| 00:06 | <zewt> | gar why is it rel=noreferrer and not rel=noreferer, heh |
| 00:07 | <gsnedders> | :) |
| 00:07 | <zewt> | bad enough to have had to teach myself to type "referer" without now having it be 50/50 |
| 00:27 | <TabAtkins> | zewt: Oh my lord, *seriously*? That's the worst thing. I mean, referer is evil for being misspelled, but we're not helping anyone by correcting it here. >_< |
| 00:27 | <zewt> | webkit sends onclick to <input type=radio> after onchange? |
| 00:27 | <zewt> | that seems impossible enough to make me think I'm doing something stupid, heh |
| 00:28 | <jamesr_> | think of the children! |
| 00:28 | <jamesr_> | zewt, if you preventDefault() the onclick, does it go back in time and un-fire the onchange? |
| 00:28 | <zewt> | no black holes reported yet |
| 00:28 | <zewt> | i get the onchange first, with the checked value changed, then cancelling the onclick restores the old checked (without firing another onchange) |
| 00:29 | <jamesr_> | oh my god! it's gone plaid^Wopera! |
| 00:29 | <zewt> | (chrome, specifically) |
| 00:29 | <zewt> | i definitely feel like giving it the raspberry |
| 00:29 | <jamesr_> | zewt, that doesn't sound right |
| 00:29 | <zewt> | indeed |
| 00:29 | <jamesr_> | does it happen for any other checkbox-y types? |
| 00:30 | <zewt> | i'll check |
| 00:30 | <zewt> | (and put my test up for stupid-typo checking) |
| 00:30 | <jamesr_> | yeah and if you can file on bugs.webkit.org that'd be great |
| 00:33 | <zewt> | happens for checkbox too |
| 00:35 | <zewt> | https://zewt.org/~glenn/test-webkit-radio-click-cancel.html |
| 00:35 | <zewt> | logs very differently in chrome and FF |
| 00:36 | <zewt> | (... though it logs .checked == true in both, so maybe I'm doing at least something wrong) |
| 00:36 | <zewt> | (might just be .checked being the value it'd be after the click during onclick, which is weird but okay, but the event order is definitely being weird) |
| 00:57 | <benjoffe> | from reading the spec it looks like an 'output' element cannot have a 'list' attribute pointing to a datalist |
| 00:57 | <benjoffe> | this sound correct? if so it's a shame, cause there's what I think is a legitimate use-case for it |
| 01:02 | <hober> | what's the use case? |
| 01:16 | <zewt> | jamesr_: want to give that a quick sanity-check? |
| 01:17 | <jamesr_> | zewt, yeah looks legit |
| 01:17 | <zewt> | bizarre |
| 01:18 | <zewt> | seems like that should be breaking all kinds of things |
| 01:18 | <jamesr_> | maybe some of the common JS libraries have hacks around this area |
| 01:20 | <zewt> | what the hell |
| 01:21 | <zewt> | i download the webkit nightly in chrome, and it's putting up this big scary "WebKit-SVN-r125178.zip is not commonly downloaded and could be dangerous" nonsense |
| 01:28 | <benjoffe> | hober: if the output shall be textual, the one I'm thinking of is for a client side password strength measurer, the output can have values either 'Too short' 'Weak' 'Strong' etc. |
| 01:31 | <zewt> | (yeah I see why .checked is true during onclick in the spec) |
| 01:31 | <imakewebthings> | can anyone point me to a spec, if it exists, that says how a browser should restore scroll position on page reload? i'm specifically interested in what events should fire but anything would be good |
| 01:37 | <Hixie> | imakewebthings: http://whatwg.org/html is part of that loop |
| 01:39 | <zewt> | jamesr_: guessing this is intentional since IE does it too (ugh) suggesting webkit copied IE, but https://bugs.webkit.org/show_bug.cgi?id=93674 |
| 01:39 | <zewt> | http://www.webkit.org/quality/reporting.html is "search bugzilla" supposed to be a joke? heh |
