w3 :: #testing

17 Mar 2017
08:52annevkSo what's the plan with documentation?
08:52annevkIs there a tracking bug?
09:06annevkjgraham: I can't ask lyzadanger for review using the GitHub UI
09:12MikeSmithannevk: no bug that Im aware of
09:12annevkI&#39;d really like to get some things fixed since currently it&#39;s a disaster with the <iframe> and finding things is still hard
09:12annevkE.g., I&#39;m having a hard time finding the .sub. documentation
09:13MikeSmithI guess we need to talk to gsnedders
09:13annevkAnd wptserve being separate doesn&#39;t help either
09:13MikeSmithyeah agreed about that
09:13MikeSmiththough those are good docs
09:14annevkYeah, could do with some more examples maybe, but they&#39;re pretty good once you know where to find them
09:29jgrahamannevk: Need to add her to the collaborators list :/
09:30jgraham(happy to do that ofc, just a silly requirement)
09:30jgrahamannevk: I&#39;m hoping the docs situation will improve when we put everything in the same repo
09:30annevkjgraham: and that happens post-css?
09:31annevkjgraham: note that I&#39;m willing to invest time if that can help
09:32annevkNow we finally realized testing is important for standards to succeed, I&#39;d very much like to do my part
09:34jgrahamyeah, as soon as CSS merges I think we can monorepo all the things. Need a plan for open PRs and such, but that will amount to close out as many as possible and do manual work to transfer the others if necessary
09:39annevkOkay, I guess I can wait a couple weeks and do other things meanwhile
09:39annevkProbably more productive to wait for that to happen
10:13jgrahamannevk: Yeah, that&#39;s what I&#39;m thinking
10:17jgrahamMikeSmith: Thanks for fixing the mirroring
10:19jgrahamMikeSmith: Maybe we should write a cron job that checks everything in submissions/ is still an open PR
10:19jgrahamand deletes it if not
10:20MikeSmithneed to set that up
10:20MikeSmithmaybe this weekend
11:21gsneddersannevk: if you have ideas of how to improve navigatability for reference without making it hard for those who are new feel free to file an issue
11:22annevkgsnedders: is this about docs? I filed a bunch of issues, I was getting impatient
11:30gsneddersannevk: yeah; most of this is down to priorities, I can&#39;t just work on docs indefinitely :)
11:32annevkThat&#39;s understood, but I&#39;m also not entirely clear that the half-way state we have now is better, but I&#39;ll wait a few weeks until everything is merged and it&#39;s easier to make improvements
11:34gsneddersdoes the iframe actually cause you actual problems?
11:34gsneddersit&#39;s ugly as sin, no denying it
11:35gsneddersbut how problematic is it?
11:36annevkgsnedders: searching doesn&#39;t work well, scrolling is terrible, it&#39;s super tiny
11:37annevkgsnedders: I basically get stuck whenever I get to that page
11:37annevkgsnedders: I actively avoid looking at documentation at the moment and copy-and-paste from elsewhere
11:38annevkgsnedders: e.g., I didn&#39;t learn about .any.js from the documentation
11:38annevkI just happened to stumble across old PRs / commits while doing triage that had it
11:43gsneddersannevk: searching works fine for me in every browser I have, it&#39;s the same size as every other bit of content on the site
11:45annevkgsnedders: good for you I guess, it&#39;s definitely cut off at the bottom here and the iframe has horizontal scrolling as well as vertical
11:46gsnedderswhat size in your window and what browser?
11:46annevkgsnedders: I don&#39;t know, bit smaller than 27&#39;&#39;
11:46annevkgsnedders: Nightly.app
11:47gsnedders(the iframe will basically always have horzinontal scrolling, it has some lines something ridiculously wide in <pre>)
11:47annevkgsnedders: anyway, I don&#39;t know why we&#39;re discussing this if we can fix it in a couple weeks
11:48gsneddersjgraham: I can&#39;t imagine moving lots of 24 year old PRs is going to be very easy
11:48* gsnedders wonders why there are so many
11:52jgrahamgsnedders: We can move them on an ad-hoc basis
11:52jgrahamThey won&#39;t magically disappear from the old location
11:53jgrahamSome of them can probably be closed
13:23lyzadangerannevk: FYI Im reviewing https://github.com/w3c/web-platform-tests/pull/2356 right now and checking the behavior of the tests out I think I may have some input as I ran into something with the way `window.opener` is treated in these contexts, as well.
13:39jgrahamannevk: If lyzadanger wants to be a member of the wpt team (the usual way to get push access/ability to be a reviewer) MikeSmith has to invite her into the org
13:39lyzadangerjgraham: Its fine if I dont have it for now, FWIW. Im happy with just commenting on this thread. But thanks!
13:40jgrahamlyzadanger: OK, if you&#39;re happy with @mentions that is easy :)
13:41* jgraham wishes github had a dashboard for PRs for which you are marked as a reviewer
13:43lyzadangerjkfdj df
13:43lyzadangerUm, yes, thank you crappy IRC client.
13:44* jgraham interprets that as &quot;the noise one makes when discovering a 20 year old feature of the web platform is still buggy&quot;
13:44annevkjgraham: https://github.com/pulls/review-requested?
