Files
t 8d5d8af297
CI / check (push) Skipped
CI / e2e (push) Skipped
ci(release): keep the changelog out of a protected branch
main is protected (enable_push: false, empty whitelist), so
@semantic-release/git's commit-back is rejected by the pre-receive
hook -- and it would be rejected *after* the tag was pushed, leaving a
tagged release the run then reports as failed. Found by trying to push
this branch to main.

Whitelisting the CI user was the alternative and is declined: it
weakens a protection someone set deliberately and lets a bot push to
main without the checks every human PR has to pass.

So the release page is the changelog. The changelog plugin now writes a
gitignored .release-notes.md, which exists only to carry the notes into
gitea-release.sh without interpolating them into a shell command, and
CHANGELOG.md is a signpost -- a file claiming to be a changelog while
silently never updating is worse than no file.

Tags are not protected, so the tag push is unaffected.
2026-08-17 19:46:48 -04:00

97 lines
3.4 KiB
Bash
Executable File

#!/usr/bin/env bash
#
# Create the Gitea release for a version semantic-release has just tagged.
#
# This is `@semantic-release/exec`'s publishCmd, and it exists because
# Gitea's API is /api/v1 and @semantic-release/github speaks GitHub's.
# That is the whole of the Gitea-shaped work: one POST.
#
# **The notes come from a file, not from an argument.** Release notes are
# rendered commit messages — arbitrary text carrying backticks, quotes and
# `$` — so interpolating ${nextRelease.notes} into a shell command would
# be an injection whose input is the commit log. @semantic-release/changelog
# has already written them to .release-notes.md by the time `publish` runs,
# so the only thing crossing the shell boundary here is a semver string,
# which is validated below anyway.
#
# That file is a gitignored build artifact, not a document: `main` is a
# protected branch, so nothing commits a changelog back to it and the
# release page is the changelog. See .releaserc.yml.
#
# Usage: scripts/gitea-release.sh <version> # e.g. 0.0.1
#
# Environment (all set by .gitea/workflows/release.yml):
# SERVER_URL https://git.ljones.me
# OWNER yonlu
# REPO yonlu/yellowjacket
# PACKAGE_TOKEN a user PAT with write access
set -euo pipefail
cd "$(dirname "$0")/.."
version="${1:?usage: gitea-release.sh <version>}"
# Validated rather than trusted: this is the one value that reaches a URL
# and a JSON document, and semantic-release is not the only thing that
# could ever call this.
if ! printf '%s' "$version" | grep -qE '^[0-9]+\.[0-9]+\.[0-9]+(-[0-9A-Za-z.-]+)?$'; then
echo "gitea-release: '$version' is not a semver version" >&2
exit 1
fi
: "${SERVER_URL:?SERVER_URL is not set}"
: "${REPO:?REPO is not set}"
: "${PACKAGE_TOKEN:?PACKAGE_TOKEN is not set}"
tag="v${version}"
# The top section of the changelog is this release's notes: everything
# from the first `## ` heading to the one after it. awk rather than sed
# so the "there is no second heading" case (the first release) needs no
# special handling.
notes=$(awk '
/^## / { seen++; if (seen > 1) exit }
seen { print }
' .release-notes.md)
if [ -z "$notes" ]; then
echo "gitea-release: found no release section at the top of .release-notes.md" >&2
echo ' the changelog plugin runs in prepare and this runs in publish, so' >&2
echo ' an empty section means the plugin order in .releaserc.yml moved.' >&2
exit 1
fi
echo "gitea-release: creating $tag from $(printf '%s' "$notes" | wc -l) lines of notes"
# jq builds the body, so a backtick or a quote in a commit subject is data
# rather than syntax.
payload=$(jq -n \
--arg tag "$tag" \
--arg name "$tag" \
--arg body "$notes" \
'{tag_name: $tag, name: $name, body: $body, draft: false, prerelease: false}')
code=$(curl -sS -o /tmp/gitea-release.out -w '%{http_code}' \
-X POST \
-H "Authorization: token ${PACKAGE_TOKEN}" \
-H "Content-Type: application/json" \
-d "$payload" \
"${SERVER_URL}/api/v1/repos/${REPO}/releases")
case "$code" in
201)
echo "gitea-release: created ${SERVER_URL}/${REPO}/releases/tag/${tag}"
;;
409)
# Already there. The correct outcome for a re-run of the same tag,
# and not a failure — the publish workflows are idempotent for the
# same reason.
echo "gitea-release: $tag already has a release; leaving it alone"
;;
*)
echo "gitea-release: POST /releases returned $code" >&2
cat /tmp/gitea-release.out >&2
exit 1
;;
esac