| 01:40 | <zewt> | did a couple searches which came back with 600 results, tried googling and like w3c it apparently blocks spiders, then gave up |
| 01:42 | <imakewebthings> | Hixie: Thanks, I suppose I should start digging in around 6.6.2 and see what I can determine from there |
| 02:23 | <Hixie> | imakewebthings: looks like it's at the bottom of the "update the session history with the new page" algorithm |
| 02:26 | <Hixie> | does anyone have an up to date implementation of the outline algorithm? |
| 03:17 | <imakewebthings> | Hixie: Thanks, I'm reading it as being left to user agents exactly what state to restore and how to go about it. The test case for why I ask: http://imakewebthings.com/sandbox/scroll-reload/ |
| 03:18 | <Hixie> | too tired to understand that right now :-) |
| 03:19 | <Hixie> | if you want the spec changed, send e-mail and explain why it's important that every UA do the same thing |
| 03:29 | <imakewebthings> | no problem, thanks. |
| 06:26 | <jgraham> | tobie: Thanks. As gsnedders said we try to report bugs that end up being site-patched, but of course for huge sites having things break for even a short amount of time is bad, so I expect a few slip through the process |
| 07:36 | <Ms2ger> | Whoa |
| 07:37 | <Ms2ger> | http://www.whatwg.org/issues/data.html?period=1 < that was a lot of respimg emails |
| 07:54 | <jgraham> | Ms2ger: No, really? |
| 07:55 | <Ms2ger> | I blissfully ignored them :) |
| 07:57 | <jgraham> | "Estimated date for last e-mail based on the data above: 1999-08-08" |
| 07:57 | <jgraham> | (on the 3 years of data) |
| 11:02 | <krijn> | O hai, nobody is missing the logs huh? :) |
| 11:02 | <Ms2ger> | I am |
| 11:03 | <krijn> | The globbot thing looks handier |
| 11:04 | <Ms2ger> | It doesn't have highlighting |
| 11:04 | <krijn> | We've finally moved, so I need to setup my server in the new office thingy |
| 11:04 | <krijn> | Hopefully somewhere next week |
| 11:04 | <Ms2ger> | Great :) |
| 11:06 | <Ms2ger> | odinho, fwiw, I'm having someone write tests for blob.slice |
| 11:06 | <odinho> | Ms2ger: yay someone |
| 11:09 | <Stevef_> | krijn: i have been missing them! |
| 11:09 | <krijn> | They will be up again! :) |
| 11:10 | <krijn> | If I can't get my server up and running again, I'll point my site to a different one and make a static copy |
| 11:12 | <Ms2ger> | odinho, also, can has a build with constructor support? :) |
| 11:15 | <odinho> | Ms2ger: Maybe I can just send you my binary here-here? :P |
| 11:16 | <Ms2ger> | Sure :) |
| 11:22 | <odinho> | Ms2ger: Ah, it's in ci-337, so desktop have it in the 12.50 build. Crazy-old, but there's a huge gap between desktop 12 and "us". |
| 11:24 | <Ms2ger> | 12.50 internal / build 1497 doesn't seem to have it |
| 11:25 | <odinho> | http://snapshot.opera.com/unix/looooooooooong_12.50-1538/ << 1538 seems to be the number |
| 11:25 | <odinho> | wtf up with looooooooong |
| 11:26 | <darobin> | maybe it supports longdesc? |
| 11:26 | <odinho> | yeah, the ultimate marketable feature ;D |
| 11:27 | <darobin> | since when has Opera been about marketable features? |
| 11:28 | <odinho> | If we put it in the url it has to be very important, does it not? |
| 11:29 | <darobin> | nah, we all know URLs are opaque and all |
| 11:29 | <odinho> | Guess G+ taught us that |
| 11:29 | <darobin> | heh |
| 11:50 | <MikeSmith> | wow krijn is alive |
| 11:50 | <krijn> | Of course! |
| 11:51 | <MikeSmith> | krijn: I thought you had retired to the cote d'azur, like Batman |
| 11:53 | <krijn> | Actually I did |
| 11:53 | <Ms2ger> | odinho, quite some test failures, still :) |
| 11:53 | <krijn> | Thought nobody noticed |
| 11:53 | <MikeSmith> | heh |
| 11:53 | <MikeSmith> | I'm glad you decided to come back and fight more crime |
