| 01:11 | AryehGregor | just got a message from an at job: Right or wrong? #whatwg [100125 13:52:46] <AryehGregor> I predict that in two years, normal people using new browsers won't need to have Flash installed. |
| 01:11 | <AryehGregor> | Arguably right, if one construes "need" narrowly enough. |
| 01:12 | <AryehGregor> | Random YouTube videos seem to work out-of-the-box in Chrome dev in incognito, so I'd say it's true. |
| 01:13 | <Hixie> | chrome dev comes with flash built in, iirc |
| 01:13 | <AryehGregor> | about:plugins says it's disabled. |
| 01:13 | <AryehGregor> | Only have Remoting Viewer, NaCl, and the PDF viewer. |
| 01:14 | <AryehGregor> | Also, right-clicking on the video says "About HTML5" in the menu. |
| 01:14 | <AryehGregor> | So it's HTML5 all right, and I got it without opting in (due to incognito). |
| 01:14 | <Dashiva> | Yeah, Youtube used to be pretty bad, but I don't recall seeing a flash-only video in quite a while |
| 01:15 | <AryehGregor> | Vimeo too. |
| 01:15 | <Dashiva> | Actually, I should say flash and/or h264-only |
| 01:15 | <AryehGregor> | (works with HTML5 out of the box) |
| 01:15 | AryehGregor | made no claims about H.264-only |
| 01:15 | <Hixie> | nice |
| 01:16 | <Dashiva> | http://downforeveryoneorjustme.org/facebook.com |
| 01:16 | <Dashiva> | Seeing that 90% of the recent log is people checking facebook.com, I wonder... |
| 01:17 | <AryehGregor> | I've browsed with Flash disabled for a long time, and the only things I can remember seeing recently that didn't work were ads. |
| 01:17 | <AryehGregor> | And not even many of that. |
| 01:17 | <AryehGregor> | So: long live HTML5! Flash is no longer important! |
| 01:17 | <AryehGregor> | Hurrah! |
| 01:17 | AryehGregor | pats everyone in the room on the back |
| 01:18 | <Hixie> | hehe |
| 01:18 | <Dashiva> | Hey |
| 01:18 | <Dashiva> | Flash games are still flash games |
| 01:23 | <Hixie> | at least on g+ some of the games are canvas-based |
| 01:36 | <divya> | Dashiva: do you have the logs for downforeveryoneorjustme? |
| 01:36 | <divya> | or is it public? >_> |
| 01:38 | <Hixie> | it's on the page |
| 02:38 | <AryehGregor> | Where's the giant warning box in the HTML spec saying not to use canvas to make a text editor? |
| 02:39 | <AryehGregor> | Ah, here. |
| 02:40 | <AryehGregor> | http://www.whatwg.org/specs/web-apps/current-work/multipage/the-canvas-element.html#best-practices |
| 02:40 | <Hixie> | it's got a whole section, not just a box :-) |
| 02:45 | <AryehGregor> | Yeah, but it's not in scary warning colors, so I didn't spot it when skimming through. |
| 02:58 | <MikeSmith> | the spec needs a EULA |
| 02:59 | <MikeSmith> | "By using the technologies we make available in this specification, you agree to not make things that are bad for the Web." |
| 03:00 | <MikeSmith> | and if you break that agreement you get kicked off the Web |
| 03:00 | <MikeSmith> | for two weeks |
| 03:21 | <Hixie> | MikeSmith: man, if we had the power to do that... |
| 07:00 | <jamesr> | how do you make a path in SVG that isn't closed? |
| 07:00 | <jamesr> | i tried just ending on an L instead of a z but it's bitching at me Problem parsing d="M 272 202 L 274 202 L" |
| 07:01 | <jamesr> | oh wait, i think i just leave off the last command |
| 07:05 | jamesr | feels dumb now |
| 07:12 | <jamesr> | ok better SVG question. i have a path with a bunch of points. is there a way to say "stroke it such that the first point is blue, the last point is green, and everything in the middle is linearly interpolated between the two based on the point's distance along the path" ? |
