Joel Maher ( :jmaher ) (UTC -8) Assignee
	Description • 11 years ago

a fallout from 923770 where the endtimestamp is in python time.time() format and not in javscript time format.

	Joel Maher ( :jmaher ) (UTC -8) Assignee
	Comment 2 • 11 years ago

Comment on attachment 821591 [details] [diff] [review]
time.time() * 1000 (1.0)

unfortunately I see a 500ms regression on ts_paint with this patch, but I don't know how to make the timestamp prior to launching any closer to the actual lauching with mozprocess.

nevertheless, this fixes a regression in tp5o0 - shutdown.

	William Lachance (:wlach)
	Comment 3 • 11 years ago

Comment on attachment 821591 [details] [diff] [review]
time.time() * 1000 (1.0)

Review of attachment 821591 [details] [diff] [review]:
-----------------------------------------------------------------

Without having investigated the details of why we need the timestamps to be in milliseconds, this looks fine to me. Moving the calculation of the initial timestamp closer to the launching of the process also sounds good, though I worry there's some other problem there that we're not addressing here. Still, looks good to go in. Can always sort out other problems or the real issue some other way.

::: talos/ttest.py
@@ +370,4 @@
>                          setup.run()
>                          setup.wait()
>  
> +                    env = os.environ.copy()

I don't see why this deserves its own line/variable :)

	Joel Maher ( :jmaher ) (UTC -8) Assignee
	Comment 4 • 11 years ago

pushed and moved the env into the arg list instead of a new line:
https://hg.mozilla.org/build/talos/rev/cc936fa80fd8