| 11:53 | <odinho> | Ms2ger: (whisper)we have a bug called "fix ms2gers' blob test failures" |
| 11:54 | <Ms2ger> | :) |
| 11:54 | <odinho> | But blob guy is on a long deserved vacation. |
| 11:55 | <krijn> | http://krijn.qontent.nl/server.jpg should have some vacation as well |
| 11:55 | <odinho> | krijn: lol, 1995 called, it wants its computer back. |
| 11:55 | <krijn> | (Note says: "Do not turn off, the internet relies on this machine! Even though no monitor or keyboard is attached, this thing is working!") |
| 11:56 | <MikeSmith> | krijn: that explains things. you're powering it from your radiator |
| 11:56 | <krijn> | I'm glad they ignored the note |
| 11:56 | <MikeSmith> | krijn, btw, I will be in Amsterdam next month for a few days. Planning to meet up with Anne and whoever feels like getting together for foods and beers or whatnot |
| 11:56 | <krijn> | odinho: :p |
| 11:56 | <krijn> | MikeSmith: oh! when? |
| 11:57 | <MikeSmith> | around 13-14-15 I think |
| 11:57 | jgraham | always assumed that the logs were running on a souped-up toaster |
| 11:57 | <krijn> | I'll be in Amsterdam then as well |
| 11:57 | <MikeSmith> | cool |
| 11:57 | <krijn> | MikeSmith: if Anne allows me to join, I'd be happy to :) |
| 11:57 | <MikeSmith> | I'll talk to him once he's back |
| 12:00 | <niloy> | Why isnt any other browser implementing setImmediate? |
| 12:01 | <krijn> | MikeSmith: oki, cool, ping me when you know more :) |
| 12:01 | <krijn> | <-- gone now o/ |
| 12:01 | <Ms2ger> | Is that Microsoft's thing? |
| 12:01 | <niloy> | yes |
| 12:01 | <jgraham> | Isn't that the thing that's precisely the same as setTimeout(f, 0) |
| 12:02 | <niloy> | yes, so is setTimeout the recommended thing? |
| 12:02 | <Ms2ger> | It was a silly thing, that much I recall |
| 12:02 | <niloy> | so does the browser automatically treat setTImeout(f, 0) to be executed in the next event loop? |
| 12:03 | <jgraham> | Well I suppose it doesn't force it to the front of the event queue |
| 12:03 | <jgraham> | But if setImmediate is supposed to do that it seems very wrong |
| 12:04 | <niloy> | nodeJS has nextTict(), the browser needs something like that too |
| 12:04 | <jgraham> | e.g. setImmediate(f1); setImmediate(f2) would lead to f2 being run before f1 |
| 12:04 | <niloy> | s/nextTict/nextTick/ |
| 12:04 | <Ms2ger> | Huh |
| 12:04 | <Ms2ger> | Opera seems to restore tabs when hitting ctrl+w |
| 12:04 | <Ms2ger> | Is that intentional? |
| 12:05 | <niloy> | nodeJS has nextTick which accepts a callback which gets executed in the next event loop |
| 12:06 | <jgraham> | It has the behaviour I said above? |
| 12:07 | <niloy> | Umm... am not sure about that, but will it break something horribly? |
| 12:07 | <Ms2ger> | Sounds like a good way to hang your browser |
| 12:07 | <niloy> | how so? 2 events will be executed one after another right? |
| 12:08 | <niloy> | but you guys know better, so I will take your word on it |
| 12:08 | <jgraham> | niloy: In a browser, not processing e.g. user input events is bad |
| 12:08 | <jgraham> | node.js doesn't really have that problem |
| 12:09 | <jgraham> | So if people start running code that jumped the event queue it could lead to badness |
| 12:10 | <jgraham> | IOW setTimeout(f, 0) seems like a better idea in a browser |
| 12:13 | <niloy> | alright, thanks |
| 12:14 | <smaug____> | requestAnimationFrame(f) is possibly even better than setTimeout(f, 0) |
| 12:15 | <smaug____> | depending on what you're going |
| 12:15 | <smaug____> | er |
| 12:15 | <smaug____> | doing |
| 12:17 | <jgraham> | Yes, if you need to sync with painting |
| 12:27 | <niloy> | I am attaching around 1000 nodes to DOM which is freezing the UI |
