| 09:31 | <annevk> | zcorpan: did you look into changing them into <details> too? |
| 09:31 | <annevk> | zcorpan: <details><summaryStatus</summary> ...</details> |
| 09:31 | <annevk> | zcorpan: I was thinking if we did that we'd drop the "Support:" paragraph |
| 09:31 | <zcorpan> | annevk: not yet. i want to PR the tweaks i already have since now it's looking pretty good |
| 09:32 | <zcorpan> | we can do a second round for details |
| 09:32 | <annevk> | zcorpan: okay, I don't care too strongly about it either |
| 09:32 | <annevk> | zcorpan: it's a little weird that what the specification does cannot really be done with <details> |
| 09:33 | <zcorpan> | annevk: it can be done surely |
| 09:35 | <annevk> | zcorpan: well in the specification you have multiple summaries in a single <details> box |
| 09:36 | <annevk> | zcorpan: I guess you could only markup the first as the overall summary and the others with some other element, but it's a little wonky |
| 09:36 | <zcorpan> | oh that. hmm yeah |
| 09:36 | <zcorpan> | or generated content, but yes |
| 09:40 | <nox> | zcorpan: The diffstat on my PR is a lie. :D |
| 09:40 | <nox> | zcorpan: I did a first commit that enables all document.write tests, |
| 09:40 | <nox> | which added dozens of unexpected failures, |
| 09:41 | <nox> | and then removed most of them with the implementation of the damn method. :) |
| 09:41 | <nox> | +402 −0 and then +191 −575 |
| 09:41 | <zcorpan> | nox: ah :-) |
| 09:43 | <nox> | zcorpan: Reminds me of https://github.com/servo/servo/pull/5839/commits/5386cb31d05da2a8c9e8a9fb42817c107880a96b, that diffstat. :) |
| 09:43 | <nox> | Followed by +521 −98,593. |
| 09:44 | <nox> | Why code lines aren't a good metric. ^ :p |
| 09:44 | <zcorpan> | nox: clearly you're just doing it wrong. maybe all expected failures should all be on one line? |
| 09:45 | <nox> | :D |
| 09:47 | <nox> | Reminds me, |
| 09:47 | <nox> | there is no writeln() test that uses only writeln(), |
| 09:47 | <nox> | they all use open() and close() and thus I can't make them pass. |
| 09:47 | <nox> | (I didn't implement them yet. open() scares me.) |
| 09:49 | <annevk> | zcorpan: btw, did you try floating? |
| 09:50 | <annevk> | zcorpan: I was curious what floating might do, but I guess that doesn't turn out nicely for wide screens |
| 09:50 | <zcorpan> | annevk: yes, i tried that first. doesn't work because there are status boxes in tables. |
| 09:51 | <zcorpan> | annevk: and indeed it won't push the box to the edge of the viewport unless we re-do other things... |
| 09:51 | <zcorpan> | annevk: plus, reflowing text when clicking the button on single-page is not going to be cheap |
| 09:51 | <annevk> | zcorpan: yeah true |
| 09:51 | <annevk> | zcorpan: let's not do that |
| 09:51 | <annevk> | zcorpan: btw, I need you to look at https://github.com/whatwg/html/pull/2098 at some point |
| 09:52 | <zcorpan> | chekcing now |
| 10:02 | <zcorpan> | annevk: commented |
| 10:05 | <zcorpan> | annevk: PRs for style tweaks, https://github.com/whatwg/resources.whatwg.org/pull/37 https://github.com/whatwg/html/pull/2106 |
| 10:33 | <annevk> | zcorpan: I'm happy with you landing those provided you tried them out, resized things a bit, and it all looks reasonable |
| 10:34 | <annevk> | Soon we'll have validation and <span> without attributes checking, pretty exciting |
| 10:37 | <zcorpan> | annevk: ok, thx |
| 10:59 | <zcorpan> | does Domenic need to update build.whatwg.org or so whenever wattsi changes? |
| 11:04 | <annevk> | zcorpan: I'm not sure what the deal is |
| 11:05 | <annevk> | zcorpan: maybe file an issue to track it? |
| 11:08 | <zcorpan> | annevk: filed https://github.com/whatwg/wattsi/issues/37 |
| 11:15 | <MikeSmith> | annevk: what’s the purpose of the omit-Origin flag? I guess first for <img> and <embed> and <object> elements that have always been allowed to cause cross-origin fetches? |
| 11:17 | <MikeSmith> | and same-origin requests? |
| 11:18 | <MikeSmith> | per spec any fetch cors-mode request is always required to send Origin? |
| 11:20 | <annevk> | MikeSmith: that is basically a bug |
