Commented out tests that would only pass with Longest()

regex/compile.go

@@ -14,20 +14,25 @@ var notDotChars []rune
 // the startState of the NFA representation of the regex, and the number of capturing
 // groups in the regex. It also contains the expression string.
 type Reg struct {
-	start     *nfaState
-	numGroups int
-	str       string
+	start         *nfaState
+	numGroups     int
+	str           string
+	preferLongest bool
 }
 
 // NumSubexp returns the number of sub-expressions in the given [Reg]. This is equivalent
 // to the number of capturing groups.
-func (r Reg) NumSubexp() int {
-	return r.numGroups
+func (re Reg) NumSubexp() int {
+	return re.numGroups
 }
 
 // String returns the string used to compile the expression.
-func (r Reg) String() string {
-	return r.str
+func (re Reg) String() string {
+	return re.str
+}
+
+func (re *Reg) Longest() {
+	re.preferLongest = true
 }
 
 const concatRune rune = 0xF0001
@@ -1135,7 +1140,7 @@ func thompson(re []postfixNode) (Reg, error) {
 	concatenate(nfa[0], &lastState)
 
 	// The string is empty here, because we add it in Compile()
-	return Reg{nfa[0], numGroups, ""}, nil
+	return Reg{nfa[0], numGroups, "", false}, nil
 
 }
 

regex/doc.go

@@ -173,6 +173,6 @@ The following features are not available in [regexp], but are supported in my en
  1. Lookarounds
  2. Numeric ranges
 
-The goal is to shorten the first list, and expand the second.
+I hope to shorten the first list, and expand the second.
 */
 package regex

regex/example_test.go

@@ -52,3 +52,40 @@ func ExampleReg_FindSubmatch() {
 	// 0	1
 	// 2	3
 }
+
+func ExampleReg_Expand() {
+	inputStr := `option1: value1
+	option2: value2`
+	regexStr := `(\w+): (\w+)`
+	templateStr := "$1 = $2\n"
+	regexComp := regex.MustCompile(regexStr, regex.RE_MULTILINE)
+	result := ""
+	for _, submatches := range regexComp.FindAllSubmatch(inputStr) {
+		result = regexComp.Expand(result, templateStr, inputStr, submatches)
+	}
+	fmt.Println(result)
+	// Output: option1 = value1
+	// option2 = value2
+
+}
+
+func ExampleReg_LiteralPrefix() {
+	regexStr := `a(b|c)d*`
+	regexComp := regex.MustCompile(regexStr)
+	prefix, complete := regexComp.LiteralPrefix()
+	fmt.Println(prefix)
+	fmt.Println(complete)
+	// Output: a
+	// false
+}
+
+func ExampleReg_Longest() {
+	regexStr := `x|xx`
+	inputStr := "xx"
+	regexComp := regex.MustCompile(regexStr)
+	fmt.Println(regexComp.FindString(inputStr))
+	regexComp.Longest()
+	fmt.Println(regexComp.FindString(inputStr))
+	// Output: x
+	// xx
+}

regex/matching.go

@@ -2,6 +2,8 @@ package regex
 
 import (
 	"fmt"
+	"strconv"
+	"unicode"
 )
 
 // A Match represents a match found by the regex in a given string.
@@ -63,8 +65,8 @@ func copyThread(to *nfaState, from nfaState) {
 
 // Find returns the 0-group of the leftmost match of the regex in the given string.
 // An error value != nil indicates that no match was found.
-func (regex Reg) Find(str string) (Group, error) {
-	match, err := regex.FindNthMatch(str, 1)
+func (re Reg) Find(str string) (Group, error) {
+	match, err := re.FindNthMatch(str, 1)
 	if err != nil {
 		return Group{}, fmt.Errorf("no matches found")
 	}
@@ -72,15 +74,27 @@ func (regex Reg) Find(str string) (Group, error) {
 }
 
 // Match returns a boolean value, indicating whether the regex found a match in the given string.
-func (regex Reg) Match(str string) bool {
-	_, err := regex.Find(str)
+func (re Reg) Match(str string) bool {
+	_, err := re.Find(str)
 	return err == nil
 }
 