| 12:28 | <jgraham> | Testcase? |
| 12:30 | <niloy> | umm, I dont have it prepared, its happening with the application I am developing at my job |
| 12:30 | <niloy> | but I guess I can make something quick on jsfiddle |
| 12:30 | <jgraham> | Well if you can provide a simple TC that reproduces the behaviour, it will be easy to help. If you can't it won't :) |
| 12:31 | <jgraham> | (fwiw the Live DOM viewer is preferred comnpared to js-fiddle in these parts) |
| 12:31 | <jgraham> | (because it is simpler) |
| 12:32 | niloy | is embarrased, doesnt know what Live DOM viewer is |
| 12:32 | <jgraham> | It's OK, I was just looking up the URL |
| 12:32 | <jgraham> | http://software.hixie.ch/utilities/js/live-dom-viewer/ |
| 12:32 | <niloy> | thanks |
| 12:33 | <jgraham> | It's not really popular in the wider web community, but it has a helpful level of simplicity |
| 12:33 | <niloy> | cool, I will check it out |
| 12:33 | <jgraham> | For example it doesn't encourage the use of jQuery-for-everything |
| 12:33 | <niloy> | hehe |
| 12:34 | <jgraham> | (10 line testcases with 10,000 lines of library code are 10,010 line testcases ;) |
| 12:34 | <niloy> | :) |
| 12:36 | <niloy> | jgraham, umm... where do I write JS? |
| 12:36 | <Ms2ger> | <script>// JS</script> |
| 12:36 | <niloy> | okay, thanks |
| 12:44 | <niloy> | I tried creating 1000 divs, it happened quiet instantly |
| 12:45 | <darobin> | niloy: FWIW one way of getting something rather close to nextTick is to use postMessage |
| 12:45 | <niloy> | I should postMessage to window? |
| 12:45 | <darobin> | setTimeout has a minimal delay that the browser enforces, postMessage doesn't (yet) |
| 12:45 | <darobin> | yeah, you postMessage to yourself, comes with the next event loop |
| 12:46 | <niloy> | Oh, thats great |
| 12:46 | <niloy> | thank you |
| 12:46 | <darobin> | at least, that used to work, it may now be defended against — let me dig it up |
| 12:46 | <darobin> | here http://dbaron.org/log/20100309-faster-timeouts |
| 12:47 | <niloy> | I am creating a tree component, so before every node inserting, I am perform some tree traversing |
| 12:47 | <niloy> | so the UI hangs even with 1000 nodes |
| 12:47 | <niloy> | I am not able to replicate with a simple for loop with 1000 div insertion |
| 12:47 | <darobin> | niloy: here's an example https://github.com/substack/node-browserify/blob/master/builtins/__browserify_process.js |
| 12:48 | <niloy> | darobin, thanks a lot |
| 12:48 | <darobin> | np |
| 12:49 | <jgraham> | In theory setTimeout only has a delay when called recursively |
| 12:49 | <darobin> | jgraham: I think it has a delay when you call it in a loop, no? |
| 12:49 | <darobin> | but recursively would be a problem here anyway |
| 12:50 | <jgraham> | Only recursively in the spec |
| 12:50 | <niloy> | I am currently calling setTimeout recursively, its solved the UI freeze problem |
| 12:51 | <niloy> | in each setTimeout call, I insert 1 node |
| 12:51 | <niloy> | and then call setTimeout again till I reach the end of array |
| 12:52 | <niloy> | It has fixed the UI freeze problem, but it feels very hackish |
| 12:52 | <darobin> | niloy: no, sometimes you have to do things like that to handle large numbers of synchronous ops |
| 12:52 | <darobin> | it's part of the platform |
| 12:52 | <darobin> | Node spoils you by enforcing it a lot of the time |
| 12:52 | <darobin> | but it's still what it does |
| 12:53 | <darobin> | if it feels too hackish, just hide it in a library ;-) |
| 12:53 | <niloy> | darobin, no no, dont get me wrong |
| 12:53 | <darobin> | actually I think that's what async.js does when in the browser |
| 12:54 | <niloy> | I just wish the browser would provide something a api natively, like setImmediate |
