all: exclude file that have ".license" file

If the file X has another file named "X.license" in the same directory,
exclude it for being processed.

The ".license" is SPDX specific file that contains only SPDX identifiers.

4 files changed, 29 insertions, 4 deletions

diff --git a/spdxconv.go b/spdxconv.go
index 7222eb4..b303070 100644
--- a/spdxconv.go
+++ b/spdxconv.go
@@ -7,6 +7,7 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"slices"
 	"strings"
 
 	"git.sr.ht/~shulhan/pakakeh.go/lib/git"
@@ -70,6 +71,7 @@ func New(path string) (conv *SPDXConv, err error) {
 	}
 
 	conv = &SPDXConv{}
+
 	conv.curDir, err = os.Getwd()
 	if err != nil {
 		return nil, fmt.Errorf(`%s: %w`, logp, err)
@@ -179,6 +181,8 @@ func (conv *SPDXConv) scanFiles(listDir []string) (listFile []string, err error)
 		`node_modules`: struct{}{},
 		`vendor`:       struct{}{},
 	}
+	var suffixLicense = `.license`
+
 	var dir string
 	var listde []os.DirEntry
 	var de os.DirEntry
@@ -191,8 +195,17 @@ func (conv *SPDXConv) scanFiles(listDir []string) (listFile []string, err error)
 		if err != nil {
 			return listFile, err
 		}
+
+		var listFileLicense = make(map[string]struct{})
+		var listName []string
 		for _, de = range listde {
 			var name = de.Name()
+			// File end with ".license" is SPDX specific file that
+			// contains only identifiers.
+			if strings.HasSuffix(name, suffixLicense) {
+				listFileLicense[name] = struct{}{}
+				continue
+			}
 			if conv.scm.IsIgnored(name) {
 				continue
 			}
@@ -200,12 +213,20 @@ func (conv *SPDXConv) scanFiles(listDir []string) (listFile []string, err error)
 			if ok {
 				continue
 			}
-			var pathName = filepath.Join(dir, name)
+			listName = append(listName, name)
+			name = filepath.Join(dir, name)
 			if de.IsDir() {
-				listDir = append(listDir, pathName)
+				listDir = append(listDir, name)
 				continue
 			}
-			listFile = append(listFile, pathName)
+			listFile = append(listFile, name)
+		}
+		// Remove the file that have the spdx .license file.
+		for x, name := range listName {
+			_, ok = listFileLicense[name+suffixLicense]
+			if ok {
+				listFile = slices.Delete(listFile, x, x+1)
+			}
 		}
 	}
 	return listFile, nil
diff --git a/testdata/loadConfig/config_exists/spdxconv.cfg b/testdata/loadConfig/config_exists/spdxconv.cfg
index e4edf89..669d38a 100644
--- a/testdata/loadConfig/config_exists/spdxconv.cfg
+++ b/testdata/loadConfig/config_exists/spdxconv.cfg
@@ -11,4 +11,4 @@ delete_line_pattern = "^(//+|#+)*\\s*$"
 delete_line_pattern = "^(//+|#+)*\\s+license that(.*)$"
 
 [match-copyright]
-pattern = "^(//+|#+)*\\s+Copyright\\s+(?<year>\\d{4},?\\s+(?<holder>.*)\\s+<*(?<email>.*)>.*$"
+pattern = "^(//+|#+)*\\s+Copyright\\s+(?<year>\\d{4}),?\\s+(?<holder>.*)\\s+<*(?<email>.*)>.*$"
diff --git a/testdata/scan/dummy b/testdata/scan/dummy
new file mode 100644
index 0000000..73fd7c8
--- /dev/null
+++ b/testdata/scan/dummy
@@ -0,0 +1,2 @@
+// SPDX-License-Identifier: GPL-3.0-only
+// SPDX-FileCopyrightText: 2025 M. Shulhan <ms@kilabit.info>
diff --git a/testdata/scan/dummy.license b/testdata/scan/dummy.license
new file mode 100644
index 0000000..73fd7c8
--- /dev/null
+++ b/testdata/scan/dummy.license
@@ -0,0 +1,2 @@
+// SPDX-License-Identifier: GPL-3.0-only
+// SPDX-FileCopyrightText: 2025 M. Shulhan <ms@kilabit.info>