ci-fairy: Support checking commits for required/disallowed patterns

In mutter and gnome-shell, we use a custom script to check that commit
messages conform to a particular pattern/style. This seems generally
useful, so add the option to provide additional patterns in a YAML file.
Signed-off-by: Florian Müllner <fmuellner@gnome.org>

.gitlab-ci.yml

@@ -154,7 +154,7 @@ ci-fairy images:
 
 .ci-fairy-tag:
   variables:
-    FDO_DISTRIBUTION_TAG: sha256-936060119966b261c07af69d9422ee970cb2f37a355ba8509dceeaee70633349
+    FDO_DISTRIBUTION_TAG: sha256-754a2c72cf4a7e258ba19bc895ad749ddac120842b7121355d9a3586cb9ea942
 
 .ci-fairy-local-image:
   extends:
@@ -292,7 +292,7 @@ test published images:
   script:
     - skopeo inspect docker://quay.io/freedesktop.org/ci-templates:buildah-aarch64-2020-12-18.2
     - skopeo inspect docker://quay.io/freedesktop.org/ci-templates:buildah-x86_64-2020-12-18.2
-    - skopeo inspect docker://quay.io/freedesktop.org/ci-templates:ci-fairy-sha256-936060119966b261c07af69d9422ee970cb2f37a355ba8509dceeaee70633349
+    - skopeo inspect docker://quay.io/freedesktop.org/ci-templates:ci-fairy-sha256-754a2c72cf4a7e258ba19bc895ad749ddac120842b7121355d9a3586cb9ea942
     - skopeo inspect docker://quay.io/freedesktop.org/ci-templates:qemu-base-x86_64-2020-12-18.2
     - skopeo inspect docker://quay.io/freedesktop.org/ci-templates:qemu-mkosi-base-x86_64-2020-12-18.2
   rules:

.gitlab-ci/ci-fairy-ci.yml

@@ -23,7 +23,7 @@ ci-fairy-base-image:
 
 .ci-fairy-tag:
   variables:
-    FDO_DISTRIBUTION_TAG: sha256-936060119966b261c07af69d9422ee970cb2f37a355ba8509dceeaee70633349
+    FDO_DISTRIBUTION_TAG: sha256-754a2c72cf4a7e258ba19bc895ad749ddac120842b7121355d9a3586cb9ea942
 
 # The actual ci-fairy image with ci-fairy installed
 # This image uses the sha of the ci-fairy script itself as tag.

doc/ci-fairy.rst

@@ -264,6 +264,20 @@ some cosmetic requirements. Other options to check for is whether a
 a line is **not** included (``--no-signed-off-by``). See the ``--help``
 output for a list fo checks available.
 
+It is possible to provide custom rule definitions in YAML format, either in
+``.gitlab-ci/commitrules.yml`` if it exists, or by specifying a file with the
+``--rules-file`` option. An example rules files:
+
+.. code-block:: yaml
+    patterns:
+      require:
+        - regex: ($CI_MERGE_REQUEST_PROJECT_URL/(-/)?(issues|merge_requests)/[0-9]+)
+          message: Commit message must contain a link to an issue or merge request
+      deny:
+        - regex: '^\S*\.[ch]:'
+          message: Commit message subject prefix should not include .c, .h etc.
+          where: subject
+
 ``ci-fairy`` uses the CI environment where available and will do the right
 thing without requiring options. It does need the ``$FDO_UPSTREAM_REPO``
 variable to be set to the upstream project's full name. An example

templates/ci-fairy.yml

@@ -31,6 +31,6 @@
 # Variables provided by this template should be considered read-only.
 #
 .fdo.ci-fairy:
-  image: quay.io/freedesktop.org/ci-templates:ci-fairy-sha256-936060119966b261c07af69d9422ee970cb2f37a355ba8509dceeaee70633349
+  image: quay.io/freedesktop.org/ci-templates:ci-fairy-sha256-754a2c72cf4a7e258ba19bc895ad749ddac120842b7121355d9a3586cb9ea942
   variables:
-    FDO_DISTRIBUTION_IMAGE: quay.io/freedesktop.org/ci-templates:ci-fairy-sha256-936060119966b261c07af69d9422ee970cb2f37a355ba8509dceeaee70633349
\ No newline at end of file
+    FDO_DISTRIBUTION_IMAGE: quay.io/freedesktop.org/ci-templates:ci-fairy-sha256-754a2c72cf4a7e258ba19bc895ad749ddac120842b7121355d9a3586cb9ea942
\ No newline at end of file

tools/ci_fairy.py

@@ -15,6 +15,7 @@ import json
 import fnmatch
 import logging
 import os
+import re
 import shutil
 import sys
 import time
@@ -220,6 +221,27 @@ class GitCommitValidator(object):
             if not is_revert:
                 self.__error(f'Commit message subject must not exceed {tw} characters')
 