| 12:55 | <niloy> | I mean, technically animations can be done by setTimeout, but the browers came with requestAnimationFrame, which is amazing |
| 12:56 | <niloy> | so I was just wish browsers would give us something so execute stuff in a async manner without any minimum delay |
| 12:57 | <darobin> | niloy: yeah, I think that makes sense |
| 12:57 | <niloy> | thanks ^_^ |
| 12:57 | <darobin> | "execute this as soon as possible, async (but you're allowed to prioritise something else if it makes more sense)" |
| 12:57 | <darobin> | I haven't looked at setImmediate tbh |
| 12:57 | <niloy> | yes, exactly ^_^ |
| 12:58 | <darobin> | maybe we could get there if setTimeout can be implemented in a more helpful fashion though (I haven't thought this through at all atm) |
| 12:59 | <niloy> | yeh, if browers remove minimum delay from setTimeout, that would become equivalent, but I guess it has technical issues |
| 13:00 | <niloy> | since I am not a browser developer, I dont know :( |
| 13:00 | <darobin> | niloy: sometimes you don't want to let JS code just flood the event loop; it wouldn't work out |
| 13:01 | <darobin> | but one probably *could* resort to smarter mechanisms than just a delay |
| 13:01 | <niloy> | yeh |
| 13:51 | <AryehGregor> | Hmm, does Google+ translate " " into " "? It should be " " -- the way they have it now adds whitespace at the beginning of lines. |
| 13:52 | AryehGregor | sends feedback |
| 15:41 | <dglazkov> | good morning, Whatwg! |
| 15:45 | <tantek> | good morning dglazkov |
| 17:11 | <Hixie> | jesus, the linked data folk are now trying to get RDF into HTML through the back door with JSON-LD |
| 17:15 | <hober> | and you're surprised? |
| 17:17 | <Hixie> | their persistence is admirable |
| 17:17 | <Hixie> | also, julian is lying in ietf-types again to get his pet ideas through, sigh |
| 17:29 | <Hixie> | hey can anyone think of technologies that have tried to replace HTML recently other than .NET and XHTML2? |
| 17:29 | <Hixie> | i'm having a mind blank |
| 17:30 | <TabAtkins> | How recently? Flash and Silverlight made attempts. |
| 17:30 | <jarek> | uhm... Adobe Air |
| 17:30 | <jarek> | it was sort of a compromise between Flash and HTML |
| 17:30 | <Hixie> | ah yes, Flash and Air |
| 17:31 | <Hixie> | (Silverlight is .NET) |
| 17:31 | <jarek> | and JavaFX |
| 17:32 | <Hixie> | ah yes, of course, how could i forget java |
| 17:45 | <jgraham> | And native apps :) |
| 18:15 | <bfrohs> | If an document contains <input autofocus> and a textarea is added later via JavaScript with the autofocus attribute, which element should be focused? Chrome focuses the textarea while Firefox focuses the input. |
| 18:16 | <bfrohs> | To test: data:text/html,<!DOCTYPE html><form><input autofocus><script>el = document.createElement("textarea");el.setAttribute("autofocus", "");document.body.appendChild(el);</script></form> |
| 18:16 | <Hixie> | bfrohs: see the spec :-) |
| 18:16 | <Hixie> | ( http://www.whatwg.org/specs/web-apps/current-work/#autofocusing-a-form-control:-the-autofocus-attribute ) |
| 18:16 | <bfrohs> | Hixie: Yes, I'm there. |
| 18:17 | <bfrohs> | My thoughts are it should be the input, as #7 aborts the steps. Am I correct? |
| 18:18 | <Hixie> | yes |
| 18:18 | <bfrohs> | Thanks much! :) |
| 18:28 | <sawrubh> | is the HT(horizontal tab) mentioned in http://tools.ietf.org/html/rfc2616#page-16 8 or 4 spaces ? |
| 18:28 | <TabAtkins> | It appears to be a tab character. U+0009. |
| 18:29 | <sawrubh> | so what is that, 4 or 8 spaces, I need to use it in a token |
| 18:29 | <TabAtkins> | It's neither. It's a *tab character*. That's a different thing than a space. |