| 07:13 | <jamesr> | i've tired specifying the paint server as a linearGradient, but that gradient seems to be evaluated in the viewboxes' space no matter what i do |
| 07:13 | <jamesr> | so i can make stuff on the left blue and stuff on the right green, but can't figure out how to make the color change as a function of progress along the path |
| 07:17 | <jamesr> | interbutts as of 2007 say SVG doesn't do that: https://groups.google.com/group/mozilla.dev.tech.svg/browse_thread/thread/6d557ad470cf9cbc/5ad0cf5b1a23188a |
| 07:17 | <jamesr> | :( |
| 08:21 | <remysharp> | Trying to work out which browser has the bug: I create an event, set a custom property on it and dispatch it. Chrome carries the custom prop through, Firefox drops it. Does the spec define what the browser should do? |
| 08:55 | <MikeSmith> | hsivonen: dunno if you're around but I wanted to ask if there's a reason why for the V.nu site you aren't serving the static files from jetty |
| 08:56 | <MikeSmith> | I mean the style.css and script.js files |
| 08:56 | <hsivonen> | MikeSmith: this way, I have less need to track Jetty security updates and less need to figure out how to make Jetty serve static files when V.nu has its own main() |
| 08:56 | <hsivonen> | I've forgotten to update script.js, haven't I? |
| 08:57 | <MikeSmith> | hsivonen: no no |
| 08:57 | <MikeSmith> | I haven't landed the new code yet so you don't need to update yet |
| 08:57 | <MikeSmith> | so I know how to make jetty serve static files |
| 08:57 | <hsivonen> | MikeSmith: I've considered reading style.css and script.js to RAM and serving them from the servlet, though. I just haven't gotten around to it |
| 08:57 | <MikeSmith> | I have the code working in my workspace |
| 08:58 | <MikeSmith> | OK |
| 08:58 | <MikeSmith> | that would be a better solution I guess |
| 08:58 | <hsivonen> | MikeSmith: I'm most worried about Jetty's generic file serving serving other files accidentally |
| 08:59 | <MikeSmith> | yeah |
| 08:59 | <MikeSmith> | well, the reason I'm looking into this is that for the w3c instance I would really like to have everything in one place |
| 08:59 | <hsivonen> | MikeSmith: I'm OK with having the code in the repo |
| 09:00 | <hsivonen> | MikeSmith: not sure if I want it enabled on the V.nu instance without seeing the code |
| 09:00 | <MikeSmith> | yeah, sure, understood |
| 09:00 | <MikeSmith> | and I can put it under the w3c flag I think |
| 09:00 | <hsivonen> | MikeSmith: ok. |
| 09:00 | <MikeSmith> | all of it |
| 09:00 | <MikeSmith> | I'll get you a patch |
| 09:01 | <hsivonen> | or more to the point, I don't want it enabled on V.nu without seeing the Jetty guts it calls |
| 09:01 | <MikeSmith> | yup |
| 09:01 | <MikeSmith> | understood |
| 09:02 | <MikeSmith> | but now I'm wondering if it's better just to pursue what you mentioned about reading style.css and script.js to RAM and serving them from the servlet |
| 09:07 | <MikeSmith> | hsivonen: btw, is there a way I can detect within the code what path jetty is running under? |
| 09:07 | <MikeSmith> | so that I don't have to use absolute paths |
| 09:08 | <MikeSmith> | right now I'm just feeding it the path through a system property |
| 09:11 | <hsivonen> | MikeSmith: I don't know. There might be a magic system property for it, but I don't see a dedicated JDK API call for it |
