[webkit-reviews] review requested: [Bug 27256] Geolocation timeout parameter is not correctly applied : [Attachment 38373] Patch 2 for bug 27256

[bugzilla-daemon at webkit.org]

steveblock at google.com has asked  for review:
Bug 27256: Geolocation timeout parameter is not correctly applied
https://bugs.webkit.org/show_bug.cgi?id=27256

Attachment 38373: Patch 2 for bug 27256
https://bugs.webkit.org/attachment.cgi?id=38373&action=review

------- Additional Comments from steveblock at google.com
> Why are tests impossible?  (I'm not suggesting they aren't, I just need more
> info as to 'why?')
Writing test for this (and all of the other Geolocation features) requires a
mock Geolocation service so that the behavior of the API is predictable for
testing. This is being tracked in Bug 28264 (which also needs review!).

> Is it possible that the handlers could do nasty things?
> like start a new request or something?  Anything that could potentially crash

> us when we return back to executing c++ after the JS callback?
Yes, a handler could start a new request, but this shouldn't be a problem. I'll
add tests for this once we have a mock Geolocation service.

> We only name arguments when they add clarity:
Fixed.

> Are
> there other webkit reviewers who know something about geolocation?
Not that I knew of. Greg Bolsinga wrote the original code, but he's not a
reviewer.

> Who uses
> this code?  Andriod?	Does iPhone use this or do they have their own thing?
Yes, this is used by Android, iPhone and Chrome.