all: fix default regex match license and copyright to ignore comment

Instead of assuming that the comment prefix and space always exists

  "^(//)\s+..."

change it to be optional, so it will works on the multi-line comment.
For example, comment and old headers in html,

  <--
  Copyright ...
  -->

there is no comment prefix and space.

5 files changed, 41 insertions, 17 deletions

diff --git a/config.go b/config.go
index e968fd4..ed6ce10 100644
--- a/config.go
+++ b/config.go
@@ -54,12 +54,12 @@ prefix = "%"
 pattern = "^.*\\.(apk|app|bz2|csv|doc|docx|exe|gif|gz|jpeg|jpg|json|pdf|png|ppt|pptx|svg|svgz|tar|tgz|xls|xlsx|zip)$"
 
 [match-license]
-pattern = "^(//+|#+)\\s+(.*)governed by a BSD-style(.*)$"
+pattern = "^(//+|#+|/\\*+|<!--+)?\\s*(.*)governed by a BSD-style(.*)$"
 license_identifier = BSD-3-Clause
-delete_line_pattern = "^(//+|#+)\\s+license that(.*)$"
+delete_line_pattern = "^(//+|#+|/\\*+|<!--+)?\\s*license that can(.*)$"
 
 [match-copyright]
-pattern = "^(//+|#+)\\s+Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<*(?<contact>.*)>.*$"
+pattern = "^(//+|#+|/\\*+|<!--+)?\\s*Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<(?<contact>.*)>.*$"
 `
 
 type config struct {
diff --git a/file.go b/file.go
index 7849d9e..798790c 100644
--- a/file.go
+++ b/file.go
@@ -187,18 +187,20 @@ func (f *file) scanCopyrightText(conv *SPDXConv) {
 			}
 			if cmc.match(string(line)) {
 				f.idxCopyrightText = x
+				f.copyrightYear = cmc.year
 				f.copyrightText = valMatch
 				return
 			}
 		}
 		for x, line := range f.bottomLines {
-			if reLicenseID.Match(line) {
+			if reCopyrightText.Match(line) {
 				f.idxCopyrightText = x - conv.cfg.MaxLineMatch
 				f.copyrightText = valExist
 				return
 			}
 			if cmc.match(string(line)) {
 				f.idxCopyrightText = x - conv.cfg.MaxLineMatch
+				f.copyrightYear = cmc.year
 				f.copyrightText = valMatch
 				return
 			}
diff --git a/spdxconv_test.go b/spdxconv_test.go
index 194276e..77929e8 100644
--- a/spdxconv_test.go
+++ b/spdxconv_test.go
@@ -64,14 +64,14 @@ func TestInit(t *testing.T) {
 			Pattern: `^.*\.(apk|app|bz2|csv|doc|docx|exe|gif|gz|jpeg|jpg|json|pdf|png|ppt|pptx|svg|svgz|tar|tgz|xls|xlsx|zip)$`,
 		}},
 		MatchLicense: []*matchLicense{{
-			Pattern:           `^(//+|#+)\s+(.*)governed by a BSD-style(.*)$`,
+			Pattern:           `^(//+|#+|/\*+|<!--+)?\s*(.*)governed by a BSD-style(.*)$`,
 			LicenseIdentifier: `BSD-3-Clause`,
 			DeleteLinePattern: []string{
-				`^(//+|#+)\s+license that(.*)$`,
+				`^(//+|#+|/\*+|<!--+)?\s*license that can(.*)$`,
 			},
 		}},
 		MatchCopyright: []*matchCopyright{{
-			Pattern: `^(//+|#+)\s+Copyright\s+(?<year>\d{4}),?\s+(?<author>.*)\s+<*(?<contact>.*)>.*$`,
+			Pattern: `^(//+|#+|/\*+|<!--+)?\s*Copyright\s+(?<year>\d{4}),?\s+(?<author>.*)\s+<(?<contact>.*)>.*$`,
 		}},
 	}
 	for _, mfc := range exp.MatchFileComment {
@@ -110,9 +110,9 @@ func TestScan(t *testing.T) {
 //spdxconv:version:v1
 //spdxconv:header:path,license_id,idx_license_id,year,copyright_id,idx_copyright_id
 //spdxconv:regular
-test.go,match,1,unknown,match,0,// ,
-test.html,default,0,unknown,default,0,<!-- ," -->"
-test.sh,match,1,unknown,match,0,# ,
+test.go,match,1,2022,match,0,// ,
+test.html,match,2,2022,match,1,<!-- ," -->"
+test.sh,match,1,2022,match,0,# ,
 //spdxconv:binary
 //spdxconv:unknown`
 
