| 04:16 | <zewt> | we need a comprehensive list of all the times people have been confused by TR |
| 04:17 | <zewt> | we need a comprehensive list of all the times people have been confused by TR |
| 04:17 | <Hixie> | be easier to make a list of times they weren't |
| 04:17 | <Hixie> | :-P |
| 04:17 | <Hixie> | be easier to make a list of times they weren't |
| 04:17 | <Hixie> | :-P |
| 09:45 | <annevk> | oh that's quite a silly bug with respect to qualified name |
| 09:45 | <annevk> | oh that's quite a silly bug with respect to qualified name |
| 09:47 | <Ms2ger> | zewt, http://wiki.whatwg.org/wiki/TR_strikes_again :) |
| 09:47 | <Ms2ger> | zewt, http://wiki.whatwg.org/wiki/TR_strikes_again :) |
| 09:48 | <annevk> | I knew we had a page |
| 09:48 | <annevk> | couldn't find it on Google |
| 09:48 | <annevk> | I knew we had a page |
| 09:48 | <annevk> | couldn't find it on Google |
| 09:49 | <webben> | when w3c publishes a spec without any versioned permalink like http://www.w3.org/WAI/PF/aria-implementation/ |
| 09:49 | <webben> | is there some way to link to that particular revision? |
| 09:50 | <webben> | when w3c publishes a spec without any versioned permalink like http://www.w3.org/WAI/PF/aria-implementation/ |
| 09:50 | <webben> | is there some way to link to that particular revision? |
| 09:50 | <annevk> | the W3C webmaster is not involved in that |
| 09:51 | <annevk> | the W3C webmaster is not involved in that |
| 09:51 | <annevk> | that's up to the individual maintaining those pages |
| 09:51 | <webben> | oh well i suspected as much |
| 09:51 | <annevk> | it should be in CVS though |
| 09:51 | <annevk> | that's up to the individual maintaining those pages |
| 09:51 | <webben> | oh well i suspected as much |
| 09:51 | <annevk> | it should be in CVS though |
| 09:52 | <webben> | that would do ... that's where i looked first ... maybe i looked in the wrong place |
| 09:52 | <webben> | that would do ... that's where i looked first ... maybe i looked in the wrong place |
| 09:53 | <webben> | why are there folders for years in w3c cvs? |
| 09:53 | <webben> | why are there folders for years in w3c cvs? |
| 09:54 | <annevk> | because that's how the W3C used to organize things |
| 09:54 | <annevk> | e.g. http://www.w3.org/1999/xhtml |
| 09:55 | <annevk> | because that's how the W3C used to organize things |
| 09:55 | <annevk> | e.g. http://www.w3.org/1999/xhtml |
| 09:55 | <annevk> | (which redirects, hey ho for copy & paste) |
| 09:55 | <annevk> | (which redirects, hey ho for copy & paste) |
| 10:00 | <annevk> | my whole twitter feed talks about some sqoot (whatever that is) screw up but no details... |
| 10:00 | <annevk> | anyone? |
| 10:00 | <annevk> | my whole twitter feed talks about some sqoot (whatever that is) screw up but no details... |
| 10:00 | <annevk> | anyone? |
| 10:00 | <izhak> | Hi, guys. The "emit the current input character" claim often can be meet in tokenizer specification, but I just cannot find what to do in case when no characters were consumed yet? |
| 10:01 | <izhak> | Hi, guys. The "emit the current input character" claim often can be meet in tokenizer specification, but I just cannot find what to do in case when no characters were consumed yet? |
| 10:01 | <mattwest> | They made a sexist joke in one of there event ads and people got angry with them for it. |
| 10:01 | <mattwest> | They made a sexist joke in one of there event ads and people got angry with them for it. |
| 10:01 | <annevk> | ta |
| 10:01 | <mattwest> | Wasn't as bad as everyone made it out to be but some of their big sponsors pulled out. |
| 10:02 | <annevk> | ta |
| 10:02 | <mattwest> | Wasn't as bad as everyone made it out to be but some of their big sponsors pulled out. |
| 10:02 | <annevk> | izhak: there's always a current input character... |
