Steve Fink [:sfink] [:s:] Assignee
	Description • 12 years ago

This may not be necessary after bug 836980, but there are many ways to generate try pushes, so it might be worth having this as a backstop.

	Steve Fink [:sfink] [:s:] Assignee
	Comment 1 • 12 years ago

And it has nothing to do with me wanting to land something with a try try again pun.

We could also make these pushes fail. Personally, I do quite a few pushes with things like |-p quantum_interference_patterns_of_supercooled_fluorine| (usually because I need pushes for testing releng changes), which has the same effect as -p none.

	Ed Morley [:emorley]
	Comment 2 • 12 years ago

Comment on attachment 709258 [details] [diff] [review]
Detect -p none

Sorry for the delay, dealing with the backlog after not one but two SSD failures (yey \o/).

Looks good - only change I would make is for this to fail the push (since as you say for testing, you can just enter junk instead of 'none'), otherwise my bet is on someone filing another bug in a few months time with "I pushed to try and ignored the response, and thought it would be successful since return code 0".

	Steve Fink [:sfink] [:s:] Assignee
	Comment 3 • 12 years ago

Pushed with the suggested change (making it fail instead of just warn)

http://hg.mozilla.org/hgcustom/hghooks/rev/3b0c66182bb0

	Phil Ringnalda (:philor)
	Comment 4 • 11 years ago

Don't forget to file the dependent IT bug to have that actually deployed.

	Ed Morley [:emorley]
	Comment 5 • 11 years ago

(This is an hg hook customisation; the trychooser counterpart landed in bug 836980. Will still need deploying in production.)

	Gregory Szorc [:gps]
	Comment 6 • 9 years ago

This appears to have been fixed years ago but nobody closed the bug.