Stream: t-compiler/wg-rls-2.0

Topic: bors on rust-analyzer


pnkfelix (Feb 12 2019 at 13:11, on Zulip):

Hey @matklad when I skimmed over other PR's to infer what the review-protocol was, I noticed a bunch have bors commenting "canceled"

pnkfelix (Feb 12 2019 at 13:11, on Zulip):

and they also, I think, all happened to have a :robot: in the comment below where you had written "bors r+" (example here)

pnkfelix (Feb 12 2019 at 13:12, on Zulip):

so my question is: is that just a coincidence? Or are you adding in the robot as a way to cause the cancellation? Or is bors adding it in when it cancels the r+ of a PR?

pnkfelix (Feb 12 2019 at 13:17, on Zulip):

and more generally, what is the bors "cancelled" signalling there? is the CI hitting some issue and forcing a retry, and thus in general it is not (necessarily) a sign that one might need to adjust one's PR in some manner? The "Canceled" comment doesn't include a link so its hard for me to tell where I would go to learn more

matklad (Feb 12 2019 at 13:19, on Zulip):

The :robot: thing is my local script which sends a PR and r+ it immediately. I want even the smallest changes to go via bors, but I don't want to open the browser for each small commit.

pnkfelix (Feb 12 2019 at 13:20, on Zulip):

Gotcha

matklad (Feb 12 2019 at 13:20, on Zulip):

As it happens, I am pretty bad with making good commits from the first try so often after I've done the :robot: thing I notice that something is wrong and go with git commit --amend && git push --force-with-lease

matklad (Feb 12 2019 at 13:21, on Zulip):

That causes bors to cancel the current build,

pnkfelix (Feb 12 2019 at 13:22, on Zulip):

ah, so the cancellation is in response to the git push --force (or at least --force-with-lease)

pnkfelix (Feb 12 2019 at 13:22, on Zulip):

gotcha

Last update: Nov 19 2019 at 17:40UTC