| 11:20 | <annevk> | MikeSmith: hard to get browsers to work on resolving it |
| 11:20 | <annevk> | Maybe I should give that another push later today |
| 11:21 | <annevk> | MikeSmith: see Origin header issue |
| 11:28 | <MikeSmith> | annevk: issue where? |
| 11:28 | <botie> | somebody said issue was that blink's callback interfaces are just callback functions |
| 11:29 | <MikeSmith> | https://github.com/whatwg/fetch/issues/225 |
| 11:29 | <MikeSmith> | “Should we send an Origin header for no-cors fetches?” from ehsan? |
| 13:06 | <annevk> | MikeSmith: yeah, sorry |
| 13:15 | <annevk> | MikeSmith: it seems https://github.com/whatwg/html/pull/2093/files can be copied to other Travis-using repositories without much trouble, that's really quite nice |
| 13:20 | <annevk> | I wonder how the W3C TAG ever got to publish https://github.com/w3ctag/client-certificates |
| 13:20 | <annevk> | Who there has the expertise? |
| 15:21 | <Ms2ger> | annevk, ping |
| 15:22 | <Ms2ger> | annevk, umping |
| 15:22 | <Ms2ger> | unping, even |
| 15:22 | <annevk> | Ms2ger: yo unyo |
| 15:22 | <Ms2ger> | I was looking for the definition of the origin of the document returned by createHTMLDocument, but it was hidden, not missing |
| 15:34 | <annevk> | zcorpan: http://speleotrove.com/decimal/decifaq.html |
| 15:39 | <zcorpan> | annevk: thanks. per http://speleotrove.com/decimal/decifaq1.html#needed it seems something like 32 digits is necessary |
| 15:40 | <annevk> | zcorpan: if we ever did something though we'd introduce something closer to Java BigDecimal/C decNumber I hope |
| 15:43 | <zcorpan> | annevk: yeah |
| 15:51 | <annevk> | Sam Ruby tried to get decimals into JavaScript at some point and failed |
| 15:51 | <annevk> | I looked that up a couple weeks ago, but I forgot again why it failed |
| 15:52 | <annevk> | Suspect details weren't worked out to the satisfaction of everyone |
| 15:52 | <Ms2ger> | annevk, https://html.spec.whatwg.org/multipage/scripting.html#appropriate-template-contents-owner-document does seem to be missing the definition of its origin, though |
| 15:52 | <annevk> | Ms2ger: nice find, file an issue? |
| 15:53 | <annevk> | Ms2ger: maybe we should start defining a "create a Document" algorithm that everyone hooks into... |
| 15:53 | <Ms2ger> | Perhaps, yes |
| 15:54 | <annevk> | Running the HTML checker on standards is a great idea MikeSmith |
| 15:54 | <annevk> | MikeSmith: already finding numerous issues that should have been solved ages ago |
| 15:54 | <annevk> | MikeSmith: Travis will soon run a lot of Java, hehe |
| 17:22 | <annevk> | Domenic: just noticed the HTML Standard doesn't use Travis to deploy yet |
| 17:31 | <MikeSmith> | annevk: gotta keep the Java spirit alive :) |
| 17:31 | <MikeSmith> | anyway glad the checker is being useful |
| 17:36 | <annevk> | MikeSmith: I did something wrong and it didn't end up failing which was somewhat surprising |
| 17:36 | <annevk> | MikeSmith: I hadn't corrected the path |
| 17:36 | <MikeSmith> | ah |
| 17:36 | <MikeSmith> | yeah I should refine it to fail in that case |
| 17:37 | <annevk> | MikeSmith: for Fetch I should use /home/travis/build/whatwg/fetch/*.html, right? |
| 17:37 | <MikeSmith> | yep |
| 17:38 | <annevk> | MikeSmith: see https://github.com/whatwg/fetch/pull/426 for my WIP |
| 17:38 | MikeSmith | looks |
| 17:48 | <annevk> | MikeSmith: it doesn't seem to work |
| 17:49 | <MikeSmith> | annevk: yeah looking at it now |
| 17:49 | <annevk> | MikeSmith: does /home/travis/build/whatwg/fetch/*.html actually go into subdirectories? |
| 17:49 | <MikeSmith> | oh |
| 17:49 | <MikeSmith> | not it does not |
| 17:49 | <annevk> | MikeSmith: okay, so for HTML it's wrong too then |
| 17:49 | <MikeSmith> | oh? |
| 17:50 | MikeSmith | looks back HTML config |
| 17:50 | <MikeSmith> | ah yeah |
| 17:51 | <annevk> | JakeA: have you ever had a need for a promise/event when document.body was inserted? |
| 17:52 | <annevk> | MikeSmith: is there an elegant fix? |
| 17:58 | <MikeSmith> | yes |
| 17:58 | <MikeSmith> | https://github.com/whatwg/html/pull/2110 |
