Can we jettison Sputnik?

# Ross.Kirsling at sony.com (2 days ago)

I’d like to verify whether there’s any specific reason that the LayoutTests/sputnik directory continues to exist.

According to Wikipedia, “All current Sputnik tests have been incorporated into ECMA's Test262 test suite.” (Here “current” evidently means June 2011.) And indeed, when making fixes to address Test262 failures in JSC, it’s quite awkward to have to update the Sputnik expectations to reflect any new “failures”—particularly given that it’s not sufficient to simply build JSC in order to run them.

Would anyone object to a patch deleting this directory outright? The only concern that occurs to me is that Test262 is not currently integrated into EWS.

Thanks! Ross

Contact us to advertise here
# Saam barati (2 days ago)

On Jan 15, 2019, at 11:50 AM, Ross.Kirsling at sony.com wrote:

Hi everybody,

I’d like to verify whether there’s any specific reason that the LayoutTests/sputnik directory continues to exist.

According to Wikipedia, “All current Sputnik tests have been incorporated into ECMA's Test262 test suite.” (Here “current” evidently means June 2011.) And indeed, when making fixes to address Test262 failures in JSC, it’s quite awkward to have to update the Sputnik expectations to reflect any new “failures”—particularly given that it’s not sufficient to simply build JSC in order to run them.

I think it’s a good idea to remove these. (An aside: in general, I’m also not a fan of JS tests that don’t rely on the DOM existing in LayoutTests.)

Would anyone object to a patch deleting this directory outright? The only concern that occurs to me is that Test262 is not currently integrated into EWS.

Maybe we should integrate these into EWS? The main downside is IIRC, these take a while to run. I think it’s ok to remove these even if we never add Test262 to EWS since we still run Test262 on bots post-commit.

Want more features?

Request early access to our private beta of readable email premium.