| 09:12 | <MikeSmith> | ok |
| 09:15 | <MikeSmith> | hsivonen: what code would I need to touch if I wanted to implement reading style.css and script.js to RAM and serving them from the servlet? |
| 09:17 | <hsivonen> | MikeSmith: nu.validator.servlet.VerifierServlet |
| 09:17 | <MikeSmith> | OK |
| 09:19 | <MikeSmith> | ah |
| 09:19 | <MikeSmith> | hsivonen: doGet? |
| 09:20 | <hsivonen> | MikeSmith: yeah |
| 09:20 | <MikeSmith> | OK |
| 09:20 | <hsivonen> | MikeSmith: you could read the files to RAM from the static initializer |
| 09:21 | <hsivonen> | MikeSmith: it's dirty, but there's already IO from static initializer in the app |
| 09:21 | <MikeSmith> | I see |
| 09:25 | <MikeSmith> | hsivonen: so the code is doing this already for serving /robots.txt right? |
| 09:25 | <hsivonen> | MikeSmith: yeah |
| 09:26 | <MikeSmith> | could this be done for serving image files as well? |
| 09:26 | <hsivonen> | MikeSmith: sure |
| 09:28 | <MikeSmith> | hsivonen: OK, well, if you prefer I can write a patch that does it this way instead of using the jetty static-file serving stuff |
| 09:28 | <hsivonen> | MikeSmith: that would be awesome |
| 09:28 | <MikeSmith> | will do then |
| 09:28 | <hsivonen> | thanks |
| 09:30 | <MikeSmith> | hsivonen: to be clear, we still need to files in the repo and the code reads them in from the filesystem, right? |
| 09:30 | <MikeSmith> | if so, my next question is, how do I get the path to the directory they're in? |
| 09:31 | <hsivonen> | MikeSmith: I'd read them from the file system to a byte[] from static {} and then serve them from RAM the way robots.txt is served from RAM |
| 09:31 | <MikeSmith> | yup, OK |
| 09:31 | <hsivonen> | MikeSmith: hmm. how are the local entities located... |
| 09:31 | <MikeSmith> | ah yeah |
| 09:31 | <hsivonen> | oh. they come from within a JAR |
| 09:31 | <MikeSmith> | yup |
| 09:31 | <MikeSmith> | I could do the same for these I guess |
| 09:32 | <hsivonen> | so one option would be moving the files so that they get baked in a JAR |
| 09:32 | <hsivonen> | yeah |
| 09:32 | <MikeSmith> | I mean, just use that same mechnaism |
| 09:32 | <MikeSmith> | OK |
| 09:32 | <hsivonen> | makes sense |
| 09:32 | <MikeSmith> | yup |
| 09:32 | <MikeSmith> | OK |
| 09:32 | <hsivonen> | hmm. though a JAR *might* make the length of the file unknown |
| 09:33 | <MikeSmith> | oh |
| 09:33 | <hsivonen> | which will result in some reallocation and copying of byte[] if you can't allocate the right-sized byte[] in advance |
| 09:33 | <MikeSmith> | oh |
| 09:33 | <MikeSmith> | why is that not also a problem for the existing files that get JARed up? |
| 09:33 | <MikeSmith> | oh |
| 09:34 | <MikeSmith> | I see what you mean |
| 09:34 | <MikeSmith> | yeah |
| 09:34 | <hsivonen> | they don't get read into a file-length byte[] |
| 09:34 | <MikeSmith> | right |
| 10:14 | <mhausenblas> | http://lists.w3.org/Archives/Public/www-archive/2012Jan/0032.html ... he he, well played, Hixie :D |
| 10:21 | <asmodai> | Who here was from the Opera team again (seems that anne's not around)? |
| 10:32 | <zcorpan> | jgraham is here |
| 10:43 | jgraham | is indeed here |
| 11:18 | <cweiske> | about the autocomplete proposal: why isn't there an openid, username and password token? |
| 11:22 | <asmodai> | jgraham: Opera mobile, it's known that the previews of tabs sometimes keep showing wrong content? |
| 11:25 | <jgraham> | asmodai: Not known to me at least. Could you file a bug, please? |