+// CompileMatch compiles expr and returns true if str contains a match of the expression.
+// It is equivalent to [regexp.Match].
+// An optional list of flags may be provided (see [ReFlag]).
+// It returns an error (!= nil) if there was an error compiling the expression.
+func CompileMatch(expr string, str string, flags ...ReFlag) (bool, error) {
+	re, err := Compile(expr, flags...)
+	if err != nil {
+		return false, err
+	}
+	return re.Match(str), nil
+}
+
 // FindAll returns a slice containing all the 0-groups of the regex in the given string.
 // A 0-group represents the match without any submatches.
-func (regex Reg) FindAll(str string) []Group {
-	indices := regex.FindAllSubmatch(str)
+func (re Reg) FindAll(str string) []Group {
+	indices := re.FindAllSubmatch(str)
 	zeroGroups := funcMap(indices, getZeroGroup)
 	return zeroGroups
 }
@@ -89,8 +103,8 @@ func (regex Reg) FindAll(str string) []Group {
 // The return value will be an empty string in two situations:
 //  1. No match was found
 //  2. The match was an empty string
-func (regex Reg) FindString(str string) string {
-	match, err := regex.FindNthMatch(str, 1)
+func (re Reg) FindString(str string) string {
+	match, err := re.FindNthMatch(str, 1)
 	if err != nil {
 		return ""
 	}
@@ -103,8 +117,8 @@ func (regex Reg) FindString(str string) string {
 // number of groups. The validity of a group (whether or not it matched anything) can be determined with
 // [Group.IsValid], or by checking that both indices of the group are >= 0.
 // The second-return value is nil if no match was found.
-func (regex Reg) FindSubmatch(str string) (Match, error) {
-	match, err := regex.FindNthMatch(str, 1)
+func (re Reg) FindSubmatch(str string) (Match, error) {
+	match, err := re.FindNthMatch(str, 1)
 	if err != nil {
 		return Match{}, fmt.Errorf("no match found")
 	} else {
@@ -121,9 +135,9 @@ func (regex Reg) FindSubmatch(str string) (Match, error) {
 //  2. Group n found a zero-length match
 //
 // A return value of nil indicates no match.
-func (regex Reg) FindStringSubmatch(str string) []string {
-	matchStr := make([]string, regex.numGroups+1)
-	match, err := regex.FindSubmatch(str)
+func (re Reg) FindStringSubmatch(str string) []string {
+	matchStr := make([]string, re.numGroups+1)
+	match, err := re.FindSubmatch(str)
 	if err != nil {
 		return nil
 	}
@@ -145,8 +159,8 @@ func (regex Reg) FindStringSubmatch(str string) []string {
 // FindAllString is the 'all' version of [FindString].
 // It returns a slice of strings containing the text of all matches of
 // the regex in the given string.
-func (regex Reg) FindAllString(str string) []string {
-	zerogroups := regex.FindAll(str)
+func (re Reg) FindAllString(str string) []string {
+	zerogroups := re.FindAll(str)
 	matchStrs := funcMap(zerogroups, func(g Group) string {
 		return str[g.StartIdx:g.EndIdx]
 	})
@@ -155,14 +169,14 @@ func (regex Reg) FindAllString(str string) []string {
 
 // FindNthMatch return the 'n'th match of the regex in the given string.
 // It returns an error (!= nil) if there are fewer than 'n' matches in the string.
-func (regex Reg) FindNthMatch(str string, n int) (Match, error) {
+func (re Reg) FindNthMatch(str string, n int) (Match, error) {
 	idx := 0
 	matchNum := 0
 	str_runes := []rune(str)
 	var matchFound bool
 	var matchIdx Match
 	for idx <= len(str_runes) {
-		matchFound, matchIdx, idx = findAllSubmatchHelper(regex.start, str_runes, idx, regex.numGroups)
+		matchFound, matchIdx, idx = findAllSubmatchHelper(re.start, str_runes, idx, re.numGroups, re.preferLongest)
 		if matchFound {
 			matchNum++
 		}
@@ -175,14 +189,14 @@ func (regex Reg) FindNthMatch(str string, n int) (Match, error) {
 }
 
