Inventory: include any repo with issues+PRs enabled instead of a fork allow-list

README.md

@@ -329,13 +329,13 @@ Instead of hand-maintaining the repo list, a repolist file can define
 puppetsync::repos_source:
   org: simp
   include: ['pupmod-*', 'puppet-*', 'rubygem-*']  # name globs
-  include_forks:                                  # forks the org maintains
-    - rubygem-simp-rspec-puppet-facts
-    - pupmod-voxpupuli-selinux
 ```
 
-* Archived and empty repos are excluded; forks are excluded unless they
-  match `include_forks`
+* Archived and empty repos are excluded
+* Repos with **issues or pull requests disabled** are excluded: the org
+  disables both on every fork that exists only as a mirror, so those flags
+  separate maintained repos (forks included) from mirrors — no fork
+  allow-list needed
 * Repos with the `puppetsync-ignore` GitHub topic are excluded — set the
   topic to opt a repo out without touching puppetsync
 * Each repo's branch comes from the API's default branch (so e.g.

data/sync/repolists/github-org.yaml

@@ -9,8 +9,10 @@
 # Filtering notes:
 #   - Archived repos and empty repos are always excluded (exclude_archived
 #     defaults to true)
-#   - Forks are excluded unless listed in include_forks — the escape hatch
-#     for forks the org actively maintains
+#   - Repos with issues OR pull requests disabled are excluded: the org
+#     disables both on every fork that exists only as a mirror, so those
+#     two flags are the signal separating maintained repos (forks
+#     included) from mirrors — no fork allow-list needed
 #   - Repos with the 'puppetsync-ignore' topic are excluded by default: set
 #     that topic on a repo in GitHub to opt it out without touching this file
 #   - Names are matched with globs; project_type filtering (the config's
@@ -22,10 +24,11 @@ puppetsync::repos_source:
     - 'pupmod-*'
     - 'puppet-*'
     - 'rubygem-*'
-  exclude: []
-  # Forks that simp actively maintains:
-  # (pupmod-voxpupuli-selinux is deliberately absent for now: its CI isn't
-  # green yet, and puppetsync can't open PRs from a fork-network sibling
-  # like silug/puppet-selinux until simp/puppetsync#83 is fixed)
-  include_forks:
-    - rubygem-simp-rspec-puppet-facts
+  exclude:
+    # Not ready for puppetsync yet: CI isn't green, and the GitHub PR
+    # stages can't use a fork-network sibling like silug/puppet-selinux
+    # until simp/puppetsync#83 is fixed
+    - pupmod-voxpupuli-selinux
+    # Never received the new baseline, and likely to be archived soon
+    # (archival will make this entry redundant)
+    - pupmod-simp-haveged

dist/puppetsync/tasks/list_github_repos.rb

@@ -6,10 +6,12 @@
 # Filtering (all driven by the `source` parameter):
 #
 #   - Archived repos are excluded (`exclude_archived`, default true)
-#   - Forks are excluded (`exclude_forks`, default true) UNLESS the repo
-#     matches an `include_forks` glob — the escape hatch for forks the org
-#     actively maintains (e.g. rubygem-simp-rspec-puppet-facts,
-#     pupmod-voxpupuli-selinux)
+#   - Repos with issues OR pull requests DISABLED are excluded:
+#     puppetsync's whole output is a PR, and the simp org disables
+#     issues+PRs on every fork that exists only as a mirror — so either
+#     flag being off separates mirrors from maintained repos (forks
+#     included). (The old `exclude_forks`/`include_forks` allow-list is
+#     retired.)
 #   - Repos with any topic in `exclude_topics` are excluded (default:
 #     ['puppetsync-ignore'] — set that topic on a repo in GitHub to opt it
 #     out without touching puppetsync)
@@ -79,24 +81,25 @@ def select_repos(repos, source)
   include_topics   = source.fetch('include_topics', nil) || []
   exclude_topics   = source.fetch('exclude_topics', nil) || ['puppetsync-ignore']
   exclude_archived = source.fetch('exclude_archived', true)
-  exclude_forks    = source.fetch('exclude_forks', true)
-  include_forks    = source.fetch('include_forks', nil) || []
+  if source.key?('exclude_forks') || source.key?('include_forks')
+    warn '== WARNING: exclude_forks/include_forks are retired and ignored — ' \
+         'inventory now includes any repo with issues+PRs enabled ' \
+         '(mirrors have them disabled org-wide)'
+  end
 
   repos.select do |repo|
     name = repo['name']
     topics = repo['topics'] || []
     next false if exclude_archived && repo['archived']
     next false if repo['size'].to_i.zero?