| 11:27 | <asmodai> | jgraham: going to try, internet is acting up at the moment |
| 11:28 | <asmodai> | major lag and packet loss due to some upstream router on the campus reaching CPU limits >_< |
| 11:30 | <jgraham> | https://bugs.opera.com/wizard/mobile/ if it helps |
| 11:32 | <asmodai> | jgraham: ANDMEX-5712 |
| 11:33 | <asmodai> | jgraham: Would a screenshot help? (assuming I can send such files to the address as well) |
| 11:35 | <jgraham> | asmodai: …maybe? I don't really do product QA so I don't know exactly what they use. I imagine a screenshot is not needed if it is 100% reproducable, but might be helpful if it isn't (to prove you had the problem :) |
| 11:36 | <wilhelm> | asmodai: Yes. |
| 11:36 | <wilhelm> | (Not that I work for Opera anymore. :) |
| 11:37 | <jgraham> | But at least you have some experience with product QA :) |
| 11:41 | <asmodai> | wilhelm: hehe, cool :) |
| 11:41 | <asmodai> | jgraham: Will check if I can show the previews and use the native screenshot button to snap it. |
| 11:43 | <jgraham> | asmodai: Thanks, this is really helpful! |
| 11:43 | <asmodai> | jgraham: np, so far opera mobile is my browser of choice on Android |
| 11:45 | <wilhelm> | asmodai: Even with the Opera hat gone, I can agree with that. (c: |
| 11:46 | <wilhelm> | asmodai: For later reference: by adding a screenshot, you're proving that the bug actually exists. That may make your report float to the top. |
| 11:47 | <jgraham> | Things I was vaugely aware of, but just saw demoed: you can actually use the mobile browser on your desktop: http://www.opera.com/developer/tools/mobile/ |
| 11:49 | <jgraham> | wilhelm: yes. Although you shouldn't take the same screenshot and post it to *every* *single* blog post whenever there's a new nightly build. |
| 11:52 | <gsnedders> | GlitchMr: But it still happens! |
| 11:52 | <gsnedders> | GlitchMr? |
| 11:52 | <gsnedders> | jgraham. |
| 11:52 | <GlitchMr> | Yeah |
| 11:52 | <gsnedders> | I can't type. |
| 11:52 | <GlitchMr> | wait.. lol. |
| 12:06 | <asmodai> | wilhelm: Good to know |
| 12:06 | <asmodai> | wilhelm: When I'm home and actually near my tablet, I'll snap and upload it |
| 12:38 | <MikeSmith> | karlcow: you walked right into that one man |
| 12:39 | <MikeSmith> | the reply from Roy on public-tracking |
| 12:39 | <MikeSmith> | or maybe you made him walk right into that one |
| 12:39 | <MikeSmith> | karlcow: were you intentionally trolling him? |
| 13:33 | <karlcow> | MikeSmith: I rarely troll anyone, or if I troll it is because the person is a friend who will be able to understand the nature of my comment. |
| 13:34 | <karlcow> | for the answer of Roy, I do not understand why it went on that reply which is barely related to what we were discussing in the tracking protection group this morning. |
| 13:34 | <karlcow> | s/it went/he went/ |
| 13:35 | <cweiske> | about the autocomplete proposal: why isn't there an openid, username and password token? |
| 13:43 | <jgraham> | cweiske: You asked that about 2.5 hours ago |
| 13:43 | <jgraham> | cweiske: I suggest asking on the list; the people driving that proposal aren't here |
| 14:19 | <cweiske> | ok |
| 14:35 | <Ms2ger> | Does IE use 1px for thin borders already? |
| 15:43 | <scott_gonzalez> | Hixie: Does WHATWG get involved with PAGs for W3C specs at all? |
| 15:44 | <scott_gonzalez> | I've got an attorney that's willing to help investigate the Apple patents for Touch Events, but the response from Rigo Wenning a bit abrupt and standoffish. |
