ci: fix AGPL license bot (runner label alpine-latest, HTTP scheme, amend + comment)

.gitea/workflows/license.yml

@@ -1,13 +1,30 @@
 name: license-headers
 
+# Enforce AGPL-3.0 headers on every file ADDED by a pull request: amend the PR
+# head commit with the headers and comment the new commit hash on the PR.
+#
+# `runs-on` must match a registered runner label. The org's only runner offers
+# `alpine-latest` (container `node:alpine`), so jobs asking for anything else
+# (e.g. `alpine-3.24`) queue forever — that is why this bot never ran before.
+#
+# No long-lived secret is needed or wanted here: Gitea mints a per-run token
+# (`secrets.GITHUB_TOKEN`) scoped to this repository, which is enough to amend
+# the PR branch and comment. Never add a PAT/SSH key — an action that runs on
+# PRs (including forks) must stay keyless, and Gitea withholds repo/org secrets
+# from fork PRs anyway.
+
 on:
-  push:
-    branches: ['**']
   pull_request:
+    types: [opened, synchronize, reopened, ready_for_review]
 
 permissions:
   contents: write
   issues: write
+  pull-requests: read
+
+concurrency:
+  group: license-headers-${{ gitea.event.pull_request.number }}
+  cancel-in-progress: true
 
 defaults:
   run:
@@ -15,104 +32,143 @@ defaults:
 
 jobs:
   addlicense:
-    runs-on: alpine-3.24
+    runs-on: alpine-latest
     steps:
       - name: Install prerequisites
         run: apk add --no-cache git curl ca-certificates jq
 
-      - name: Checkout
+      - name: Checkout pull request head branch
         env:
           TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
           set -e
-          REF="${{ gitea.event.pull_request.head.ref || gitea.ref_name }}"
+          REF="${{ gitea.event.pull_request.head.ref }}"
+          HEAD_REPO="${{ gitea.event.pull_request.head.repo.full_name }}"
+          BASE_REPO="${{ gitea.repository }}"
+          if [ "$HEAD_REPO" != "$BASE_REPO" ]; then
+            echo "Fork PR ($HEAD_REPO -> $BASE_REPO): the bot cannot push to a fork; skipping."
+            echo "fork_pr=true" >> "$GITHUB_ENV"
+            exit 0
+          fi
+          # Keep the scheme: the dev Gitea is HTTP-only, so a hardcoded https://
+          # here fails with "TLS connect error: wrong version number".
           SERVER="${{ gitea.server_url }}"
-          SERVER="${SERVER#https://}"; SERVER="${SERVER#http://}"
+          SERVER="${SERVER%/}"
+          SCHEME="${SERVER%%://*}"
+          HOST="${SERVER#*://}"
           git clone --depth 1 --branch "$REF" \
-            "https://oauth2:${TOKEN}@${SERVER}/${{ gitea.repository }}.git" .
+            "${SCHEME}://oauth2:${TOKEN}@${HOST}/${BASE_REPO}.git" .
+          # Lease value for the later force push: the exact remote head we based on.
+          git ls-remote origin "refs/heads/${REF}" | cut -f1 > /tmp/head_before.txt
+          echo "checked out ${REF} at $(git rev-parse HEAD)"
 
       - name: Install addlicense
+        if: env.fork_pr != 'true'
         run: |
           set -e
           curl -sSL -o /tmp/addlicense.tar.gz \
             https://github.com/google/addlicense/releases/download/v1.2.0/addlicense_v1.2.0_Linux_x86_64.tar.gz
           tar -xzf /tmp/addlicense.tar.gz -C /tmp
 
-      - name: Add AGPL-3.0 headers
+      - name: Add AGPL-3.0 headers to newly added files
-        run: |
+        if: env.fork_pr != 'true'
-          set -e
-          set -- -f .license-header.tpl -c "Dcntrlized" -y "$(date +%Y)"
-          while IFS= read -r line; do
-            case "$line" in
-              "" | \#*) continue ;;
-            esac
-            set -- "$@" -ignore "$line"
-          done < .licenseignore
-          /tmp/addlicense "$@" .
-
-      - name: Commit and push changes
         env:
           TOKEN: ${{ secrets.GITHUB_TOKEN }}
         run: |
           set -e
-          REF="${{ gitea.event.pull_request.head.ref || gitea.ref_name }}"
           SERVER="${{ gitea.server_url }}"
-          SERVER="${SERVER#https://}"; SERVER="${SERVER#http://}"
+          SERVER="${SERVER%/}"
+          REPO="${{ gitea.repository }}"
+          PR="${{ gitea.event.pull_request.number }}"
+
+          # Files added by this PR, straight from the API (no base history needed).
+          : > /tmp/added.txt
+          page=1
+          while : ; do
+            body="$(curl -sS -H "Authorization: token ${TOKEN}" \
+              "${SERVER}/api/v1/repos/${REPO}/pulls/${PR}/files?limit=100&page=${page}")"
+            n="$(printf '%s' "$body" | jq 'length')"
+            [ "$n" = "0" ] && break
+            printf '%s' "$body" | jq -r '.[] | select(.status=="added") | .filename' >> /tmp/added.txt
+            [ "$n" -lt 100 ] && break
+            page=$((page + 1))
+          done
+
+          # addlicense validates -ignore with doublestar and silently skips file
+          # types it does not understand, so pass every added file and let
+          # .licenseignore be the only escape hatch.
+          set -- -v -f .license-header.tpl -c "Dcntrlized" -y "$(date +%Y)"
+          if [ -f .licenseignore ]; then
+            while IFS= read -r line; do
+              case "$line" in ""|\#*) continue ;; esac
+              set -- "$@" -ignore "$line"
+            done < .licenseignore
+          fi
+          found=0
+          while IFS= read -r f; do
+            [ -n "$f" ] || continue
+            [ -f "$f" ] || continue
+            found=1
+            /tmp/addlicense "$@" "$f"
+          done < /tmp/added.txt
+          [ "$found" = "1" ] || echo "No newly added files."
+
           if git diff --quiet; then
             echo "no_changes=true" >> "$GITHUB_ENV"
-            echo "commit_sha=" >> "$GITHUB_ENV"
+          else
-            echo "No license header changes."
+            echo "no_changes=false" >> "$GITHUB_ENV"
-            exit 0
           fi
-          git config user.name "Dcntrlized Auto CopyLeft Bot"
-          git config user.email "[email protected]"
-          git remote set-url origin "https://oauth2:${TOKEN}@${SERVER}/${{ gitea.repository }}.git"
-          git add -A
-          git commit -m "chore: add AGPL-3.0 license headers"
-          SHA="$(git rev-parse HEAD)"
-          git push origin HEAD:refs/heads/$REF
-          echo "no_changes=false" >> "$GITHUB_ENV"
-          echo "commit_sha=$SHA" >> "$GITHUB_ENV"
 