+    # puppetsync's whole output is a PR; a repo with PRs (or issues)
+    # disabled is a mirror (the org turns both off on mirror forks).
+    # Absent fields => include, so older API responses/fixtures don't
+    # silently empty the inventory.
+    next false unless repo.fetch('has_pull_requests', true) && repo.fetch('has_issues', true)
     next false if glob_match?(name, exclude_globs)
     next false if topics.any? { |topic| exclude_topics.include?(topic) }
 
-    # A fork listed in include_forks is an explicit allow-list entry: it
-    # both clears the fork gate AND bypasses the include filters below
-    # (listing it means "I want this fork", whatever its name)
-    explicitly_included_fork = repo['fork'] && glob_match?(name, include_forks)
-    next false if repo['fork'] && exclude_forks && !explicitly_included_fork
-    next true if explicitly_included_fork
-
     glob_match?(name, include_globs) || topics.any? { |topic| include_topics.include?(topic) }
   end
 end

spec/tasks/list_github_repos_spec.rb

@@ -5,23 +5,41 @@
   let(:org_repos) do
     [
       { 'name' => 'pupmod-simp-aide', 'full_name' => 'simp/pupmod-simp-aide',
-        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 500 },
+        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 500,
+        'has_issues' => true, 'has_pull_requests' => true },
       { 'name' => 'puppet-gpasswd', 'full_name' => 'simp/puppet-gpasswd',
-        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 200 },
+        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 200,
+        'has_issues' => true, 'has_pull_requests' => true },
       { 'name' => 'pupmod-simp-ntpd', 'full_name' => 'simp/pupmod-simp-ntpd',
-        'fork' => false, 'archived' => true, 'default_branch' => 'master', 'topics' => [], 'size' => 300 },
+        'fork' => false, 'archived' => true, 'default_branch' => 'master', 'topics' => [], 'size' => 300,
+        'has_issues' => true, 'has_pull_requests' => true },
+      # Maintained forks: issues + PRs enabled
       { 'name' => 'rubygem-simp-rspec-puppet-facts', 'full_name' => 'simp/rubygem-simp-rspec-puppet-facts',
-        'fork' => true, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 100 },
+        'fork' => true, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 100,
+        'has_issues' => true, 'has_pull_requests' => true },
       { 'name' => 'pupmod-voxpupuli-selinux', 'full_name' => 'simp/pupmod-voxpupuli-selinux',
-        'fork' => true, 'archived' => false, 'default_branch' => 'simp-master', 'topics' => [], 'size' => 400 },
+        'fork' => true, 'archived' => false, 'default_branch' => 'simp-master', 'topics' => [], 'size' => 400,
+        'has_issues' => true, 'has_pull_requests' => true },
+      # Mirror fork: issues + PRs disabled org-wide
       { 'name' => 'puppetlabs-apache', 'full_name' => 'simp/puppetlabs-apache',
-        'fork' => true, 'archived' => false, 'default_branch' => 'main', 'topics' => [], 'size' => 900 },
+        'fork' => true, 'archived' => false, 'default_branch' => 'main', 'topics' => [], 'size' => 900,
+        'has_issues' => false, 'has_pull_requests' => false },
+      # Only ONE of the two flags off still means mirror-ish: skip
+      { 'name' => 'pupmod-simp-halfmirror', 'full_name' => 'simp/pupmod-simp-halfmirror',
+        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 70,
+        'has_issues' => false, 'has_pull_requests' => true },
       { 'name' => 'pupmod-simp-optout', 'full_name' => 'simp/pupmod-simp-optout',
-        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => ['puppetsync-ignore'], 'size' => 100 },
+        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => ['puppetsync-ignore'], 'size' => 100,
+        'has_issues' => true, 'has_pull_requests' => true },
       { 'name' => 'topic-tagged-tool', 'full_name' => 'simp/topic-tagged-tool',
-        'fork' => false, 'archived' => false, 'default_branch' => 'main', 'topics' => ['simp-baseline'], 'size' => 50 },
+        'fork' => false, 'archived' => false, 'default_branch' => 'main', 'topics' => ['simp-baseline'], 'size' => 50,
+        'has_issues' => true, 'has_pull_requests' => true },
+      # No flags at all (older API shape): must still be included
