914713 - Default timeout for wait_for_port is 3000s (50 minutes)

Status: RESOLVED FIXED

(Remote Protocol :: Marionette, defect)

Description

[James Graham [:jgraham]]

The default timeout of 3000s seems excessively long. Possibly it is supposed to be 3s? In that case it would make sense to retry more than once per second.

(It is also not clear why there is a 5s wait after the port has been opened).

Comment 1

[Malini Das [:mdas] - Away, not checking bugmail]

I think 30seconds is perfect for wait for port, since it accommodates the case where we restart b2g.

As a first bug:
To make this change, you'll need to look at (mozilla-central source_testing/marionette/client/marionette/marionette.py and modify the default value for the wait_for_port function and how it uses the default value (ie: what it passes to the timedelta function)

Comment 2

[Soumya Deb [:Debloper]]

Attachment: v1.patch

Comment 3

[Malini Das [:mdas] - Away, not checking bugmail]

Comment on attachment 802621 [details] [diff] [review]
v1.patch

Review of attachment 802621 [details] [diff] [review]:
-----------------------------------------------------------------

Great, thanks! I've tested this locally against the unit test suite and it works as expected. Will land in m-i/m-c shortly.

Comment 4

[Malini Das [:mdas] - Away, not checking bugmail]

I've pushed it to m-i:
https://hg.mozilla.org/integration/mozilla-inbound/rev/e0da7141fa04


by the way, Debloper, I had to add your username to the patch so you'd be recognized as the one who wrote the change. You should add your username to your hg configuration so your username will show up in future patches automatically: https://developer.mozilla.org/en-US/docs/Installing_Mercurial#Basic_configuration

Comment 5

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/e0da7141fa04

Comment 6

[Soumya Deb [:Debloper]]

(In reply to Malini Das [:mdas] from comment #4)
> I've pushed it to m-i:
> https://hg.mozilla.org/integration/mozilla-inbound/rev/e0da7141fa04
Awesome, thanks!

> by the way, Debloper, I had to add your username to the patch so you'd be
> recognized as the one who wrote the change.
/me checks the hgrc config... *seems alright*
/me checks the patch again... *no author info*
/me isn't sure what went wrong, but sure as hell it's one such case!
Thanks, bytheway!