-      - name: Comment on pull request
+      - name: Amend the commit and force-push
-        if: env.no_changes == 'false'
+        if: env.fork_pr != 'true' && env.no_changes == 'false'
+        env:
+          TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: |
+          set -e
+          REF="${{ gitea.event.pull_request.head.ref }}"
+          SERVER="${{ gitea.server_url }}"
+          SERVER="${SERVER%/}"
+          SCHEME="${SERVER%%://*}"
+          HOST="${SERVER#*://}"
+          git config user.name "Dcntrlized Auto CopyLeft Bot"
+          git config user.email "[email protected]"
+          git add -A
+          git commit --amend --no-edit
+          SHA="$(git rev-parse HEAD)"
+          printf '%s\n' "$SHA" > /tmp/commit_sha.txt
+          echo "amended to $SHA"
+          BEFORE="$(cat /tmp/head_before.txt)"
+          git push --force-with-lease="refs/heads/${REF}:${BEFORE}" \
+            "${SCHEME}://oauth2:${TOKEN}@${HOST}/${{ gitea.repository }}.git" \
+            "HEAD:refs/heads/${REF}"
+
+      - name: Comment on the pull request
+        if: env.fork_pr != 'true' && env.no_changes == 'false'
         env:
           TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          COMMIT_SHA: ${{ env.commit_sha }}
-          COMMIT_URL: "${{ gitea.server_url }}/${{ gitea.repository }}/commit/${{ env.commit_sha }}"
         run: |
           set -e
           SERVER="${{ gitea.server_url }}"
+          SERVER="${SERVER%/}"
           REPO="${{ gitea.repository }}"
-          if [ "${{ gitea.event_name }}" = "pull_request" ]; then
+          PR="${{ gitea.event.pull_request.number }}"
-            PR_NUMBER="${{ gitea.event.pull_request.number }}"
+          SHA="$(cat /tmp/commit_sha.txt)"
-          else
+          cat > /tmp/pr-comment.md <<EOF
-            OWNER="${REPO%%/*}"
-            BRANCH="${{ gitea.ref_name }}"
-            PR_NUMBER="$(curl -sS -H "Authorization: token ${TOKEN}" \
-              "${SERVER}/api/v1/repos/${REPO}/pulls?state=open&limit=50" \
-              | jq -r --arg b "$BRANCH" 'map(select(.head.ref == $b))[0].number // empty')"
-            if [ -z "$PR_NUMBER" ]; then
-              echo "No open pull request for branch ${BRANCH}; skipping comment."
-              exit 0
-            fi
-          fi
-          cat > /tmp/pr-comment.md <<'EOF'
           ## Uh oh! Unlicensed code detected!
 
           You just tried to push **unlicensed code**! This repository is covered by the
           [AGPL-3.0](https://www.gnu.org/licenses/agpl-3.0.html) license, and all code
           requires a license header.
 
-          Thankfully, our helpful bot just created a commit
+          Thankfully, the bot caught it and amended your commit
-          ([`COMMIT_SHA`](COMMIT_URL)) with the correct licensing!
+          ([\`${SHA}\`](${SERVER}/${REPO}/commit/${SHA})) with the correct licensing.
 
-          If you think this is incorrect, tag **@Admin** so we can take a look!
+          If this looks wrong, tag **@Admin** so we can take a look!
           EOF
-          sed -i "s|COMMIT_SHA|${COMMIT_SHA}|g; s|COMMIT_URL|${COMMIT_URL}|g" /tmp/pr-comment.md
           BODY="$(jq -n --rawfile b /tmp/pr-comment.md '{body:$b}')"
           curl -sS -X POST \
-            "${SERVER}/api/v1/repos/${REPO}/issues/${PR_NUMBER}/comments" \
+            "${SERVER}/api/v1/repos/${REPO}/issues/${PR}/comments" \
             -H "Authorization: token ${TOKEN}" \
             -H "Content-Type: application/json" \
             -d "$BODY"
+          echo "comment posted"