| 15:44 | <scott_gonzalez> | I'm not sure if there's some other/better way for the attorney to provide help. |
| 16:11 | AryehGregor | grumbles about his flaky router |
| 16:59 | <AryehGregor> | Is the user-select property specced anywhere these days? |
| 17:00 | <Ms2ger> | If CSSUI doesn't have it, probably not |
| 17:01 | <AryehGregor> | Sad. |
| 17:01 | <Ms2ger> | CSS? Yes |
| 17:02 | <MikeSmith> | hsivonen: sent you a patch for review, for the serve-stuff-from-RAM thing we talked about earlier |
| 17:04 | <MikeSmith> | I don't know what the best way is to read from a file into a byte array, but the way I did seems sane |
| 17:04 | <MikeSmith> | I didn't do the JAR thing because I found that actually could use a relative path |
| 17:06 | <hsivonen> | MikeSmith: ok. I'll take a look tomorrow. gotta go now |
| 17:07 | <MikeSmith> | thanks |
| 17:07 | <MikeSmith> | no rush |
| 17:54 | <crankharder> | ideas what "NOT_FOUND_ERR: DOM IDBDatabase Exception 3" might mean? I can't find any docs on it |
| 18:55 | <AryehGregor> | Huh . . . -webkit-perspective behaves weirdly. |
| 18:56 | <AryehGregor> | It seems like its value for practical purposes is equal to the value of the last ancestor with -webkit-perspective not set to "none". So <div style="-webkit-perspective:100px">, <div style="-webkit-perspective:100px"><div style="-webkit-perspective:100px">, and <div style="-webkit-perspective:100px"><div style="-webkit-perspective:none"> are all equivalent. |
| 18:56 | <AryehGregor> | This seems very weird. |
| 18:57 | <AryehGregor> | In Gecko it's just the parent's value that's used. |
| 18:57 | <AryehGregor> | I'd have expected they accumulate, so it's like transform: perspective(100px) . . . |
| 18:57 | AryehGregor | wonders what the use-case is |
| 19:20 | <AryehGregor> | Oh, awesome, WebKit's getBoundingClientRect() just doesn't match its rendering here, so it gets an unjustified pass. |
| 19:21 | <Ms2ger> | Sounds like webkit |
| 19:21 | <AryehGregor> | I found cases where that happens in Gecko too. |
| 19:23 | <AryehGregor> | https://bugzilla.mozilla.org/show_bug.cgi?id=718809 |
| 19:23 | <Ms2ger> | Shh ;) |
| 19:41 | <AryehGregor> | I wish I had a way to run reftests from JavaScript. :( |
| 19:42 | <AryehGregor> | Why does no one but Gecko seem to support writing SVG-in-data-URLs to canvas and then reading back the results? |
| 19:42 | <AryehGregor> | Extremely sad. |
| 19:42 | <AryehGregor> | Running reftests in arbitrary browsers is so painful as it stands. |
| 19:44 | AryehGregor | notes that all browsers would fail his autogenerated reftests anyway, because they all have slight errors in rounding |
| 19:44 | <Philip`> | AryehGregor: Security, presumably |
| 19:45 | <Philip`> | (In particular the difficulty of being absolutely certain that it's secure) |
| 19:45 | <AryehGregor> | Philip`, browsers that support SVG-in-<img> don't run any script or load any resources from the network, so it's pretty hard for it to not be secure. |
| 19:45 | <AryehGregor> | I guess it could reveal prefs and such. |
| 19:45 | <AryehGregor> | Hmm, I wonder if it evades the Gecko mitigation of link-color examination? |
| 19:46 | AryehGregor | fidgets |
| 19:50 | <AryehGregor> | They seem to have thought of that. Sad. |
| 19:50 | <AryehGregor> | (or, good) |
| 20:00 | <Hixie> | wait, what? |
| 20:00 | <AryehGregor> | Wait, what what? |
| 20:00 | <Hixie> | being able to reference hidden="" content from aria attributes is controversial now? |
