Skip to content
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

More accurate release notes for backport releases #181

Open
christopher-dG opened this issue Nov 18, 2020 · 4 comments
Open

More accurate release notes for backport releases #181

christopher-dG opened this issue Nov 18, 2020 · 4 comments
Labels
rewrite Will be done in the Julia rewrite of TagBot (see #55), but not before.

Comments

@christopher-dG
Copy link
Member

christopher-dG commented Nov 18, 2020

Suppose I make the following releases, in order: 1.0.0, 2.0.0, 1.0.1

  • Issues/PRs included in release notes for 1.0.0 are from beginning of time -> 1.0.0.
  • For 2.0.0, they're from 1.0.0 to 2.0.0
  • For 1.0.1, they're from 1.0.0 to 1.0.1 because 1.0.0 is the previous version according to SemVer, not 2.0.0. But usually, most of the issues/PRs included in the release notes are actually to do version 2.0.0, and thus not relevant to 1.0.1. The generator collects stuff solely based on close/merge time.

An idea: Most backport releases are not made on the main branch, and we can use that to help prune the issues/PRs.

  • Get the the commit being registered from the registry PR.
  • Figure out what branch that commit is on. For normal releases, this is basically always the default branch, but for backports, often it's release-A.B or similar and not on the default branch.
  • Ignore all pull requests that were not merged into this branch.

Pruning issues is harder, since they don't have a base branch of their own. But they're often associated with (closed by) PRs, and we could select only the issues that were closed by a PR that was merged into the release branch. But errors in that approach result in lost information rather than extra information, and it's a lot easier to remove stuff manually than to add it in manually. And ultimately, that approach gets the same or less as just taking all issues from the previous chronological release.

So maybe the best bet is to take only PRs that were merged into the release branch since the previous SemVer release, and take all issues that were closed since the last chronological release.

@c42f
Copy link

c42f commented Nov 27, 2020

Great summary. I've hit this a couple of times in the opposite order:

1.0.0 -> 1.0.1 -> 2.0.0

where 1.0.1 is a tiny release made on a branch, but gets attributed most of what 2.0.0 should include. Then the 2.0.0 release notes include hardly anything.

@maleadt
Copy link

maleadt commented Mar 26, 2021

Is there a workaround for this? On multiple occasions (e.g. see above) I've had users confused about a feature that Tagbot had put in the release notes of a backport release while it's only available on master, not even in any release.

@christopher-dG christopher-dG added the rewrite Will be done in the Julia rewrite of TagBot (see #55), but not before. label Apr 22, 2021
@oxinabox
Copy link
Contributor

oxinabox commented Aug 1, 2021

It would be nice if in doubt tagbot would put nothing user viable.
(Could generate the full text and hide in a markdown comment so if package mainainer wants to use some of it to write the release notes they can edit it in)

I think detecting the in doubt cases is probably easier?

@fonsp
Copy link
Contributor

fonsp commented Nov 23, 2021

This will still mean that issues and PRs get 'swallowed' by a backport release (1.0.1 in the example) and won't show up in the release that actually contains these fixes, right?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
rewrite Will be done in the Julia rewrite of TagBot (see #55), but not before.
Projects
None yet
Development

No branches or pull requests

5 participants