all: split the delete_line_pattern into before and after

While at it, also add configuration for delete line before and after
for match-copyright section.

1 files changed, 124 insertions, 41 deletions

diff --git a/file.go b/file.go
index 798790c..742290d 100644
--- a/file.go
+++ b/file.go
@@ -35,7 +35,8 @@ var reCopyrightText = regexp.MustCompile(`^(//+|#+|/\*+|<!--+)?\s?SPDX-FileCopyr
 // REUSE-IgnoreEnd
 
 type file struct {
-	matchLicense *matchLicense
+	matchCopyright *matchCopyright
+	matchLicense   *matchLicense
 
 	path string
 
@@ -226,7 +227,7 @@ func (f *file) apply(conv *SPDXConv) (err error) {
 		return err
 	}
 
-	f.applyDelete()
+	f.applyDelete(&conv.cfg)
 
 	// REUSE-IgnoreStart
 	line := fmt.Sprintf("%sSPDX-License-Identifier: %s%s",
@@ -360,19 +361,18 @@ func (f *file) applyCopyrightText(conv *SPDXConv) (err error) {
 		}
 		// Verify that the line actually match with one of
 		// match-copyright pattern.
-		var cmc *matchCopyright
-		for _, cmc = range conv.cfg.MatchCopyright {
-			if cmc.match(string(line)) {
-				f.copyrightYear = cmc.year
-				f.copyrightText = cmc.String()
+		for _, f.matchCopyright = range conv.cfg.MatchCopyright {
+			if f.matchCopyright.match(string(line)) {
+				f.copyrightYear = f.matchCopyright.year
+				f.copyrightText = f.matchCopyright.String()
 				break
 			}
 		}
-		if cmc == nil {
-			return fmt.Errorf(`%s: line %d does not match with any of match-license pattern`,
-				f.path, f.idxLicenseID)
+		if f.matchCopyright == nil {
+			return fmt.Errorf(`%s: line %d does not match with any of match-copyright pattern`,
+				f.path, f.idxCopyrightText)
 		}
-		if f.idxLicenseID >= 0 {
+		if f.idxCopyrightText >= 0 {
 			f.topLines[x] = nil
 		} else {
 			f.bottomLines[x] = nil
@@ -381,48 +381,131 @@ func (f *file) applyCopyrightText(conv *SPDXConv) (err error) {
 	return nil
 }
 
-func (f *file) applyDelete() {
-	var reDeleteLine []*regexp.Regexp
+func (f *file) applyDelete(cfg *config) {
+	var startat int
+	var lines [][]byte
 	if f.matchLicense != nil {
-		reDeleteLine = f.matchLicense.reDeleteLine
+		if f.idxLicenseID >= 0 {
+			startat = f.idxLicenseID
+			lines = f.topLines
+		} else {
+			startat = f.idxLicenseID + cfg.MaxLineMatch
+			lines = f.bottomLines
+		}
+		f.deleteLineBefore(f.matchLicense.reDeleteLineBefore, lines, startat-1)
+		f.deleteLineAfter(f.matchLicense.reDeleteLineAfter, lines, startat+1)
 	}
-	var lines [][]byte
-	for _, line := range f.topLines {
-		if line == nil {
-			continue
+	if f.matchCopyright != nil {
+		if f.idxCopyrightText >= 0 {
+			startat = f.idxCopyrightText
+			lines = f.topLines
+		} else {
+			startat = f.idxCopyrightText + cfg.MaxLineMatch
+			lines = f.bottomLines
 		}
-		var found bool
-		for _, redel := range reDeleteLine {
-			if redel.Match(line) {
-				found = true
-				break
-			}
+		f.deleteLineBefore(f.matchCopyright.reDeleteLineBefore, lines, startat-1)
+		f.deleteLineAfter(f.matchCopyright.reDeleteLineAfter, lines, startat+1)
+	}
+	f.trimDeletedLines()
+}
+
+func (f *file) deleteLineBefore(listre []*regexp.Regexp, lines [][]byte, startat int) {
+	if len(listre) == 0 {
+		return
+	}
+	var idxre = 0
+	var re = listre[idxre]
+	var line []byte
+	for x := startat; x >= 0; x-- {
+		if lines[x] == nil {
+			return
 		}
-		if found {
-			continue
+		line = lines[x]
+		if !re.Match(line) {
+			// Once the regex does not match with line, return
+			// immediately.
+			return
+		}
+		lines[x] = nil
+		idxre++
+		if idxre == len(listre) {
+			break
 		}
-		lines = append(lines, line)
+		re = listre[idxre]
 	}
-	f.topLines = lines
+	// Continue deleting with the next lines.
+	for x := len(f.lines) - 1; x >= 0; x-- {
+		line = f.lines[x]
+		if !re.Match(line) {
+			return
+		}
+		f.lines[x] = nil
+		idxre++
+		if idxre == len(listre) {
+			break
+		}
+		re = listre[idxre]
+	}
+}
 
-	lines = nil
-	for _, line := range f.bottomLines {
-		if line == nil {
-			continue
+func (f *file) deleteLineAfter(listre []*regexp.Regexp, lines [][]byte, startat int) {
+	if len(listre) == 0 {
+		return
+	}
+	var idxre = 0
+	var re = listre[idxre]
+	var line []byte
+	for x := startat; x < len(lines); x++ {
+		if lines[x] == nil {
+			return
 		}
-		var found bool
-		for _, redel := range reDeleteLine {
-			if redel.Match(line) {
-				found = true
-				break
-			}
+		line = lines[x]
+		if !re.Match(line) {
+			// Once the regex does not match with line, return
+			// immediately.
+			return
+		}
+		lines[x] = nil
+		idxre++
+		if idxre == len(listre) {
+			break
 		}
-		if found {
+		re = listre[idxre]
+	}
+	// Continue deleting with the next lines.
+	for x := 0; x < len(f.lines); x++ {
+		line = f.lines[x]
+		if !re.Match(line) {
+			return
+		}
+		f.lines[x] = nil
+		idxre++
+		if idxre == len(listre) {
+			break
+		}
+		re = listre[idxre]
+	}
+}
+
+func (f *file) trimDeletedLines() {
+	var newlines [][]byte
+	var lines [][]byte
+	if f.idxLicenseID >= 0 {
+		lines = f.topLines
+	} else {
+		lines = f.bottomLines
+	}
+	for x := 0; x < len(lines); x++ {
+		if lines[x] == nil {
 			continue
 		}
-		lines = append(lines, line)
+		newlines = append(newlines, lines[x])
+	}
+	if f.idxLicenseID >= 0 {
+		f.topLines = newlines
+	} else {
+		f.bottomLines = newlines
 	}
-	f.bottomLines = lines
 }
 
 // insertEmptyLine insert empty line after SPDX identifiers or any comments after it.