| 07:49 | <smaug____> | blink has the some 0.x% usage limit for removing APIs |
| 07:50 | <smaug____> | anyone recalls what that number is |
| 07:50 | <smaug____> | was it something like 0.01? |
| 07:50 | <smaug____> | or 0.1 |
| 07:50 | <smaug____> | foolip: ^ |
| 07:55 | <annevk> | smaug____: 0.003 iirc |
| 07:55 | <annevk> | smaug____: but they use it more as an indicator of sorts these days, not as a hard line |
| 07:55 | <smaug____> | sure |
| 07:57 | <foolip> | TabAtkins: do you mean the thing I've added to lots of specs? |
| 07:57 | <foolip> | foolip: I pasted that in https://github.com/tabatkins/bikeshed/issues/1001, but I imagine that's not what you're asking for? |
| 07:58 | <foolip> | smaug____: we used to say that 0.03% was some magic limit, but we've removed mention of that from everywhere now I hope |
| 07:58 | <foolip> | certainly chromestatus.com no longer has a different color for things below that cutoff |
| 07:59 | <foolip> | smaug____: things with >1% usage have been removed, it all depends on likely breakage in the wild |
| 07:59 | <smaug____> | ok |
| 09:11 | <espadrine> | I presume NPAPI was a bit above |
| 09:34 | <smaug____> | annevk: how documents get destroyed isn't spec'ed anywhere, right? I mean, how the DOM tree is unraveled and so. Since internally there are operations similar to 'remove' http://w3c.github.io/webcomponents/spec/custom/#concept-node-remove but of course we don't want to call the custom element callbacks then. |
| 09:47 | <annevk> | smaug____: no, expectation is that is not observable |
| 12:29 | <TabAtkins> | foolip: No, you had a gist or codepen or something containing a template for a wpt test, that you'd start from when writing one. |
| 12:32 | <jgraham> | TabAtkins: http://searchfox.org/mozilla-central/source/testing/web-platform/mach_commands.py#165 might be similar to what you want |
| 12:37 | <foolip> | TabAtkins: ah, that was from https://docs.google.com/document/d/1LSuLWJDP02rlC9bOlidL6DzBV5kSkV5bW5Pled8HGC8/edit?usp=sharing |
| 12:37 | <foolip> | and https://jsbin.com/pazubak/edit?html,output is the template |
| 12:37 | <TabAtkins> | Excellent. |
| 14:04 | <zcorpan> | TabAtkins: note that the test needs to use /resources/testharness.js as the URL, unlike in that template |
| 14:12 | <Domenic> | smaug____: annevk: yeah that limit is not real anymore and really should stop being repeated. We have no limit; it's all case by case judgment. |
| 14:12 | <Domenic> | I see foolip said that too |
| 14:15 | <Domenic> | annevk: what do you mean by "the test (subtest two in particular) you pointed to above should also be fixed" ? |
| 14:16 | <annevk> | Domenic: it does an array.length assert where it equals 5 and then shortly after compares the array to another array of 4 items |
| 14:17 | <annevk> | Domenic: looks like that was my fault |
| 14:17 | <annevk> | Domenic: patch coming up |
| 14:18 | <Domenic> | OK. So LGTM on spec patch but will wait to see test changes as I think it'll be easier to tell what's going on there with a patch. |
| 14:20 | <annevk> | Domenic: you already changed the existing tests last time around |
| 14:20 | <Domenic> | Sure, but now we have a full rewrite, and maybe also some fix about 5 vs. 4, I'm not sure what the actual plan is anymore |
| 14:21 | <annevk> | Domenic: I don't understand |
| 14:21 | <Domenic> | So the plan is not to merge https://github.com/w3c/web-platform-tests/pull/7266 , right? Since that tests something we haven't agreed on. |
| 14:22 | <Domenic> | https://github.com/whatwg/html/pull/3013#issuecomment-327493192 says that's the plan, but I'm assuming it's not the plan anymore. |
| 14:23 | <Domenic> | Then some test (still not sure which one) has some 5 vs. 4 issue, so you're thinking about submitting a separate patch for that? Is that the new plan for testing whatwg/html#3013? |
| 14:23 | <Domenic> | Or do we not need to test whatwg/html#3013 at all since it was already tested way back when? |
| 14:23 | <annevk> | Domenic: those tests need to be reviewed, but I don't think that should block this simple bug in the HTML Standard |
| 14:23 | <Domenic> | Right, as I said, spec patch LGTM, but I'll wait to see what's going on with test changes. |
| 14:24 | <annevk> | Domenic: yeah, I don't think we necessarily need tests for this change |
| 14:24 | <annevk> | Domenic: I submitted a patch for the 5/4 thing |
| 14:24 | <annevk> | Domenic: https://github.com/w3c/web-platform-tests/pull/7267 |
| 14:24 | <annevk> | Domenic: the assert is redundant with the assert_array_equals, which does that length check already |
| 14:24 | <Domenic> | Got it |
| 14:27 | <annevk> | If I don't forget I'll look into those tests some more tomorrow |
| 17:10 | <TabAtkins> | Domenic: While not real, the limit is still an okay eyeball figure, I think. |
| 22:59 | <MikeSmith> | TabAtkins: I'm wondering why in http://wpt.fyi/css the css-page-3 subdir doesn’t show up, despite being in https://github.com/w3c/web-platform-tests/tree/master/css/css-page-3 |
| 22:59 | <MikeSmith> | is there something different about those tests that would cause them to not get picked up? |
| 23:00 | <TabAtkins> | MikeSmith: I haven't a single clue about any of that. |
| 23:00 | <MikeSmith> | like, those tests can’t be automated |
| 23:00 | <MikeSmith> | OK |
| 23:04 | <MikeSmith> | I guess maybe they’re print-only |
| 23:05 | <MikeSmith> | anyway the specific thing I was looking for is a @bottom-right test |
| 23:05 | <TabAtkins> | They, those tests appear to all be manual tests that you need to print and verify yourself. |
| 23:05 | <MikeSmith> | and looking at http://w3c-test.org/css/work-in-progress/hp/margin-boxes-012.xht it seems like it’s maybe a print-only thign |
| 23:05 | <MikeSmith> | yeah |
| 23:06 | <MikeSmith> | > This test requires 3 pages. When printed, the green box on this page should be equidistant from the blue boxes above and below it. |