Keep legal files when vendoring tools (#8)

* Keep legal files when vendoring tools

* Bump version to v1.1.0

* Use a broader pile of legal file hints

clean.go

@@ -77,15 +77,12 @@ func clean(pkgs []string) {
 			return err
 		}
 
-		// Delete files in packages that aren't marked as "keep",
-		// and any non-go or test files.
+		// Delete files in packages that aren't in packages we need to build the
+		// tools, and any non-go, non-legal files in packages we *do* need.
 		if info.Mode().IsRegular() {
 			pkg = filepath.Dir(pkg)
 			_, keptPkg := keep[pkg]
-			isGo := strings.HasSuffix(path, ".go")
-			isAssembly := strings.HasSuffix(path, ".s")
-			isTest := strings.HasSuffix(path, "_test.go")
-			if !keptPkg || !(isGo || isAssembly) || isTest {
+			if !(keptPkg && keepFile(path)) {
 				toDelete = append(toDelete, path)
 			}
 			return nil
@@ -113,3 +110,55 @@ func clean(pkgs []string) {
 		}
 	}
 }
+
+func keepFile(filename string) bool {
+	if strings.HasSuffix(filename, "_test.go") {
+		return false
+	}
+	if strings.HasSuffix(filename, ".go") {
+		return true
+	}
+	if strings.HasSuffix(filename, ".s") {
+		return true
+	}
+	if isLegalFile(filename) {
+		return true
+	}
+	return false
+}
+
+var commonLegalFilePrefixes = []string{
+	"licence", // UK spelling
+	"license", // US spelling
+	"copying",
+	"unlicense",
+	"copyright",
+	"copyleft",
+	"authors",
+	"contributors",
+	"readme", // often has a license inline
+}
+
+var commonLegalFileSubstrings = []string{
+	"legal",
+	"notice",
+	"disclaimer",
+	"patent",
+	"third-party",
+	"thirdparty",
+}
+
+func isLegalFile(filename string) bool {
+	base := strings.ToLower(filepath.Base(filename))
+	for _, p := range commonLegalFilePrefixes {
+		if strings.HasPrefix(base, p) {
+			return true
+		}
+	}
+	for _, s := range commonLegalFileSubstrings {
+		if strings.Contains(base, s) {
+			return true
+		}
+	}
+	return false
+}

clean_test.go

@@ -0,0 +1,54 @@
+package main
+
+import (
+	"path/filepath"
+	"testing"
+)
+
+func TestIsLegalFile(t *testing.T) {
+	testcase := func(filename string, want bool) {
+		t.Run(filename, func(t *testing.T) {
+			have := isLegalFile(filename)
+			if have != want {
+				t.Fail()
+			}
+		})
+	}
+
+	testcase("license.md", true)
+	testcase("license.txt", true)
+	testcase("LICENSE", true)
+	testcase("LICENCE", true)
+	testcase("LICENSE.md", true)
+	testcase(filepath.Join("pkg", "LICENSE.md"), true)
+	testcase("LEGAL", true)
+	testcase("README", true)
+	testcase("COPYING", true)
+	testcase("COPYRIGHT", true)
+	testcase("UNLICENSE", true)
+	testcase("PATENTS", true)
+
+	testcase(filepath.Join("pat", "ents"), false)
+	testcase("picture.jpeg", false)
+}
+
+func TestKeep(t *testing.T) {
+	testcase := func(filename string, want bool) {
+		t.Run(filename, func(t *testing.T) {
+			have := keepFile(filename)
+			if have != want {
+				t.Fail()
+			}
+		})
+	}
+
+	testcase("program.go", true)
+	testcase("program_test.go", false)
+	testcase(filepath.Join("pkg", "program.go"), true)
+	testcase(filepath.Join("pkg", "program_test.go"), false)
+
+	testcase("assembly.s", true)
+	testcase("notassembly.as", false)
+	testcase("picture.gif", false)
+	testcase("LICENSE.md", true)
+}

main.go

@@ -8,7 +8,7 @@ import (
 	"path/filepath"
 )
 
-const version = "v1.0.3"
+const version = "v1.1.0"
 
 var cacheDir = ""
 

retool_test.go

@@ -121,6 +121,12 @@ func TestRetool(t *testing.T) {
 
 		// Now the binary should be installed
 		assertBinInstalled(t, dir, "retool")
+
+		// Legal files should be kept around
+		_, err = os.Stat(filepath.Join(dir, "_tools", "src", "github.com", "twitchtv", "retool", "LICENSE"))
+		if err != nil {
+			t.Error("missing license file")
+		}
 	})
 
 	t.Run("dep_added", func(t *testing.T) {