diff --git a/testdata/Apply_test.txt b/testdata/Apply_test.txt
index 611f654..123dc28 100644
--- a/testdata/Apply_test.txt
+++ b/testdata/Apply_test.txt
@@ -17,7 +17,7 @@ pattern = "^.*\\.(adoc|asciidoc|c|cc|cpp|cs|dart|go|h|hh|hpp|java|js|jsx|jsonc|k
 prefix = "//"
 
 [match-file-comment]
-pattern = "^.*\\.(bash|csh|dockerfile|env|gitignore|hcl|ipynb|make|pl|pm|py|ps1|rb|sh|tf|yaml|yml|zsh)$"
+pattern = "^.*\\.(aff|bash|csh|dockerfile|env|gitignore|hcl|ipynb|make|pl|pm|py|ps1|rb|sh|tf|yaml|yml|zsh)$"
 prefix = "#"
 
 [match-file-comment]
@@ -48,12 +48,12 @@ prefix = "%"
 pattern = "^.*\\.(apk|app|bz2|csv|doc|docx|exe|gif|gz|jpeg|jpg|json|pdf|png|ppt|pptx|svg|svgz|tar|tgz|xls|xlsx|zip)$"
 
 [match-license]
-pattern = "^(//+|#+)\\s+(.*)governed by a BSD-style(.*)$"
+pattern = "^(//+|#+|/\\*+|<!--+)?\\s*(.*)governed by a BSD-style(.*)$"
 license_identifier = BSD-3-Clause
-delete_line_pattern = "^(//+|#+)\\s+license that(.*)$"
+delete_line_pattern = "^(//+|#+|/\\*+|<!--+)?\\s*license that can(.*)$"
 
 [match-copyright]
-pattern = "^(//+|#+)\\s+Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<*(?<contact>.*)>.*$"
+pattern = "^(//+|#+|/\\*+|<!--+)?\\s*Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<(?<contact>.*)>.*$"
 
 >>> without_spdx_license_id.go
 1
@@ -162,5 +162,27 @@ pattern = "^(//+|#+)\\s+Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<*(?<
 2
 3
 
+>>> with.html
+<!--
+Copyright 2022, Shulhan <ms@kilabit.info>. All rights reserved.
+Use of this source code is governed by a BSD-style
+license that can be found in the LICENSE file.
+-->
+
+1
+2
+3
+
+<<< with.html
+<!-- SPDX-License-Identifier: BSD-3-Clause -->
+<!-- SPDX-FileCopyrightText: 2022 Shulhan <ms@kilabit.info> -->
+<!--
+
+-->
+
+1
+2
+3
+
 <<< END
 // REUSE-IgnoreEnd
diff --git a/testdata/scan/spdxconv.cfg b/testdata/scan/spdxconv.cfg
index 59ae864..3b137b6 100644
--- a/testdata/scan/spdxconv.cfg
+++ b/testdata/scan/spdxconv.cfg
@@ -43,9 +43,9 @@ prefix = "%"
 pattern = "^.*\\.(apk|app|bz2|csv|doc|docx|exe|gif|gz|jpeg|jpg|json|pdf|png|ppt|pptx|svg|svgz|tar|tgz|xls|xlsx|zip)$"
 
 [match-license]
-pattern = "^(//+|#+)\\s+(.*)governed by a BSD-style(.*)$"
+pattern = "^(//+|#+|/\\*+|<!--+)?\\s*(.*)governed by a BSD-style(.*)$"
 license_identifier = BSD-3-Clause
-delete_line_pattern = "^(//+|#+)\\s+license that(.*)$"
+delete_line_pattern = "^(//+|#+|/\\*+|<!--+)?\\s*license that can(.*)$"
 
 [match-copyright]
-pattern = "^(//+|#+)\\s+Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<*(?<contact>.*)>.*$"
+pattern = "^(//+|#+|/\\*+|<!--+)?\\s*Copyright\\s+(?<year>\\d{4}),?\\s+(?<author>.*)\\s+<(?<contact>.*)>.*$"