| 20:00 | <Hixie> | wtf |
| 20:01 | <Hixie> | i feel like i am in some sort of badly written comedy |
| 20:03 | <Hixie> | sicking: dude your bug about hidden="" and aria-* is getting me in trouble |
| 20:04 | <Ms2ger> | Hixie, would've been nice if you hadn't hijacked my obsolete specs thread, btw ;) |
| 20:15 | <sicking> | Hixie: i won't have time to spend on that stuff unfortunately :( |
| 20:16 | <sicking> | Hixie: also, allowing aria-* to point to hidden stuff is the right thing to do, so i don't understand why it's trouble? |
| 20:47 | <heycam> | jamesr_, there's no easy way to do that currently :( |
| 20:47 | <heycam> | jamesr_, (the gradients along the polyline) |
| 20:47 | <heycam> | jamesr_, the new gradient meshes we'll have should make it possible, though still not simple to write |
| 21:18 | AryehGregor | writes http://aryeh.name/tmp/css-test/contributors/aryehgregor/incoming/viewer.html as a crude makeshift reftest-checker, in about 20 minutes |
| 21:18 | <AryehGregor> | Needs some refinement, I guess. |
| 21:18 | <AryehGregor> | But I'm not going to submit reftests if I can't easily check whether they pass in implementations . . . |
| 21:20 | <gsnedders> | Write a runner using WebDriver! |
| 21:24 | <Hixie> | sicking: ditto on both counts |
| 21:24 | <Hixie> | sicking: like i said before you joined, i feel like i am in some sort of badly written comedy |
| 21:34 | <smaug____> | is anne on vacation ? |
| 21:47 | <Hixie> | smaug____: yes |
| 21:48 | <jamesr_> | heycam, looks like the closest i can get today is to build up the path in lots of little overlapping pieces (2-3 segments) and color them individually |
| 21:48 | <jamesr_> | to get something kinda close |
| 21:49 | <heycam> | jamesr_, yeah |
| 21:49 | <heycam> | jamesr_, btw did you see the mail about rAF |
| 21:49 | <jamesr_> | about last call? |
| 21:49 | <heycam> | jamesr_, yeah |
| 21:49 | <jamesr_> | yeah i saw it, haven't had time to look in detail |
| 21:49 | <jamesr_> | gotta remember what the LC reqs are |
| 21:49 | heycam | has not been paying attention to the spec at all |
| 21:49 | <heycam> | just that all the issues are closed, really |
| 21:50 | <jamesr_> | well that's easy enough to accomplish :) |
| 21:50 | <jamesr_> | biggest problem with rAF spec-wise is the lack of a test suite, imo |
| 21:51 | <heycam> | yeah, technically that's not required unless you want to transition from CR to the next one up |
| 21:51 | <heycam> | but the earlier the better obviously |
| 21:51 | <heycam> | anyway if all the issues we had are actually closed, we can say go for it for LC |
| 21:51 | <jamesr_> | will have to check |
| 21:51 | <heycam> | I might fiddle with the IDL though first to take account of recent webidl changes |
| 21:51 | <jamesr_> | i think there are some issues of the form "think about doing BLAH" |
| 21:52 | <heycam> | ok |
| 21:52 | <jamesr_> | and they can probably be closed with either "thought about it, don't want to do it" or "let's think about it later" |
| 21:52 | <heycam> | yeah |
| 21:52 | <heycam> | it's probably fine to leave those called out in the spec, if you wanted to attract attention to them |
| 21:53 | <jamesr_> | as notes? |
| 21:53 | <heycam> | yeah |
| 21:53 | <jamesr_> | sg |
| 21:53 | <jamesr_> | i'll try to go through open issues and then reply to that email tomorrow |
| 21:54 | <jamesr_> | today is a mess for me |
| 21:54 | <heycam> | doubt there's a hurry |