| 17:58 | <MikeSmith> | sorry was just careless about it before |
| 17:59 | <MikeSmith> | please try that invocation in the fetch travis.yml |
| 18:00 | <MikeSmith> | the --skip-non-html option is because by default vnu.jar will try to parse and check any file you give it or any file it finds when recursing through directory |
| 18:01 | <MikeSmith> | regardless of whether it has a .html or .htm or .xhtml or .xht extension |
| 18:02 | <MikeSmith> | --skip-non-html causes it to only check files with those extensions |
| 18:07 | <gsnedders> | foolip, Domenic, rbyers: https://bugs.chromium.org/p/chromium/issues/detail?id=668172 covers sorta cases where Blink has diverged from the HTML spec wrt parsing; the linked log is p. useful |
| 18:07 | <gsnedders> | foolip, Domenic, rbyers: basically, a bunch of the removed SVG stuff has had it's parser special cases removed |
| 18:10 | <MikeSmith> | annevk: if you give it the --verbose option it will report all filenamesーboth the names of HTML files it actually checks and a warning message for each non-HTML file (if there any) it skips |
| 18:16 | <annevk> | MikeSmith: done, we'll see how it goes |
| 18:16 | <annevk> | MikeSmith: didn't add --verbose for now |
| 18:17 | <MikeSmith> | dang |
| 18:18 | <MikeSmith> | https://travis-ci.org/whatwg/fetch/builds/179524939 exited with 0 |
| 18:18 | <MikeSmith> | will try it locally |
| 18:21 | <annevk> | MikeSmith: per https://github.com/whatwg/fetch/pull/426 one of the checks did fail |
| 18:21 | <annevk> | MikeSmith: kinda curious |
| 18:22 | <MikeSmith> | ah yeah |
| 18:22 | <MikeSmith> | I noticed that with HTML as well |
| 18:22 | <annevk> | MikeSmith: so it would fail on master, but doesn't fail within the PR |
| 18:22 | <annevk> | MikeSmith: that's something, but not great |
| 18:23 | <MikeSmith> | continuous-integration/travis-ci/push will fail if there’s an error on the PR branch but continuous-integration/travis-ci/pr will not |
| 18:23 | <MikeSmith> | I don’t understand why |
| 18:24 | <MikeSmith> | I guess I don’t actually understand why there are two to begin with |
| 18:27 | <MikeSmith> | ah is it because it skips running the deploy.sh script for PR branches? |
| 18:27 | <MikeSmith> | and the deploy.sh script is what produces the generated HTML? |
| 18:28 | <annevk> | MikeSmith: no we run it for both |
| 18:28 | <annevk> | MikeSmith: there's different output, but it always runs |
| 18:29 | <annevk> | MikeSmith: is the directory recursion infinite? |
| 18:29 | <MikeSmith> | OK then yeah I don’t understand why the checker finds no errors for the continuous-integration/travis-ci/pr one |
| 18:29 | <MikeSmith> | yeah it is infinite |
| 18:29 | <annevk> | Let me re-run it just in case |
| 18:29 | <MikeSmith> | OK |
| 18:30 | <MikeSmith> | anyway another thing to note is that the warnings will not cause the travis build to faile |
| 18:30 | <MikeSmith> | because the warnings do not cause vnu.jar to exit non-zero |
| 18:31 | <annevk> | That's okay, I wonder where Fetch uses <details> now, wasn't aware of that |
| 18:31 | <MikeSmith> | OK |
| 18:31 | <MikeSmith> | if you ever do want to suppress warnings you can give it the --errors-only option |
| 18:31 | <annevk> | Hmm, I cannot find details |
| 18:31 | <MikeSmith> | oh |
| 18:31 | <MikeSmith> | maybe bikeshed |
| 18:31 | <MikeSmith> | adds something |
| 18:32 | <annevk> | even locally I can't find "<details" in the generated HTML |
| 18:33 | <annevk> | Can that warning be generated for other reasons? |
| 18:33 | <MikeSmith> | no |
| 18:33 | <MikeSmith> | should be something on line 90 |
| 18:33 | <MikeSmith> | it claims |
| 18:33 | <MikeSmith> | line 90 column 5 |
| 18:34 | <MikeSmith> | trying to test it with bikeshed locally, I get: |
| 18:34 | <MikeSmith> | $ bikeshed |
| 18:34 | <MikeSmith> | FATAL ERROR: Found unmatched text macro [SNAPSHOT-LINK]. Correct the macro, or escape it with a leading backslash. |
| 18:34 | <MikeSmith> | ✘ Did not generate, due to fatal errors |
| 18:34 | <annevk> | MikeSmith: you need to run make |
| 18:34 | <MikeSmith> | ah |