| 19:17 | <zewt> | was text-outline a thing that disappeared? it's something webvtt would want pretty badly (and seems to think still exists, since it's mentioned as a supported style) |
| 19:19 | <zewt> | can sort of fake it with text-shadow, but ... badly |
| 19:19 | <zewt> | particularly with serif fonts |
| 19:28 | <Hixie> | zewt: it's in some css spec or other |
| 19:30 | <zewt> | it's in http://www.w3.org/TR/2007/WD-css3-text-20070306/ but not the current ED, if it's moved somewhere else I havn't found it |
| 19:44 | <Hixie> | wtf opera |
| 19:44 | <Hixie> | 'cellIndex' in document.createElement("th") => true |
| 19:44 | <Hixie> | document.createElement("th").cellIndex => undefined |
| 19:44 | <Hixie> | how is that even possible |
| 19:44 | <tantek> | maybe it depends on which site you're pretending to be? |
| 19:44 | <tantek> | some per-site quirkiness? |
| 19:44 | <gavinc> | property exists, has no value? |
| 19:45 | <Hixie> | if the property exists, the getter should return a value of the property's type |
| 19:45 | <Hixie> | i don't understand how you even write code that can end up with the getter returning undefined |
| 19:45 | <Hixie> | i mean, you have to actually work to do that, as far as i can tell |
| 19:46 | <Hixie> | given that C++ doesn't have "undefined" as a value for "long" |
| 19:46 | <Ms2ger> | Ah, you got there :) |
| 19:46 | <Ms2ger> | Apparently Opera doesn't do dom bindings |
| 19:46 | <Hixie> | Ms2ger: thanks for the test case and results, btw, makes my life WAY easier |
| 19:47 | <Ms2ger> | I try :) |
| 19:47 | <Hixie> | i haven't tested IE for myself |
| 19:47 | <Hixie> | i'm scared to given what you reported as the result |
| 19:47 | <Hixie> | ("semi-random numbers") |
| 19:48 | <Ms2ger> | It might be that's only IE10 |
| 19:48 | <Ms2ger> | And I have no idea if there is any newer IE10 snapshot than the one that I managed to get my hands on somehow |
| 19:48 | <Hixie> | testing on IE is a huge pain for me (have to spin up a vm in a faraway cluster, etc) |
| 19:49 | <Hixie> | but anyway, doesn't matter here |
| 19:49 | <Hixie> | since there's clearly no interop! |
| 19:49 | <Ms2ger> | If only they released IE for Linux... |
| 19:49 | <Ms2ger> | For the long ignored spec editors' market |
| 19:50 | <Hixie> | actually i'm on mac currently |
| 19:50 | <Ms2ger> | IE5.5 for mac, then? :) |
| 19:50 | <Hixie> | probably not useful :-P |
| 19:53 | Ms2ger | files on other browsers |
| 20:08 | <Yuhong> | I am thinking of a protocol on top of WebSockets that with the user's permission, allow to read and write files on the user's hard drive. |
| 20:09 | <Yuhong> | The idea is that the browser would allow the user to select files just like a file upload, but instead of uploading the file, return a token. |
| 20:09 | <Yuhong> | That can be later used to tell the browser to read or write the file. |
| 20:10 | <zewt> | ... File? |
| 20:10 | <Yuhong> | What do you mean? |
| 20:11 | <Yuhong> | The idea to allow a Google Docs like to be created that do not store the files on the server. |
| 20:11 | <zewt> | file api gives a token (File) to read files (and filesystem API allows writing files, though that's moving slowly) |
| 20:11 | <zewt> | (not sure what WebSockets has to do with files) |
| 20:12 | <Yuhong> | Did not realize such an API already exist. |
| 20:13 | <Yuhong> | Did not realize such an API already exist. |
| 20:15 | <Ms2ger> | Bye |
| 20:15 | <zewt> | ragequit? |
| 20:15 | <zewt> | "relax, have some dip" |
| 20:15 | <Ms2ger> | It's an interesting figure, I have to admit |
| 20:15 | <Yuhong> | Ms2ger: Win8 just RTMed. |
| 20:16 | <Ms2ger> | I see |
| 20:16 | <zewt> | does that mean "crashed" |