13:44jgrahamannevk: Oooh
13:45jgrahamApparently it just isn&#39;t linked from anywhere obvious to me
13:45annevkjgraham: next to the search box there&#39;s &quot;Pull requests&quot; and then you need another click to get here
13:46lyzadangerannevk: This is a relevant section of the spec for that PR, correct? https://html.spec.whatwg.org/multipage/browsers.html#crossoriginproperties-(-o-)
13:46lyzadangerJust checking before I make a potentially idiotic comment
13:47jgrahamannevk: Thanks! Still holding out for something that looks like the Bugzilla ui with a giant red thing on each page :)
13:48annevklyzadanger: that&#39;s part of it, there&#39;s also various things around the navigate algorithm involved, and how the noreferrer algorithm interacts with them
13:48annevklyzadanger: I&#39;ll be back in an hour or so, gonna take a break
13:49lyzadangerannevk: Cool, thanks!
13:52lyzadanger(My comment is from the specs perspective without concerning myself with how its implemented as such)
15:15gsneddersMikeSmith: awake?
15:15MikeSmithgsnedders: yeah but need to step away for a bit
15:16MikeSmithwill be back around later
15:16MikeSmithwant me to ping you then?
15:16gsneddersMikeSmith: depends how much later :)
15:16gsneddersMikeSmith: but sure
15:21MikeSmithstill around 2 hours from now?
15:22MikeSmithI have a call from somewhere between 2:30pm US/East (3 hours from now) to 5pm, but will be free again after that ends
15:23MikeSmithbut 2 hours fro now I still have that window of at least 1 hour before Id need to be on that call
15:24MikeSmithanyway, will ping you when Im back
15:29gsneddersMikeSmith: that wfm
16:15gsneddersokay, so git log --follow doesn&#39;t handle the case where files were renamed in a merge commit
17:45smcgruerqq: Should we include prefixed css if necessary for some vendors? Working on position:sticky tests atm and Safari has at least some support but requires the prefixed -webkit-sticky version.
17:46gsnedderssmcgruer: no
17:46smcgruerAck, fair enough
17:47jgrahamIn that specific case definitely no
17:51annevkjgraham: folks suggesting to include ringmark into wpt: https://bugzilla.mozilla.org/show_bug.cgi?id=748248
17:51annevkjgraham: left a rather generic comment
17:54gsneddersannevk: that seems like it&#39;ll add lots of duplicated tests
17:54gsneddersand I&#39;m p. sure we don&#39;t want to get into perf testing
17:55jgrahamannevk: Well done for not replying with &quot;argh, run away&quot;, I guess
17:55jgrahamBut yeah, that seems like a sensible comment. Pity facebook didn&#39;t do it for us.
17:57jgraham(there&#39;s also a fail that the regression reported there didn&#39;t end up with a wpt test written, but more mochitest)
17:57jgraham(so much work to do :()
17:58gsneddersjgraham: what would it take to have a similar goal to Blink to have a given percentage of all new tests be in wpt?
17:58annevkNow that would be an interesting commit policy change
18:00jgrahamgsnedders: Not sure. I would like us to achieve that, and blink taking the lead may be helpful in proving the idea
18:00MikeSmithgsnedders: back now
18:25gsneddersjgraham: r? ^^
18:30jgrahamDammit, forgot that reviewable doesn&#39;t let travis complete
18:31gsneddersTravis completed! Just not AppVeyor :P
18:32gsneddersjeffcarp: just looked at the dashboard again, coming along nicely!
18:33jgrahamWhere is the data in that up to date?
18:34jgraham(it looks pretty great btw)
18:35jgraham(some kind of search would be nice, I hear Google are good at that? ;)
18:43jgrahamApparently the serviceworker results at least don&#39;t match our internal ones
18:45gsneddershttps://travis-ci.org/w3c/web-platform-tests/jobs/212230766#L383 stability checker seems to be dying repeatedly on csswg_test_merged
18:46gsneddersand not because of too many tests, AFAICT
18:46gsneddershttps://travis-ci.org/w3c/web-platform-tests/builds/212230763 is the run from the latest push
18:46gsneddersthe lint will fail because of https://github.com/w3c/csswg-test/pull/1239 having not landed
18:47gsneddersaside from that, I think that&#39;s it more or less ready to go
18:48jgrahamgsnedders: Is fe5901f6c401fafd84736082c6176b373503cd55 a revision in the tree?
18:49* jgraham checks
18:49jgrahamWell /me checks after pulling 170000 objects
18:49gsneddersit shouldn&#39;t have to repull that many?
18:50gsneddersoh, actually maybe this time is does
18:50jgrahamWell it is :)
18:51gsneddersjgraham: it&#39;s one of the new root commits
18:51gsneddershence it is true: fe5901f6c401fafd84736082c6176b373503cd55^ isn&#39;t in the tree because it has no parent
18:54jgrahamgsnedders: Ah, well then my assumption that all commits have a parent doesn&#39;t hold
18:54jgrahamgsnedders: In any case it isn&#39;t really going to work to run the stability checker here
18:57gsneddersjgraham: oh, totally
18:57gsneddersinterestingly, there&#39;s a commit 0dc74715852432b, &quot;New repository initialized by cvs2svn.&quot;, in the middle of fairly linear history!
18 Mar 2017
No messages
Last message: 11 days and 8 hours ago