liner, word separators: add configurable word separators

common.go

@@ -17,22 +17,23 @@ import (
 )
 
 type commonState struct {
-	terminalSupported bool
-	outputRedirected  bool
-	inputRedirected   bool
-	history           []string
-	historyMutex      sync.RWMutex
-	completer         WordCompleter
-	columns           int
-	killRing          *ring.Ring
-	ctrlCAborts       bool
-	r                 *bufio.Reader
-	tabStyle          TabStyle
-	multiLineMode     bool
-	cursorRows        int
-	maxRows           int
-	shouldRestart     ShouldRestart
-	needRefresh       bool
+	terminalSupported    bool
+	outputRedirected     bool
+	inputRedirected      bool
+	history              []string
+	historyMutex         sync.RWMutex
+	completer            WordCompleter
+	columns              int
+	killRing             *ring.Ring
+	ctrlCAborts          bool
+	r                    *bufio.Reader
+	tabStyle             TabStyle
+	multiLineMode        bool
+	cursorRows           int
+	maxRows              int
+	shouldRestart        ShouldRestart
+	needRefresh          bool
+	wordSeparatorChecker WordSeparatorChecker
 }
 
 // TabStyle is used to select how tab completions are displayed.
@@ -253,3 +254,12 @@ func (s *State) promptUnsupported(p string) (string, error) {
 	}
 	return string(linebuf), nil
 }
+
+// WordSeparatorChecker returns true if a rune should be consider as a word separator
+type WordSeparatorChecker func(r rune) bool
+
+// SetWordSeparatorChecker sets a function that Liner will call to determine
+// the end/beginning of a word
+func (s *State) SetWordSeparatorChecker(ch WordSeparatorChecker) {
+	s.wordSeparatorChecker = ch
+}

line.go

@@ -819,7 +819,7 @@ mainLoop:
 				// Remove word separators to the left
 				var buf []rune // Store the deleted chars in a buffer
 				for {
-					if pos == 0 || !isWordSeparator(line[pos-1]) {
+					if pos == 0 || !s.isWordSeparator(line[pos-1]) {
 						break
 					}
 					buf = append(buf, line[pos-1])
@@ -828,7 +828,7 @@ mainLoop:
 				}
 				// Remove non-word separators to the left
 				for {
-					if pos == 0 || isWordSeparator(line[pos-1]) {
+					if pos == 0 || s.isWordSeparator(line[pos-1]) {
 						break
 					}
 					buf = append(buf, line[pos-1])
@@ -906,10 +906,10 @@ mainLoop:
 						if leftKnown {
 							atWordSeparator = wordSeparatorLeft
 						} else {
-							atWordSeparator = isWordSeparator(line[pos])
+							atWordSeparator = s.isWordSeparator(line[pos])
 						}
 
-						wordSeparatorLeft = isWordSeparator(line[pos-1])
+						wordSeparatorLeft = s.isWordSeparator(line[pos-1])
 						leftKnown = true
 
 						if !atWordSeparator && wordSeparatorLeft {
@@ -936,10 +936,10 @@ mainLoop:
 						if hereKnown {
 							wordSeparatorLeft = atWordSeparator
 						} else {
-							wordSeparatorLeft = isWordSeparator(line[pos-1])
+							wordSeparatorLeft = s.isWordSeparator(line[pos-1])
 						}
 
-						atWordSeparator = isWordSeparator(line[pos])
+						atWordSeparator = s.isWordSeparator(line[pos])
 						hereKnown = true
 
 						if atWordSeparator && !wordSeparatorLeft {
@@ -996,15 +996,15 @@ mainLoop:
 				// Remove word separators to the right
 				var buf []rune // Store the deleted chars in a buffer
 				for {
-					if pos == len(line) || !isWordSeparator(line[pos]) {
+					if pos == len(line) || !s.isWordSeparator(line[pos]) {
 						break
 					}
 					buf = append(buf, line[pos])
 					line = append(line[:pos], line[pos+1:]...)
 				}
 				// Remove non-word separators to the right
 				for {
-					if pos == len(line) || isWordSeparator(line[pos]) {
+					if pos == len(line) || s.isWordSeparator(line[pos]) {
 						break
 					}
 					buf = append(buf, line[pos])
@@ -1175,3 +1175,11 @@ func (s *State) tooNarrow(prompt string) (string, error) {
 	}
 	return s.promptUnsupported(prompt)
 }
+
+func (s *State) isWordSeparator(r rune) bool {
+	if s.wordSeparatorChecker == nil {
+		s.wordSeparatorChecker = SpaceWordSeparatorChecker
+	}
+
+	return s.wordSeparatorChecker(r)
+}

word_separator_checker.go

@@ -0,0 +1,30 @@
+package liner
+
+import "unicode"
+
+// SpaceWordSeparatorChecker (default) returns true if r is a unicode-space
+func SpaceWordSeparatorChecker(r rune) bool {
+	return unicode.IsSpace(r)
+}
+
+// PunctWordSeparatorChecker returns true if r is a unicode punctuation character
+func PunctWordSeparatorChecker(r rune) bool {
+	return unicode.IsPunct(r)
+}
+
+// CombineWordSeparatorChecker combines checkers
+// eg. line.SetWordSeparatorChecker(liner.CombineWordSeparatorChecker(
+//       liner.SpaceWordSeparatorChecker,
+//       liner.PunctWordSeparatorChecker,
+//     ))
+func CombineWordSeparatorChecker(checkers ...WordSeparatorChecker) WordSeparatorChecker {
+	return func(r rune) bool {
+		for _, isSeparator := range checkers {
+			if isSeparator(r) {
+				return true
+			}
+		}
+
+		return false
+	}
+}

word_separator_checker_test.go

@@ -0,0 +1,42 @@
+package liner
+
+import (
+	"testing"
+)
+
+func TestSeparators(t *testing.T) {
+	// test spaces
+	spaces := []rune{' ', '	'}
+	for _, r := range spaces {
+		if !SpaceWordSeparatorChecker(r) {
+			t.Errorf("'%s' was not recognized by the space separator", string(r))
+		}
+	}
+
+	// test punctuation character
+	puncts := []rune{'(', ')', '{', '}'} // etc
+	for _, r := range puncts {
+		if !PunctWordSeparatorChecker(r) {
+			t.Errorf(`'%s' was not recognized by the "punctuation" separator`, string(r))
+		}
+	}
+
+	// test combination of them
+	combinedChecker := CombineWordSeparatorChecker(
+		SpaceWordSeparatorChecker,
+		PunctWordSeparatorChecker,
+	)
+	for _, r := range append(puncts, spaces...) {
+		if !combinedChecker(r) {
+			t.Errorf("'%s' was not recognized by the combined separator", string(r))
+		}
+	}
+
+	// test some letters
+	justLetters := []rune{'a', 'b', 'c'}
+	for _, r := range justLetters {
+		if combinedChecker(r) {
+			t.Errorf("'%s' was recognized as a space or a punctuation char", string(r))
+		}
+	}
+}