| 20:16 | <Yuhong> | Release to Manufacturing. |
| 20:17 | <Yuhong> | It will behttp://arstechnica.com/information-technology/2012/08/windows-8-released-to-manufacturing-with-build-9200/ |
| 20:17 | <Yuhong> | http://arstechnica.com/information-technology/2012/08/windows-8-released-to-manufacturing-with-build-9200/ |
| 20:17 | <Yuhong> | Has the key dates for when you will be able to get it. |
| 20:17 | <zewt> | does anybody not hate win8? heh |
| 20:18 | <zewt> | it's like they went "well, people liked XP, hated Vista, liked win7, so it's time to do something horrible again!" |
| 20:23 | <Yuhong> | Found the File API, thank you. |
| 20:24 | <Yuhong> | And File API: Writer. |
| 20:41 | <zewt> | when did google decide it was okay to just ignore search terms entirely, thus becoming worthless |
| 20:42 | <zewt> | haha http://wiki.whatwg.org/wiki/Timed_tracks#Architecture i wonder who made this illustration and thought, "there, that clarifies everything!" |
| 21:06 | <Hixie> | zewt: that was me, but iirc it wasn't intended to clarify things but more to illustrate what a mess it was :-) |
| 21:26 | <smaug____> | what does chromium bug status Icebox mean? |
| 21:27 | smaug____ | noticed a valid bug he had filed long ago got that status |
| 21:28 | <jgraham> | Hixie: Looks like a rather straighforward bug. |
| 21:42 | <jamesr_> | smaug____, it's a bot |
| 21:42 | <jamesr_> | smaug____, if it did something bad link me the bug and i can fix it |
| 21:44 | <smaug____> | jamesr_: http://code.google.com/p/chromium/issues/detail?id=75770 I don't have chromium to verify if that is still valid |
| 21:44 | <smaug____> | oh, and the test might need update to not report the very last progress event |
| 21:51 | <zewt> | bleh I hate hitting obvious bugs that every browser has, because hope of getting them fixed drops through the floor |
| 21:51 | <zewt> | re: cancelling mousedown preventing mouse thresholding from being applied to clicks (in IE, WebKit *and* Firefox) |
| 21:52 | <zewt> | (not opera) |
| 22:02 | <zewt> | Hixie: at least it's not a flowchart :) |
| 22:08 | <Hixie> | jgraham: sure, it's just the implications regarding the infrastructure that i am worried about :-) |
| 22:30 | <TabAtkins> | Hixie: In the CSS that establishes quoting, what's the reasoning for doing ":root:lang(X), :not(:lang(X)) > :lang(X)"? It seems like this is equivalent to just writing ":lang(X)". |
| 22:32 | <TabAtkins> | Possible answer: it avoids overriding author's rules placed somewhere in the middle of a tree of same-lang elements. |
| 22:35 | <TabAtkins> | Hixie: Actually, that's almost certainly the correct answer. So never mind the question, I answered it myself. |
| 23:13 | <jamesr_> | smaug____, i can't tell that chrome's behavior is non-conformant here (because the spec text is super fuzzy) |
| 23:13 | <jamesr_> | what does "about every 50ms" mean? |
| 23:14 | <jamesr_> | is it supposed to fire progress even if no bytes were transmitted? |
| 23:15 | <smaug____> | jamesr_: nope. when I filed that bug chrome was firing events way more often than that |
| 23:15 | <smaug____> | events shouldn't happen more often that 50ms |
| 23:15 | <smaug____> | except the last one right before loadend |
| 23:15 | <smaug____> | s/often that/often than every/ |
| 23:18 | <jamesr_> | more often than "about 50ms" |
| 23:18 | <jamesr_> | what's the rationale for that limitation? |
| 23:18 | <smaug____> | to not fire event like every ms |
| 23:18 | <jamesr_> | i'd guess naively that it's firing based on feedback from the network stack |
| 23:19 | <smaug____> | yeah, gecko used to have that behavior too |
| 23:19 | <jamesr_> | and what platform are you testing on? |
