Added tests and implementation to fix idsearcher with long prefix lines

Signed-off-by: Steve Winslow <swinslow@gmail.com>

5 files changed, 64 insertions, 8 deletions

diff --git a/testdata/project4/has-id-to-ignore.txt b/testdata/project4/has-id-to-ignore.txt
new file mode 100644
index 0000000..bd63676
--- /dev/null
+++ b/testdata/project4/has-id-to-ignore.txt
@@ -0,0 +1,8 @@
+this file actually contains some code talking about IDs, and should be ignored
+
+if file.contains("SPDX-License-Identifier: GPL-2.0") {
+    // do something...
+
+}
+
+We don't want to pick up that line as the short-form ID for this file.
diff --git a/testdata/project4/has-id.txt b/testdata/project4/has-id.txt
new file mode 100644
index 0000000..5545f2c
--- /dev/null
+++ b/testdata/project4/has-id.txt
@@ -0,0 +1,3 @@
+// SPDX-License-Identifier: Apache-2.0 OR GPL-2.0-or-later
+
+somebody was kind enough to stick an spdx short-form ID in here
diff --git a/testdata/project4/has-mix-of-ids.txt b/testdata/project4/has-mix-of-ids.txt
new file mode 100644
index 0000000..b6aa5d1
--- /dev/null
+++ b/testdata/project4/has-mix-of-ids.txt
@@ -0,0 +1,12 @@
+// SPDX-License-Identifier: MIT
+
+the short-form ID above is correct, and should be picked up.
+
+the following code is talking about IDs, and should be ignored
+
+if file.contains("SPDX-License-Identifier: GPL-2.0") {
+    // do something...
+
+}
+
+We don't want to pick up that line as the short-form ID for this file.
diff --git a/v0/idsearcher/idsearcher.go b/v0/idsearcher/idsearcher.go
index 2eab298..e025f3b 100644
--- a/v0/idsearcher/idsearcher.go
+++ b/v0/idsearcher/idsearcher.go
@@ -150,15 +150,18 @@ func searchFileIDs(filePath string) ([]string, error) {
 
 	scanner := bufio.NewScanner(f)
 
-	// build the scan string this way, so that we can run idsearcher on itself
-	// without picking up these lines as IDs...
-	tag1 := "SPDX-License-"
-	tag2 := "Identifier:"
-	tag := tag1 + tag2
-
 	for scanner.Scan() {
-		if strings.Contains(scanner.Text(), tag) {
-			strs := strings.SplitAfterN(scanner.Text(), tag, 2)
+		if strings.Contains(scanner.Text(), "SPDX-License-Identifier:") {
+			strs := strings.SplitN(scanner.Text(), "SPDX-License-Identifier:", 2)
+
+			// if prefixed by more than n characters, it's probably not a
+			// short-form ID; it's probably code to detect short-form IDs.
+			// Like this function itself, for example  =)
+			prefix := stripTrash(strs[0])
+			if len(prefix) > 5 {
+				continue
+			}
+
 			// stop before trailing */ if it is present
 			lidToExtract := strs[1]
 			lidToExtract = strings.Split(lidToExtract, "*/")[0]
diff --git a/v0/idsearcher/idsearcher_test.go b/v0/idsearcher/idsearcher_test.go
index 4399faa..10178f9 100644
--- a/v0/idsearcher/idsearcher_test.go
+++ b/v0/idsearcher/idsearcher_test.go
@@ -364,6 +364,36 @@ func TestCanStripTrailingStarSlash(t *testing.T) {
 	}
 }
 
+func TestCanIgnoreShortFormIDWhenTooManyPrefixChars(t *testing.T) {
+	filePath := "../../testdata/project4/has-id-to-ignore.txt"
+
+	ids, err := searchFileIDs(filePath)
+	if err != nil {
+		t.Fatalf("expected nil error, got %v", err)
+	}
+
+	if len(ids) != 0 {
+		t.Fatalf("expected len 0, got %d", len(ids))
+	}
+}
+
+func TestCanPickJustTheRightID(t *testing.T) {
+	filePath := "../../testdata/project4/has-mix-of-ids.txt"
+
+	ids, err := searchFileIDs(filePath)
+	if err != nil {
+		t.Fatalf("expected nil error, got %v", err)
+	}
+
+	if len(ids) != 1 {
+		t.Fatalf("expected len 1, got %d", len(ids))
+	}
+
+	if ids[0] != "MIT" {
+		t.Errorf("expected %v, got %v", "MIT", ids[0])
+	}
+}
+
 func TestCannotFindShortFormIDWhenAbsent(t *testing.T) {
 	filePath := "../../testdata/project2/no-id.txt"