+    def _match_pattern_rule(self, subject, rule):
+        if rule.get('where', 'message') == 'subject':
+            msg = subject
+        else:
+            msg = self.commit.message
+
+        regex = str(rule['regex'])
+        return re.search(regex, msg, re.MULTILINE)
+
+    def check_patterns(self, rules):
+        subject = self.commit.message.split('\n')[0]
+        patterns = rules.get('patterns', {})
+
+        for rule in patterns.get('require', []):
+            if self._match_pattern_rule(subject, rule) is None:
+                self.__error(rule.get('message', f'Required pattern not found: {rule["regex"]}'))
+
+        for rule in patterns.get('deny', []):
+            if self._match_pattern_rule(subject, rule) is not None:
+                self.__error(rule.get('message', f'Disallowed pattern found: {rule["regex"]}'))
+
     @property
     def failed(self):
         return bool(self.errors)
@@ -1011,9 +1033,11 @@ def generate_template(ctx, config, root, template, output_file, verify):
               help='Require a Signed-off-by tag to be (or not be) present')
 @click.option('--textwidth', type=int, default=80,
               help='Require the commit message subject to be less than N characters wide.')
+@click.option('--rules-file', type=click.File(), required=False,
+              help='Custom rules definitions in YAML format')
 @click.option('--junit-xml', help='junit output file to write to')
 @click.pass_context
-def check_commits(ctx, commit_range, branch, signed_off_by, textwidth, junit_xml):
+def check_commits(ctx, commit_range, branch, signed_off_by, textwidth, rules_file, junit_xml):
     '''
     Check a commit range for some properties. Some checks are always
     performed, others can be changed with commandline arguments.
@@ -1061,6 +1085,17 @@ def check_commits(ctx, commit_range, branch, signed_off_by, textwidth, junit_xml
         commit_range = f'{remote_prefix}{target_branch}..{sha}'
     logger.debug(f'Using commit range {commit_range}')
 
+    if rules_file is None:
+        default_rules = Path('.gitlab-ci/commitrules.yml')
+        if default_rules.exists():
+            rules_file = default_rules.open()
+
+    if rules_file is not None:
+        yml = os.path.expandvars(rules_file.read())
+        rules = yaml.safe_load(yml)
+    else:
+        rules = None
+
     failures = []
     for commit in repo.iter_commits(commit_range):
         validator = GitCommitValidator(commit)
@@ -1069,6 +1104,8 @@ def check_commits(ctx, commit_range, branch, signed_off_by, textwidth, junit_xml
             validator.check_sob(signed_off_by)
         if textwidth > 0:
             validator.check_text_width(textwidth)
+        if rules is not None:
+            validator.check_patterns(rules)
 
         if validator.failed:
             failures.append(validator)

tools/test_ci_fairy.py

@@ -834,6 +834,15 @@ def test_commits_sob(caplog, gitlab_default_env):
         assert 'Missing "Signed-off-by: author information"' in ''.join(caplog.messages)
 
 
+FORMAT_RULES = '''
+patterns:
+  deny:
+    - regex: '^[^:]*: [a-z]'
+      message: Commit message subject should be properly Capitalized.
+      where: subject
+'''
+
+
 def test_commits_msgformat(caplog, gitlab_default_env):
     env = gitlab_default_env
     env['CI'] = ''
@@ -915,6 +924,29 @@ def test_commits_msgformat(caplog, gitlab_default_env):
         # Revert line may have a longer line than allowed
         assert result.exit_code == 0
 
+        b = repo.create_head('mybranch5', 'master')
+        repo.head.reference = b
+        repo.head.reset(index=True, working_tree=True)
+        with open('test', 'w') as fd:
+            fd.write('bar')
+        repo.index.add(['test'])
+        repo.index.commit('foo: bar')
+
+        result = runner.invoke(ci_fairy.ci_fairy, args + ['--rules-file', '-'], input=FORMAT_RULES)
+        assert result.exit_code == 1
+        assert 'Commit message subject should be properly Capitalized.' in ''.join(caplog.messages)
+
+        b = repo.create_head('mybranch6', 'master')
+        repo.head.reference = b
+        repo.head.reset(index=True, working_tree=True)
+        with open('test', 'w') as fd:
+            fd.write('bar')
+        repo.index.add(['test'])
+        repo.index.commit('foo: Bar\n\nfoo: bar')
+
+        result = runner.invoke(ci_fairy.ci_fairy, args + ['--rules-file', '-'], input=FORMAT_RULES)
+        assert result.exit_code == 0
+
 
 def test_commits_gitlabemail(caplog, gitlab_default_env):
     env = gitlab_default_env