 // FindAllSubmatch returns a slice of matches in the given string.
-func (regex Reg) FindAllSubmatch(str string) []Match {
+func (re Reg) FindAllSubmatch(str string) []Match {
 	idx := 0
 	str_runes := []rune(str)
 	var matchFound bool
 	var matchIdx Match
 	indices := make([]Match, 0)
 	for idx <= len(str_runes) {
-		matchFound, matchIdx, idx = findAllSubmatchHelper(regex.start, str_runes, idx, regex.numGroups)
+		matchFound, matchIdx, idx = findAllSubmatchHelper(re.start, str_runes, idx, re.numGroups, re.preferLongest)
 		if matchFound {
 			indices = append(indices, matchIdx)
 		}
@@ -191,7 +205,7 @@ func (regex Reg) FindAllSubmatch(str string) []Match {
 	return indices
 }
 
-func addStateToList(str []rune, idx int, list []nfaState, state nfaState, threadGroups []Group, visited []nfaState) []nfaState {
+func addStateToList(str []rune, idx int, list []nfaState, state nfaState, threadGroups []Group, visited []nfaState, preferLongest bool) []nfaState {
 	if stateExists(list, state) || stateExists(visited, state) {
 		return list
 	}
@@ -199,32 +213,32 @@ func addStateToList(str []rune, idx int, list []nfaState, state nfaState, thread
 
 	if state.isKleene || state.isQuestion {
 		copyThread(state.splitState, state)
-		list = addStateToList(str, idx, list, *state.splitState, threadGroups, visited)
+		list = addStateToList(str, idx, list, *state.splitState, threadGroups, visited, preferLongest)
 		copyThread(state.next, state)
-		list = addStateToList(str, idx, list, *state.next, threadGroups, visited)
+		list = addStateToList(str, idx, list, *state.next, threadGroups, visited, preferLongest)
 		return list
 	}
 	if state.isAlternation {
 		copyThread(state.next, state)
-		list = addStateToList(str, idx, list, *state.next, threadGroups, visited)
+		list = addStateToList(str, idx, list, *state.next, threadGroups, visited, preferLongest)
 		copyThread(state.splitState, state)
-		list = addStateToList(str, idx, list, *state.splitState, threadGroups, visited)
+		list = addStateToList(str, idx, list, *state.splitState, threadGroups, visited, preferLongest)
 		return list
 	}
 	state.threadGroups = append([]Group{}, threadGroups...)
 	if state.assert != noneAssert {
-		if state.checkAssertion(str, idx) {
+		if state.checkAssertion(str, idx, preferLongest) {
 			copyThread(state.next, state)
-			return addStateToList(str, idx, list, *state.next, state.threadGroups, visited)
+			return addStateToList(str, idx, list, *state.next, state.threadGroups, visited, preferLongest)
 		}
 	}
 	if state.groupBegin {
 		state.threadGroups[state.groupNum].StartIdx = idx
-		return addStateToList(str, idx, list, *state.next, state.threadGroups, visited)
+		return addStateToList(str, idx, list, *state.next, state.threadGroups, visited, preferLongest)
 	}
 	if state.groupEnd {
 		state.threadGroups[state.groupNum].EndIdx = idx
-		return addStateToList(str, idx, list, *state.next, state.threadGroups, visited)
+		return addStateToList(str, idx, list, *state.next, state.threadGroups, visited, preferLongest)
 	}
 	return append(list, state)
 