+      { 'name' => 'pupmod-simp-legacyfields', 'full_name' => 'simp/pupmod-simp-legacyfields',
+        'fork' => false, 'archived' => false, 'default_branch' => 'master', 'topics' => [], 'size' => 60 },
       { 'name' => 'empty-repo', 'full_name' => 'simp/empty-repo',
-        'fork' => false, 'archived' => false, 'default_branch' => 'main', 'topics' => [], 'size' => 0 },
+        'fork' => false, 'archived' => false, 'default_branch' => 'main', 'topics' => [], 'size' => 0,
+        'has_issues' => true, 'has_pull_requests' => true },
     ]
   end
 
@@ -35,47 +53,42 @@ def urls(result)
     result['repos_config'].keys
   end
 
-  it 'excludes archived repos, forks, empty and opted-out repos by default' do
+  it 'includes maintained forks and excludes mirrors, archived, empty, and opted-out repos by default' do
     result = run_list({ 'org' => 'simp' })
 
     expect(urls(result)).to contain_exactly(
       'https://github.com/simp/pupmod-simp-aide',
       'https://github.com/simp/puppet-gpasswd',
       'https://github.com/simp/topic-tagged-tool',
+      # forks with issues+PRs enabled are maintained repos, not mirrors
+      'https://github.com/simp/rubygem-simp-rspec-puppet-facts',
+      'https://github.com/simp/pupmod-voxpupuli-selinux',
+      # missing flags (older API shape) must not exclude
+      'https://github.com/simp/pupmod-simp-legacyfields',
     )
   end
 
-  it 'keeps forks matching include_forks (the maintained-fork exceptions)' do
-    result = run_list(
-      'org' => 'simp',
-      'include_forks' => ['rubygem-simp-rspec-puppet-facts', 'pupmod-voxpupuli-selinux'],
-    )
+  it 'skips a repo when EITHER issues or pull requests are disabled' do
+    result = run_list({ 'org' => 'simp' })
 
-    expect(urls(result)).to include(
-      'https://github.com/simp/rubygem-simp-rspec-puppet-facts',
-      'https://github.com/simp/pupmod-voxpupuli-selinux',
-    )
-    expect(urls(result)).not_to include('https://github.com/simp/puppetlabs-apache')
+    expect(urls(result)).not_to include('https://github.com/simp/puppetlabs-apache')     # both off
+    expect(urls(result)).not_to include('https://github.com/simp/pupmod-simp-halfmirror') # issues off
   end
 
-  it 'include_forks alone suffices: listed forks bypass the include globs' do
-    # Mirrors the shipped github-org.yaml shape, but with include globs that
-    # do NOT match one of the fork exceptions
-    result = run_list(
-      'org' => 'simp',
-      'include' => ['pupmod-*'],
-      'include_forks' => ['rubygem-simp-rspec-puppet-facts'],
-    )
+  it 'warns that the retired include_forks/exclude_forks keys are ignored' do
+    stdout, stderr, status = run_task('list_github_repos.rb',
+                                      'source' => { 'org' => 'simp', 'include_forks' => ['x'] }, 'repos' => org_repos)
 
-    expect(urls(result)).to include('https://github.com/simp/rubygem-simp-rspec-puppet-facts')
-    expect(urls(result)).not_to include('https://github.com/simp/puppet-gpasswd') # non-fork still needs include
+    expect(status).to be_success, stderr
+    expect(stderr).to include('retired and ignored')
+    # ...and they really are ignored: mirrors stay out, maintained forks stay in
+    result = JSON.parse(stdout)
+    expect(result['repos_config'].keys).to include('https://github.com/simp/pupmod-voxpupuli-selinux')
+    expect(result['repos_config'].keys).not_to include('https://github.com/simp/puppetlabs-apache')
   end
 
   it 'takes each branch from the API default_branch' do
-    result = run_list(
-      'org' => 'simp',
-      'include_forks' => ['pupmod-voxpupuli-selinux'],
-    )
+    result = run_list({ 'org' => 'simp' })
 
     expect(result['repos_config']['https://github.com/simp/pupmod-voxpupuli-selinux']).to eq('branch' => 'simp-master')
     expect(result['repos_config']['https://github.com/simp/pupmod-simp-aide']).to eq('branch' => 'master')
@@ -87,6 +100,8 @@ def urls(result)
     expect(urls(result)).to contain_exactly(
       'https://github.com/simp/pupmod-simp-aide',
       'https://github.com/simp/puppet-gpasswd',
+      'https://github.com/simp/pupmod-voxpupuli-selinux',
+      'https://github.com/simp/pupmod-simp-legacyfields',
     )
   end