all: get the copyright year from git history

If the line that match with pattern on match-copyright does not contains
year, or there is no match, try to get the year from the first commit of
the file using "git log --follow ..." command.

If no commit history or its not using git, use default copyright year from
configuration.

1 files changed, 49 insertions, 15 deletions

diff --git a/file.go b/file.go
index 742290d..11d2154 100644
--- a/file.go
+++ b/file.go
@@ -6,11 +6,14 @@ package spdxconv
 import (
 	"bytes"
 	"fmt"
+	"log"
 	"os"
 	"path/filepath"
 	"regexp"
 	"slices"
+	"strconv"
 	"strings"
+	"time"
 
 	libos "git.sr.ht/~shulhan/pakakeh.go/lib/os"
 )
@@ -20,17 +23,16 @@ const (
 	valDefault = `default` // Use the default value from configuration.
 	valExist   = `exist`   // The license/copyright exist in the file.
 	valMatch   = `match`   // One of the pattern in match-xxx found in file.
-	valUnknown = `unknown`
 )
 
 // REUSE-IgnoreStart
 
 // reLicenseID regex to detect SPDX license identifier with or without
 // comment prefix.
-var reLicenseID = regexp.MustCompile(`^(//+|#+|/\*+|<!--+)?\s?SPDX-License-Identifier:\s*(.*)\s*$`)
+var reLicenseID = regexp.MustCompile(`^(//+|#+|/\*+|<!--+|--+)?\s?SPDX-License-Identifier:\s*(.*)\s*$`)
 
 // reCopyrightText regex to detect SPDX copyright text.
-var reCopyrightText = regexp.MustCompile(`^(//+|#+|/\*+|<!--+)?\s?SPDX-FileCopyrightText:\s*(.*)\s*$`)
+var reCopyrightText = regexp.MustCompile(`^(//+|#+|/\*+|<!--+|--+)?\s?SPDX-FileCopyrightText:\s*(.*)\s*$`)
 
 // REUSE-IgnoreEnd
 
@@ -41,8 +43,6 @@ type file struct {
 	path string
 
 	// commentPrefix used as prefix to SPDX identifier.
-	// The comment prefix is detected automatically from the first N
-	// lines of file.
 	commentPrefix string
 	commentSuffix string
 
@@ -77,11 +77,8 @@ func newFile(path string, maxLine int) (f *file, err error) {
 	}
 
 	f = &file{
-		path:          path,
-		licenseID:     valDefault,
-		copyrightYear: valUnknown,
-		copyrightText: valDefault,
-		isBinary:      libos.IsBinaryStream(content),
+		path:     path,
+		isBinary: libos.IsBinaryStream(content),
 	}
 	if f.isBinary {
 		return f, nil
@@ -125,6 +122,7 @@ func (f *file) scan(conv *SPDXConv) {
 	}
 	f.scanLicenseID(conv)
 	f.scanCopyrightText(conv)
+	f.getYearFromSCM(conv)
 }
 
 // detectComment get comment prefix and suffix using the "match-file-comment"
@@ -149,6 +147,7 @@ func (f *file) detectComment(cfg *config) {
 }
 
 func (f *file) scanLicenseID(conv *SPDXConv) {
+	f.licenseID = valDefault
 	for _, cml := range conv.cfg.MatchLicense {
 		for x, line := range f.topLines {
 			if reLicenseID.Match(line) {
@@ -175,10 +174,10 @@ func (f *file) scanLicenseID(conv *SPDXConv) {
 			}
 		}
 	}
-	f.licenseID = valDefault
 }
 
 func (f *file) scanCopyrightText(conv *SPDXConv) {
+	f.copyrightText = valDefault
 	for _, cmc := range conv.cfg.MatchCopyright {
 		for x, line := range f.topLines {
 			if reCopyrightText.Match(line) {
@@ -207,7 +206,45 @@ func (f *file) scanCopyrightText(conv *SPDXConv) {
 			}
 		}
 	}
-	f.copyrightText = valDefault
+}
+
+func (f *file) getYearFromSCM(conv *SPDXConv) {
+	if f.copyrightText == valExist {
+		// SPDX for copyright exist, skip it. Even though maybe there
+		// is no year.
+		return
+	}
+	if f.copyrightYear != `` {
+		return
+	}
+	var logp = `getYearFromSCM`
+
+	f.copyrightYear = conv.cfg.CopyrightYear
+
+	logs, err := conv.scm.LogFollow(f.path, ``)
+	if err != nil {
+		log.Printf(`%s %s: %s`, logp, f.path, err)
+		return
+	}
+	if len(logs) == 0 {
+		log.Printf(`%s %s: empty commit history`, logp, f.path)
+		return
+	}
+
+	first_commit := logs[len(logs)-1]
+	fields := strings.Split(first_commit, `,`)
+	if len(fields) < 2 {
+		log.Printf(`%s %s: cannot parse commit log %q`, logp, f.path, first_commit)
+		return
+	}
+
+	sec, err := strconv.ParseInt(fields[1], 10, 64)
+	if err != nil {
+		log.Printf(`%s %s: cannot parse commit timestamp: %s`, logp, f.path, err)
+		return
+	}
+
+	f.copyrightYear = strconv.Itoa(time.Unix(sec, 0).Year())
 }
 
 // apply the SPDX identifier to file.
@@ -242,9 +279,6 @@ func (f *file) apply(conv *SPDXConv) (err error) {
 	}
 	f.topLines = slices.Insert(f.topLines, f.idxLicenseID, rawline)
 
-	if f.copyrightYear == valUnknown {
-		f.copyrightYear = conv.cfg.CopyrightYear
-	}
 	if f.copyrightYear != `` {
 		f.copyrightYear += ` `
 	}