-
Notifications
You must be signed in to change notification settings - Fork 334
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Create a GitHub release for each action release #2517
Conversation
Must make sure this release is not marked as `latest` or else it will interfere with the CLI bundle releases also included in this repo.
"${VERSION}" \ | ||
--latest=false \ | ||
-t "${VERSION}" \ | ||
-F CHANGELOG.md |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Will this include just the release notes for the latest release?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This creates release notes for all releases. Looks like this: https://github.com/aeisenberg/codeql-action/releases/tag/v1.0.0
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think it would be great if we could pull out just the relevant section. We already have some code to do that in the .github/update-release-branch.py
script which we could reuse.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Good idea. I'll what I can do. There should also be an easy link to the full changelog.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hi @aeisenberg, is this something you still plan to address, or shall we review this PR as is?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sorry...this slipped through. Let me take a look at this before you review.
Here is a release created using the workflow: https://github.com/aeisenberg/codeql-action/releases Here is the run that created it: https://github.com/aeisenberg/codeql-action/actions/runs/11267116373/job/31331695828 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM broadly, 2 minor comments.
"$VERSION" \ | ||
--latest=false \ | ||
-t "$VERSION" \ | ||
-F "$PARTIAL_CHANGELOG" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
⛏️ Would it make sense to change the -F
to the more explicit --notes-file
?
Reveals the intent better, and if it's part of a script it might be more convenient for future readers of the code.
if line.startswith('## '): | ||
if found_first_section: | ||
break | ||
found_first_section = True |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The way this works is by appending lines until the a second H2
markdown header is found, as I understand it.
In the test run, the release notes it created contain the header [UNRELEASED]
. I think that these headers are processed before the release, but does it make sense to have any sort of post-processing/manipulation of the header in case we come across the [UNRELEASED]
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Hmmm...I'd originally thought that this was because I didn't start on a release branch, but now I'm not so sure. I'll have to take a deeper look.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
OK...there were two things going on:
- The run was not on a release branch, so we don't see the actual version number. This is expected.
- However, I originally created the changelog too late in the process after the header was changed back to
UNRELEASED
.
I've now fixed 2. I also added a few cleanups.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM.
run: | | ||
python .github/workflows/script/prepare_changelog.py CHANGELOG.md "$VERSION" > $PARTIAL_CHANGELOG | ||
|
||
echo "::group::Partial CHANGELOG" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
👍🏻 I like this change! Makes it easier to see what the changelog looked like at the time of generation!
@@ -108,6 +108,17 @@ jobs: | |||
# - `--force` since we're overwriting the `vN` tag | |||
git push origin --atomic --force refs/tags/"${VERSION}" refs/tags/"${major_version_tag}" | |||
|
|||
- name: Prepare partial Changelog | |||
env: | |||
PARTIAL_CHANGELOG: "${{ runner.temp }}/partial_changelog.md" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just one nitpick, can be merged as is:
I think we can pull this at a higher level in the workflow file, and then the PARTIAL_CHANGELOG
variable will be shared between the different steps. The benefit this brings is that if at any point we need to change it, it will be changed at both of the steps that require the same file, instead of having to change two separate definitions of the variable, minimising the chance of the two variables being out of sync and causing a bug by accident.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I tried that earlier and unfortunately, and runner.temp
is not initialized at the job level. I thought this way was safer than trying to hard code the temp directory.
Must make sure this release is not marked as
latest
or else it will interfere with the CLI bundle releases also included in this repo.See https://cli.github.com/manual/gh_release_create
Merge / deployment checklist