| 23:19 | <smaug____> | that is an old bug :) |
| 23:20 | <jamesr_> | i'm getting 50/51 on your test page on 22.0.1221.1 linux |
| 23:20 | <smaug____> | ok, if you get that, then ok |
| 23:20 | <smaug____> | perhaps it is then fixed |
| 23:20 | <jamesr_> | i think we fixed it: http://code.google.com/searchframe#OAMlx_jo-ck/src/third_party/WebKit/Source/WebCore/xml/XMLHttpRequestProgressEventThrottle.h&exact_package=chromium&q=xmlhttprequest%2050&ct=rc&cd=4&sq= |
| 23:20 | <smaug____> | k |
| 23:20 | <smaug____> | thanks then :) |
| 23:21 | <smaug____> | (the bug should be marked then fixed) |
| 23:21 | <jamesr_> | yup, did that |
| 23:21 | smaug____ | doesn't have chromium installed because it is always so hard to find the nightly .zip files |
| 23:22 | <jamesr_> | what platform are ya on? on windows/mac the canary channel is easy |
| 23:23 | <smaug____> | linux |
| 23:23 | <jamesr_> | i use dev channel normally (of course i normally have a fresh ToT build i produced myself) |
| 23:23 | <jamesr_> | dev's ~weekly which is not too bad |
| 23:24 | <smaug____> | I just never find the .zip file when I try to find it |
| 23:24 | <smaug____> | and once I find it, I end up using the same build for weeks or months because I forgot where I downloaded it :) |
| 23:24 | <smaug____> | (and because the .zip doesn't auto-update) |
| 23:25 | <jamesr_> | it's linux man, just use the .deb |
| 23:25 | <smaug____> | well, not .zip, but tar.gz or whatever |
| 23:25 | <smaug____> | just simple package |
| 23:26 | <smaug____> | which I can unpack |
| 23:26 | <smaug____> | though, I think it is .zip in chromium case |
| 23:26 | <jamesr_> | sure, but why? |
| 23:27 | <smaug____> | last time I used .rpm or some such, it polluted my update process |
| 23:27 | <smaug____> | I don't want to get notified when there is update to chromium |
| 23:28 | <jamesr_> | it's kinda broken to get notified about any package update, it should just install |
| 23:28 | <zewt> | heh firefox is trying to copy chrome's autoupdate, but still does it badly |
| 23:28 | <smaug____> | possibly I'm just too used to Firefox nightlies |
| 23:28 | <zewt> | re: middle of development, i load firefox to test, and it stops for fifteen seconds with an "installing update" while I sit there losing my train of thought |
| 23:29 | <smaug____> | zewt: on linux Firefox's autoupdate is way nicer |
| 23:29 | <smaug____> | than chromiums |
| 23:29 | <zewt> | i never notice chrome's updating in windows, which is the only possible way autoupdate is okay |
| 23:30 | <smaug____> | that is the Firefox update works on linux, and should work on windows too |
| 23:32 | <jamesr_> | it's a bit ironic that out of win/mac/linux, linux is the system with the oldest and most deeply integrated system update mechanism |
| 23:32 | <jamesr_> | but it's the worst updating experience of any chromium |
| 23:32 | <jamesr_> | or chrome, rather |
| 23:33 | <zewt> | heh, linux *had* to, since binary compatibility is so bad in linux |
| 23:33 | <jamesr_> | YOTLD |
| 23:33 | <zewt> | so distributing prebuilt packages is a massive pain |
| 23:33 | <jamesr_> | for instance i'll wager if if major distros ever actually do the gtk2->gtk3 switch, a lot of apps will statically link gtk2 in |
| 23:34 | <zewt> | i've always had to save VMs of linux build systems for building binary patches later, since otherwise it's often a complete nightmare of working around library, runtime linker, etc. incompatibilities |
| 23:34 | <zewt> | and this is on systems where i controlled the whole target system |
| 23:35 | <zewt> | vs. windows where you can run most stuff built for win95 in win7, heh |
| 23:37 | <zewt> | (granted not necesarily the reverse) |
