appended any more unique values // if num_appended == 0 { // break if we haven't appended any more unique values
break // break
} // }
} // }
for _, state := range tempstates { // for _, state := range tempstates {
if state.isEmpty && (state.assert == noneAssert || state.checkAssertion(str, idx)) && state.isLast { // if state.isEmpty && (state.assert == noneAssert || state.checkAssertion(str, idx)) && state.isLast {
return true // return true
} // }
} // }
return false // return false
} //}
// Prunes the slice by removing overlapping indices. // Prunes the slice by removing overlapping indices.
func pruneIndices(indices []Match) []Match { func pruneIndices(indices []Match) []Match {
@@ -376,17 +376,26 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
// Alternation - enqueue left then right state, and continue // Alternation - enqueue left then right state, and continue
if currentState.isAlternation { if currentState.isAlternation {
leftState := currentState.leftState if currentState.isKleene { // Reverse order of adding things
copyThread(leftState, currentState) rightState := currentState.splitState
currentStates = append(currentStates, *currentState.leftState) copyThread(rightState, currentState)
rightState := currentState.rightState currentStates = append(currentStates, *currentState.splitState)
copyThread(rightState, currentState) leftState := currentState.next
currentStates = append(currentStates, *currentState.rightState) copyThread(leftState, currentState)
currentStates = append(currentStates, *currentState.next)
} else {
leftState := currentState.next
copyThread(leftState, currentState)
currentStates = append(currentStates, *currentState.next)
rightState := currentState.splitState
copyThread(rightState, currentState)
currentStates = append(currentStates, *currentState.splitState)
}
continue continue
} }
// 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 { //&& currentState.groupBegin == false && currentState.groupEnd == false { if !currentState.isAlternation && currentState.isEmpty && currentState.assert == noneAssert { //&& currentState.groupBegin == false && currentState.groupEnd == false {
isEmptyAndNoAssertion = true isEmptyAndNoAssertion = true
} }
@@ -396,12 +405,7 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
if isEmptyAndNoAssertion || foundMatch { if isEmptyAndNoAssertion || foundMatch {
allMatches := make([]nfaState, 0) allMatches := make([]nfaState, 0)
for _, v := range currentState.transitions { allMatches = append(allMatches, *(currentState.next))
dereferenced := funcMap(v, func(s *nfaState) nfaState {
return *s
})
allMatches = append(allMatches, dereferenced...)
}
slices.Reverse(allMatches) slices.Reverse(allMatches)
for i := range allMatches { for i := range allMatches {
copyThread(&allMatches[i], currentState) copyThread(&allMatches[i], currentState)
@@ -419,24 +423,11 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
} }
if currentState.isLast && len(nextStates) == 0 { // Last state reached if currentState.isLast && len(nextStates) == 0 { // Last state reached
if foundMatch { currentState.threadGroups[0].EndIdx = idx
if currentState.assert != noneAssert { if idx == currentState.threadGroups[0].StartIdx {
currentState.threadGroups[0].EndIdx = idx idx += 1
} else {
currentState.threadGroups[0].EndIdx = idx + 1
}
if idx == currentState.threadGroups[0].StartIdx {
idx += 1
}
return true, currentState.threadGroups, idx
} else if isEmptyAndNoAssertion {
currentState.threadGroups[0].EndIdx = idx
if idx == currentState.threadGroups[0].StartIdx {
idx++
}
return true, currentState.threadGroups, idx
} }
return true, currentState.threadGroups, idx
} }
} }
currentStates = append([]nfaState{}, nextStates...) currentStates = append([]nfaState{}, nextStates...)