@@ -233,7 +247,7 @@ func addStateToList(str []rune, idx int, list []nfaState, state nfaState, thread
 // Helper for FindAllMatches. Returns whether it found a match, the
 // first Match it finds, and how far it got into the string ie. where
 // the next search should start from.
-func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups int) (bool, Match, int) {
+func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups int, preferLongest bool) (bool, Match, int) {
 	// Base case - exit if offset exceeds string's length
 	if offset > len(str) {
 		// The second value here shouldn't be used, because we should exit when the third return value is > than len(str)
@@ -248,7 +262,7 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
 	// If the first state is an assertion, makes sure the assertion
 	// is true before we do _anything_ else.
 	if start.assert != noneAssert {
-		if start.checkAssertion(str, offset) == false {
+		if start.checkAssertion(str, offset, preferLongest) == false {
 			i++
 			return false, []Group{}, i
 		}
@@ -256,7 +270,7 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
 
 	start.threadGroups = newMatch(numGroups + 1)
 	start.threadGroups[0].StartIdx = i
-	currentStates = addStateToList(str, i, currentStates, *start, start.threadGroups, nil)
+	currentStates = addStateToList(str, i, currentStates, *start, start.threadGroups, nil, preferLongest)
 	var match Match = nil
 	for idx := i; idx <= len(str); idx++ {
 		if len(currentStates) == 0 {
@@ -273,10 +287,12 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
 			if currentState.isLast {
 				currentState.threadGroups[0].EndIdx = idx
 				match = append([]Group{}, currentState.threadGroups...)
-				break
-			} else if !currentState.isAlternation && !currentState.isKleene && !currentState.isQuestion && !currentState.groupBegin && !currentState.groupEnd { // Normal character or assertion
-				if currentState.contentContains(str, idx) {
-					nextStates = addStateToList(str, idx+1, nextStates, *currentState.next, currentState.threadGroups, nil)
+				if !preferLongest {
+					break
+				}
+			} else if !currentState.isAlternation && !currentState.isKleene && !currentState.isQuestion && !currentState.groupBegin && !currentState.groupEnd && currentState.assert == noneAssert { // Normal character
+				if currentState.contentContains(str, idx, preferLongest) {
+					nextStates = addStateToList(str, idx+1, nextStates, *currentState.next, currentState.threadGroups, nil, preferLongest)
 				}
 			}
 		}
@@ -291,3 +307,68 @@ func findAllSubmatchHelper(start *nfaState, str []rune, offset int, numGroups in
 	}
 	return false, []Group{}, i + 1
 }
+
+// Expand appends template to dst, expanding any variables in template to the relevant capturing group.
+//
+// A variable is of the form '$n', where 'n' is a number. It will be replaced by the contents of the n-th capturing group.
+// To insert a literal $, do not put a number after it. Alternatively, you can use $$.
+// src is the input string, and match must be the result of [Reg.FindSubmatch].
+func (re Reg) Expand(dst string, template string, src string, match Match) string {
+	templateRuneSlc := []rune(template)
+	srcRuneSlc := []rune(src)
+	i := 0
+	for i < len(templateRuneSlc) {
+		c := templateRuneSlc[i]
+		if c == '$' {
+			i += 1
+			// The dollar sign is the last character of the string, or it is proceeded by another dollar sign
+			if i >= len(templateRuneSlc) || templateRuneSlc[i] == '$' {
+				dst += "$"
+				i++
+			} else {
+				numStr := ""
+				for unicode.IsDigit(templateRuneSlc[i]) {
+					numStr += string(templateRuneSlc[i])
+					i++
+				}
+				if numStr == "" {
+					dst += "$"
+				} else {
+					num, _ := strconv.Atoi(numStr)
+					if num < len(match) {
+						dst += string(srcRuneSlc[match[num].StartIdx:match[num].EndIdx])
+					} else {
+						dst += "$" + numStr
+					}
+				}
+			}
+		} else {
+			dst += string(c)
+			i++
+		}
+	}
+	return dst
+}
+
+// LiteralPrefix returns a string that must begin any match of the given regular expression.
+// The second return value is true if the string comprises the entire expression.
+func (re Reg) LiteralPrefix() (prefix string, complete bool) {
+	state := re.start
+	if state.assert != noneAssert {
+		state = state.next
+	}
+	for !(state.isLast) && (!state.isAlternation) && len(state.content) == 1 && state.assert == noneAssert {
+		if state.groupBegin || state.groupEnd {
+			state = state.next
+			continue
+		}
+		prefix += string(rune(state.content[0]))
+		state = state.next
+	}
+	if state.isLast {
+		complete = true
+	} else {
+		complete = false
+	}
+	return prefix, complete
+}

regex/nfa.go

@@ -133,7 +133,7 @@ func resetThreadsHelper(state *nfaState, visitedMap map[*nfaState]bool) {
 
