not-owned-checker: Add git-ls-tree implementation with subdirectory s…

…upport (#141)

README.md

@@ -131,6 +131,7 @@ Use the following environment variables to configure the application:
 | <tt>OWNER_CHECKER_ALLOW_UNOWNED_PATTERNS</tt> | `true`                        | 	Specifies whether CODEOWNERS may have unowned files. For example: <br> <br>  `/infra/oncall-rotator/                    @sre-team` <br>  `/infra/oncall-rotator/oncall-config.yml` <br> <br>  The `/infra/oncall-rotator/oncall-config.yml` file is not owned by anyone.                                                                                                                                                                                       |
 | <tt>OWNER_CHEKER_OWNERS_MUST_BE_TEAMS</tt>    | `false`                       | Specifies whether only teams are allowed as owners of files.                                                                                                                                                                                                                                                                                                                                                                                                    |
 | <tt>NOT_OWNED_CHECKER_SKIP_PATTERNS</tt>      | -                             | The comma-separated list of patterns that should be ignored by `not-owned-checker`. For example, you can specify `*` and as a result, the `*` pattern from the **CODEOWNERS** file will be ignored and files owned by this pattern will be reported as unowned unless a later specific pattern will match that path. It's useful because often we have default owners entry at the begging of the CODOEWNERS file, e.g. `*       @global-owner1 @global-owner2` |
+| <tt>NOT_OWNED_CHECKER_SUBDIRECTORIES</tt>      | -                             | The comma-separated list of subdirectories to check in `not-owned-checker`. When specified, only files in the listed subdirectories will be checked if they do not have specified owners in CODEOWNERS.                                                                                                                                                                                                                                                                                                    |
 
  <b>*</b> - Required
 

action.yml

@@ -57,6 +57,10 @@ inputs:
     default: "false"
     required: false
 
+  not_owned_checker_subdirectories:
+    description:  "Only check listed subdirectories for CODEOWNERS ownership that don't have owners."
+    required: false
+
 runs:
   using: 'docker'
   image: 'docker://ghcr.io/mszostok/codeowners-validator:v0.7.2'

docs/gh-action.md

@@ -73,6 +73,9 @@ jobs:
 
           # Specifies whether only teams are allowed as owners of files.
           owner_checker_owners_must_be_teams: "false"
+
+          # Only check listed subdirectories for CODEOWNERS ownership that don't have owners.
+          not_owned_checker_subdirectories: ""
 ```
 
 The best is to run this as a cron job and not only if you applying changes to CODEOWNERS file itself, e.g. the CODEOWNERS file can be invalidate when you removing someone from the organization.

internal/check/not_owned_file.go

@@ -16,11 +16,13 @@ import (
 )
 
 type NotOwnedFileConfig struct {
-	SkipPatterns []string `envconfig:"optional"`
+	SkipPatterns   []string `envconfig:"optional"`
+	Subdirectories []string `envconfig:"optional"`
 }
 
 type NotOwnedFile struct {
-	skipPatterns map[string]struct{}
+	skipPatterns   map[string]struct{}
+	subDirectories []string
 }
 
 func NewNotOwnedFile(cfg NotOwnedFileConfig) *NotOwnedFile {
@@ -30,7 +32,8 @@ func NewNotOwnedFile(cfg NotOwnedFileConfig) *NotOwnedFile {
 	}
 
 	return &NotOwnedFile{
-		skipPatterns: skip,
+		skipPatterns:   skip,
+		subDirectories: cfg.Subdirectories,
 	}
 }
 
@@ -130,7 +133,7 @@ func (c *NotOwnedFile) GitRemoveIgnoredFiles(repoDir string) error {
 		pipe.ChDir(repoDir),
 		pipe.Line(
 			pipe.Exec("git", "ls-files", "-ci", "--exclude-standard", "-z"),
-			pipe.Exec("xargs", "-0", "git", "rm", "--cached"),
+			pipe.Exec("xargs", "-0", "-r", "git", "rm", "--cached"),
 		),
 	)
 
@@ -168,9 +171,12 @@ func (c *NotOwnedFile) GitResetCurrentBranch(repoDir string) error {
 }
 
 func (c *NotOwnedFile) GitListFiles(repoDir string) (string, error) {
+	args := []string{"ls-files"}
+	args = append(args, c.subDirectories...)
+
 	gitls := pipe.Script(
 		pipe.ChDir(repoDir),
-		pipe.Exec("git", "ls-files"),
+		pipe.Exec("git", args...),
 	)
 
 	stdout, stderr, err := pipe.DividedOutput(gitls)

tests/integration/integration_test.go

@@ -215,6 +215,16 @@ func TestCheckFailures(t *testing.T) {
 				"NOT_OWNED_CHECKER_SKIP_PATTERNS": "*",
 			},
 		},
+		{
+			name: "notowned_sub_dirs",
+			envs: Envs{
+				"PATH":                             os.Getenv("PATH"), // need to be set to find the `git` binary
+				"CHECKS":                           "disable-all",
+				"EXPERIMENTAL_CHECKS":              "notowned",
+				"NOT_OWNED_CHECKER_SKIP_PATTERNS":  "*",
+				"NOT_OWNED_CHECKER_SUBDIRECTORIES": "notowned/dir",
+			},
+		},
 	}
 	for _, tc := range tests {
 		t.Run(tc.name, func(t *testing.T) {

tests/integration/testdata/TestCheckFailures/notowned.golden.txt

@@ -1,7 +1,8 @@
 ==> Executing [Experimental] Not Owned File Checker (<duration>)
-    [err] Found 3 not owned files (skipped patterns: "*"):
+    [err] Found 4 not owned files (skipped patterns: "*"):
             * .gitignore
             * CODEOWNERS
             * action.yml
+            * notowned/dir/example/sample.txt
 
 1 check(s) executed, 1 failure(s)

tests/integration/testdata/TestCheckFailures/notowned_sub_dirs.golden.txt

@@ -0,0 +1,5 @@
+==> Executing [Experimental] Not Owned File Checker (<duration>)
+    [err] Found 1 not owned files (skipped patterns: "*"):
+            * notowned/dir/example/sample.txt
+
+1 check(s) executed, 1 failure(s)