| 18:34 | <annevk> | MikeSmith: or ./build.sh --local |
| 18:34 | <MikeSmith> | k |
| 18:35 | <annevk> | MikeSmith: sorry, deploy.sh |
| 18:35 | <MikeSmith> | ok |
| 18:36 | <smaug> | no wpt tests for IntersectionObserver o_O |
| 18:36 | <annevk> | MikeSmith: sorry, I think you're right "Skipping deploy for a pull request; the branch build will suffice" |
| 18:36 | <MikeSmith> | annevk: 90 <details class="annoying-warning" open=""> |
| 18:37 | <MikeSmith> | annevk: yeah I think HTML does that too? skips it I mean |
| 18:37 | <annevk> | MikeSmith: aaah, that's a snapshot only thing |
| 18:37 | <MikeSmith> | OK |
| 18:37 | <annevk> | MikeSmith: not sure what HTML does exactly |
| 18:37 | <MikeSmith> | ok |
| 18:37 | <MikeSmith> | so yeah for fetch it is checking all the .html files, all versions/subdirs |
| 18:39 | <MikeSmith> | if you want to skip some subdirs then I guess you would need to have travis call vnu.jar on only the subtrees in there you want |
| 18:40 | <annevk> | MikeSmith: I'm happy with this |
| 18:40 | <MikeSmith> | OK |
| 18:40 | <MikeSmith> | sorry for the churn |
| 18:41 | <MikeSmith> | yeah anyway as long as a PR is making it run continuous-integration/travis-ci/push that will fail if there are errors, and the github UI will show the red X and all that |
| 18:41 | <MikeSmith> | so that gets us what we want |
| 18:42 | <MikeSmith> | about that <details> warning maybe it is time I make the checker stop emitting that anyway |
| 18:42 | <MikeSmith> | the table warnings are real though |
| 18:46 | <annevk> | MikeSmith: the table warnings were only because of my bogus table to cause an error I think |
| 18:46 | <annevk> | MikeSmith: the build now passes, I'll double check for any emitted warnings |
| 18:47 | <MikeSmith> | super |
| 18:47 | <annevk> | Just warnings for <details> |
| 18:47 | <annevk> | Very nice |
| 18:47 | <MikeSmith> | excellent |
| 18:48 | <MikeSmith> | OK with that I will step away for sleep |
| 18:48 | <MikeSmith> | if there are any other problems I can look in a few hours |
| 18:49 | <annevk> | MikeSmith: sleep a little longer than a few hours pretty please |
| 18:49 | <MikeSmith> | heh |
| 18:49 | <MikeSmith> | will try |
| 20:08 | <foolip> | gsnedders: thanks for noticing and filing, let's see if it gets triaged without poking. |
| 20:09 | <foolip> | gsnedders: are these tests also in web-platform-tests, and should they be? |
| 20:14 | <gsnedders> | foolip: they are |
| 20:15 | <gsnedders> | foolip: hence them being in third_party/WebKit/LayoutTests/imported/wpt |
| 20:15 | <foolip> | gsnedders: did you check if those tests are imported in Blink and failing as expected? there may already be a bug tracking those failures |
| 20:17 | <gsnedders> | foolip: they are, they are, I didn't find one |
| 20:18 | <foolip> | gsnedders: ok :) |
| 20:20 | gsnedders | really wishes those tests were better organised, but invalidating everyone's expectations isn't great either |
| 22:09 | <nox> | https://github.com/servo/servo/blob/4d93ee134c84d72f73d7a5556bd2dcaf36e5ca99/components/script/dom/servoparser/mod.rs#L43-L54 <- How I handled the insertion point |
| 22:09 | <nox> | With https://github.com/servo/servo/blob/4d93ee134c84d72f73d7a5556bd2dcaf36e5ca99/components/script/dom/servoparser/mod.rs#L163-L172 |
| 22:28 | <JakeA> | annevk: um, not specifically, but I tend to use script[defer]. If I was using async, a way to know a particular element was present would be handy, but I guess mutation observers provide that |
| 22:28 | <zcorpan> | nox: 👍 |
| 22:28 | <nox> | zcorpan: Thanks. :) |
| 22:33 | <nox> | zcorpan: Now if we could get a decision for https://github.com/whatwg/html/issues/1954... :) |
| 22:41 | <zcorpan> | nox: the black-box equivalent of Gecko, Chromium, WebKit, seems reasonable |
| 22:42 | <nox> | What does that mean? |
| 22:44 | <zcorpan> | that you should get "1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19," for my test case in Servo |
| 22:44 | <nox> | Oh. |
| 23:23 | <zcorpan> | annevk: interest in https://www.w3.org/Bugs/Public/show_bug.cgi?id=28555 ? |