 // Checks if the given state's assertion is true. Returns true if the given
 // state doesn't have an assertion.
-func (s nfaState) checkAssertion(str []rune, idx int) bool {
+func (s nfaState) checkAssertion(str []rune, idx int, preferLongest bool) bool {
 	if s.assert == alwaysTrueAssert {
 		return true
 	}
@@ -183,7 +183,7 @@ func (s nfaState) checkAssertion(str []rune, idx int) bool {
 			strToMatch = string(runesToMatch)
 		}
 
-		regComp := Reg{startState, s.lookaroundNumCaptureGroups, s.lookaroundRegex}
+		regComp := Reg{startState, s.lookaroundNumCaptureGroups, s.lookaroundRegex, preferLongest}
 		matchIndices := regComp.FindAll(strToMatch)
 
 		numMatchesFound := 0
@@ -210,9 +210,9 @@ func (s nfaState) checkAssertion(str []rune, idx int) bool {
 }
 
 // Returns true if the contents of 's' contain the value at the given index of the given string
-func (s nfaState) contentContains(str []rune, idx int) bool {
+func (s nfaState) contentContains(str []rune, idx int, preferLongest bool) bool {
 	if s.assert != noneAssert {
-		return s.checkAssertion(str, idx)
+		return s.checkAssertion(str, idx, preferLongest)
 	}
 	if idx >= len(str) {
 		return false

regex/re_test.go

@@ -25,7 +25,9 @@ var reTests = []struct {
 	{"a*b", nil, "qwqw", []Group{}},
 	{"(abc)*", nil, "abcabcabc", []Group{{0, 9}, {9, 9}}},
 	{"((abc)|(def))*", nil, "abcdef", []Group{{0, 6}, {6, 6}}},
-	{"(abc)*|(def)*", nil, "abcdef", []Group{{0, 3}, {3, 6}, {6, 6}}},
+	// This match will only happen with Longest()
+	// {"(abc)*|(def)*", nil, "abcdef", []Group{{0, 3}, {3, 6}, {6, 6}}},
+	{"(abc)*|(def)*", nil, "abcdef", []Group{{0, 3}, {3, 3}, {4, 4}, {5, 5}, {6, 6}}},
 	{"b*a*a", nil, "bba", []Group{{0, 3}}},
 	{"(ab)+", nil, "abcabddd", []Group{{0, 2}, {3, 5}}},
 	{"a(b(c|d)*)*", nil, "abccbd", []Group{{0, 6}}},
@@ -537,7 +539,9 @@ var groupTests = []struct {
 	{"(a+)|(a)", nil, "aaaa", []Match{[]Group{{0, 4}, {0, 4}, {-1, -1}}}},
 	{"(a+)(aa)", nil, "aaaa", []Match{[]Group{{0, 4}, {0, 2}, {2, 4}}}},
 	{"(aaaa)|(aaaa)", nil, "aaaa", []Match{[]Group{{0, 4}, {0, 4}, {-1, -1}}}},
-	{"(aaa)|(aaaa)", nil, "aaaa", []Match{[]Group{{0, 4}, {-1, -1}, {0, 4}}}},
+	// This match will only happen with Longest()
+	//	{"(aaa)|(aaaa)", nil, "aaaa", []Match{[]Group{{0, 4}, {-1, -1}, {0, 4}}}},
+	{"(aaa)|(aaaa)", nil, "aaaa", []Match{[]Group{{0, 3}, {0, 3}, {-1, -1}}}},
 	{"(aaaa)|(aaa)", nil, "aaaa", []Match{[]Group{{0, 4}, {0, 4}, {-1, -1}}}},
 	{"(a)|(aa)", nil, "aa", []Match{[]Group{{0, 1}, {0, 1}}, []Group{{1, 2}, {1, 2}}}},
 	{"(a?)a?", nil, "b", []Match{[]Group{{0, 0}, {0, 0}}, []Group{{1, 1}, {1, 1}}}},