w3 :: #webdriver

16 Mar 2017
12:21MikeSmithAutomatedTester: simonstewart I need to respond to plh about What would qualify as enough adequate implementation experience?
12:21simonstewartI'm not even sure what the question means
12:21simonstewartWe have multiple implementations out there already, slowly creeping towards being done
12:22simonstewartThe lead two Mozilla's geckodriver and Selenium's IEDriver
12:22simonstewartBut there are folks with "implementation experience" at MS, Google, and Apple.
12:22AutomatedTesterMozilla have 2 independent implementations in the works
12:22AutomatedTesterwith Gecko and Servo
12:22AutomatedTesterunless jgraham tells me otherwise ;)
12:22MikeSmithright, hes asking for something more like, This Candidate Recommendation will have met its exit criteria when there are at least two interoperable implementations of each feature in the spec
12:23MikeSmithso can we say that?
12:23simonstewartI think so
12:23MikeSmithOK
12:24simonstewartjimevans has a draft of Actions working in the IEDriver, and it's in geckodriver already
12:24simonstewartThe New Session stuff is in progress at Mozilla and on Jim's list
12:24simonstewartThose are the main areas where we've diverged from the original wire protocol
12:24AutomatedTesterGeckoDriver is mostly correctness work going on
12:25simonstewartSo it feels like we're close to "interoperable" already.
12:25simonstewartWhether they're also conformant is another matter entirely :)
12:27simonstewartLGTM
12:32jgrahamsimonstewart: So let's be clear that "two" here doesn't mean "one remote end and one local end"
12:32jgrahamIt means "teo remote ends"
12:32jgraham*two
12:33jgrahamOr that's how i would expect plh to interpret it
12:33jgrahamSo it's more like "geckodriver and chromedriver"
12:33jgrahamOr whichever other driver we expect to conform first
12:34AutomatedTesterjgraham: thats what simonstewart was talking about with IEDriver
12:34AutomatedTesterthats the Selenium remote end of IE
12:34jgrahamAh IEDriver
12:34jgrahamOK, for some reason I thought he meant jimevan's client
12:34MikeSmithso I also need to get back to Ralph about the open issues
12:34jgraham*jimevans'
12:35MikeSmithwhat do we do with an issue like https://github.com/w3c/webdriver/issues/864 as far as CR?
12:35MikeSmiththat looks like a normative change
12:36MikeSmithas does https://github.com/w3c/webdriver/issues/863
12:36AutomatedTesterMikeSmith: thats underspecified
12:36AutomatedTesterIts not a change to how things would work
12:37AutomatedTesterwe have specified how we serialise out of the browser
12:37AutomatedTesterfor #863
12:37AutomatedTesterMikeSmith: and #864 is an edge case that was missed
12:37jgrahamYeah 863 is just a bug
12:37MikeSmithOK
12:38jgrahamI think 862 is a minor normative change
12:38AutomatedTesterMikeSmith: we are not planning on changing how things work, we are going to be adding clarifications or bug fixes
12:38MikeSmithso maybe I should create an underspecified label?
12:38MikeSmithI just made a clarification label
12:38MikeSmithand used it for https://github.com/w3c/webdriver/issues/844
12:38AutomatedTesterMikeSmith: I can do that and then work through the bugs for you
12:39MikeSmithAutomatedTester: that would be great
12:39MikeSmithfeel free to rename or delete clarification
12:39AutomatedTesterI am going to make lunch in a sec
12:39AutomatedTester
12:39MikeSmithOK
12:39AutomatedTesterand then will get on it
12:39MikeSmiththanks
12:41MikeSmithI guess underspecified and clarification may mean the same thing but perhaps underspecified seems better in that we can use it to make normative changes while still asserting that they are not changing behavior but instead just more precisely specifying it
13:26simonstewart+1
14:36AutomatedTesterMikeSmith: sorry, havent had chance to do it yet
14:37AutomatedTestermy day has run away from me
14:37MikeSmithno worries
19:29jimevansso if a remote end fails to merge capabilities during a new session call, what happens?
19:29jgrahamFails in what way?
19:30jgrahamBut presumably it returns SessionNotCreated
19:30jgrahamAlthough I haven't checked the spec
19:30jimevansattempts to merge a property from secondary that already exists in primary. the merge capabilities algorithm says it should fail.
19:31jimevanshold on.
19:31jimevansnever mind.
19:31jimevansi overlooked the phrase &quot;trying to call <algorithm>&quot;
19:34jgrahamAh
19:34jgrahamYeah it&#39;s a bit subtle, but cuts out a lot of boilerplate
19:34jgrahamMaybe tomorrow i will actually have time to write some tests!
22:55* MikeSmith checks the state of labels in the issue tracker
22:56AutomatedTesterMikeSmith: sorry...
22:56MikeSmithAutomatedTester: didnt expect you to still be around at this hour!
22:58MikeSmithAutomatedTester: well for I can reply to Ralph with a general statement that those issues are all (almost all?) issues where the spec is just underspecified, and we will get them explicitly labeled as such soon, but we would like to not block the transition to CR on that getting done
22:58AutomatedTesterMikeSmith: sounds great!
23:24MikeSmithOK message sent https://lists.w3.org/Archives/Public/www-archive/2017Mar/0005.html
23:25MikeSmith> After examining those issues, the determination is that they fall into
23:25MikeSmithcategories of either just editorial clarifications or else cases where a
23:25MikeSmithpart of the spec is underspecified to the point that while some additional
23:25MikeSmithnormative text may be needed, its addition will not change the
23:25MikeSmithimplementation requirements but instead will, e.g., remove any ambiguities.
23:25MikeSmith> we think that none of them are anything that should block
23:25MikeSmiththe transition to CR, so wed like to go ahead with the transition, with
23:25MikeSmiththat understanding.
23:28AutomatedTestergreat
23:28AutomatedTesterI have labelled them
23:29AutomatedTesterI will go through them tomorrow to check
23:29AutomatedTesterbug == mistake in spec
23:29AutomatedTesterclarification == underspecified or needs more prose to explain
17 Mar 2017
No messages
   
Last message: 155 days and 17 hours ago