< wumpus>
so many windows PRs lately, that's great, but we could really use a windows reviewer
< MarcoFalke>
ryanofsky: It is noise. Also, you can check the edit history to see that nothing changed.
< ryanofsky>
can draftbot be prevented from showing up in edit history when it is not actually changing anything? (it seems impossible to tell in 11640 whether or not it made changes due to incomplete history)
< ryanofsky>
also imo it would be nice if draftbot refrained from adding "needs rebase" comments, and just stuck to applying labels and maybe pinging pr authors privately
< ryanofsky>
i'd like to be able to keep my rebase history in a comment that get updated instead of having this "needs rebase" "rebased" "needs rebase" "rebased" ping pong
< MarcoFalke>
ryanofsky: The "needs rebase" comments get deleted after you did the rebase
< MarcoFalke>
Also no need to reply to the bot
< MarcoFalke>
Though, I can change it to ping privately if you have an idea on how to do that
< harding>
MarcoFalke: to ping privately, the bot could just pull the email address of the committer from the most recent commit message and email them. Alternatively, the bot could open an issue on the repository they're making their PR from, e.g. github.com/harding/bitcoin/issues/new
< MarcoFalke>
The email wouldn't go through the GitHub filter, so likely couldn't be sorted into different inboxes. And issues are usually disabled for forked repos
< harding>
What's the GitHub filter? You're correct on the second point; I think that didn't used to be the case.
< harding>
Oh, I see, you're saying that people who have rules setup already wouldn't be able to reuse those rules for the rebase notifications.
< MarcoFalke>
jup
< MarcoFalke>
githubs send from:"notifications@github.com" to:"bitcoin@noreply.github.com" with cc:"comment@noreply.github.com" and cc you
< harding>
MarcoFalke: maybe just batch all rebase notifications for a particular email address into one daily notification. That's not too spamy, I think? A single list of todos is kind of nice anyway.
< MarcoFalke>
Hmm, I could just disable it for pull requests that are highly likely to conflict
< MarcoFalke>
Which I believe are only #10973 and #10102