| 10:02 | <annevk> | izhak: there's always a current input character... |
| 10:03 | <izhak> | annevk: can you point me an item that defines that |
| 10:04 | <izhak> | annevk: can you point me an item that defines that |
| 10:04 | <izhak> | annevk: ah.! yes. |
| 10:04 | <izhak> | As it always I'm stupid:) |
| 10:04 | <izhak> | annevk: ah.! yes. |
| 10:04 | <izhak> | As it always I'm stupid:) |
| 10:05 | <izhak> | annevk: Just was considering a case when 0x0000 was consumed. (And unconciously thought it's not a character :) ) |
| 10:06 | <izhak> | annevk: Just was considering a case when 0x0000 was consumed. (And unconciously thought it's not a character :) ) |
| 10:07 | <annevk> | ah, the only special thing is EOF |
| 10:07 | <annevk> | ah, the only special thing is EOF |
| 10:14 | <annevk> | zcorpan: go go Quirks Mode |
| 10:14 | <annevk> | zcorpan: go go Quirks Mode |
| 10:15 | <annevk> | so I wonder if we rename limited quirks mode and no-quirks mode as suggested, how will HTML be changed? |
| 10:15 | <annevk> | did Hixie veto the renaming? |
| 10:15 | <annevk> | so I wonder if we rename limited quirks mode and no-quirks mode as suggested, how will HTML be changed? |
| 10:15 | <annevk> | did Hixie veto the renaming? |
| 10:16 | Ms2ger | is no fan of renaming stuff |
| 10:17 | Ms2ger | is no fan of renaming stuff |
| 10:18 | <annevk> | ok lets wontfix that bug then |
| 10:18 | <annevk> | ok lets wontfix that bug then |
| 10:25 | <annevk> | does anyone know if there are bugs on browsers for non-ASCII case-insensitive matching of attribute and element names in Selectors and DOM? |
| 10:25 | <annevk> | Gecko does the correct thing btw |
| 10:25 | <annevk> | does anyone know if there are bugs on browsers for non-ASCII case-insensitive matching of attribute and element names in Selectors and DOM? |
| 10:25 | <annevk> | Gecko does the correct thing btw |
| 10:25 | <annevk> | since Firefox 4 reportedly |
| 10:26 | <annevk> | since Firefox 4 reportedly |
| 10:26 | <zcorpan> | i might have filed bugs on opera for this |
| 10:26 | <zcorpan> | i might have filed bugs on opera for this |
| 10:32 | <annevk> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1307 testcase by Leif Halvard Silli |
| 10:32 | <annevk> | http://software.hixie.ch/utilities/js/live-dom-viewer/saved/1307 testcase by Leif Halvard Silli |
| 10:35 | <annevk> | I filed https://bugs.webkit.org/show_bug.cgi?id=81752 on WebKit |
| 10:36 | <annevk> | I filed https://bugs.webkit.org/show_bug.cgi?id=81752 on WebKit |
| 10:36 | <annevk> | and added TabAtkins to the cc list for good measure |
| 10:36 | <annevk> | guess I'll file an equivalent bug for DOM |
| 10:36 | <annevk> | and added TabAtkins to the cc list for good measure |
| 10:36 | <annevk> | guess I'll file an equivalent bug for DOM |
| 10:38 | <annevk> | DOM bug is https://bugs.webkit.org/show_bug.cgi?id=81753 |
| 10:38 | <annevk> | DOM bug is https://bugs.webkit.org/show_bug.cgi?id=81753 |
| 10:42 | <Ms2ger> | Bah, looks like we're stuck with the global scope polluter :( |
| 10:42 | <Ms2ger> | Bah, looks like we're stuck with the global scope polluter :( |
| 10:46 | <annevk> | which one? |
| 10:47 | <annevk> | which one? |
| 10:52 | <Workmon> | TabAtkins: Assuming the decision to disallow/ignore !important in @style holds, it seems there would be no way to override an !important author style |
| 10:52 | <Workmon> | TabAtkins: Assuming the decision to disallow/ignore !important in @style holds, it seems there would be no way to override an !important author style |
| 10:52 | <Ms2ger> | That's the way it should be |
| 10:53 | <Ms2ger> | That's the way it should be |
| 10:53 | <Workmon> | author, not user |
| 10:53 | <annevk> | you can use !important in style |
| 10:54 | <Workmon> | author, not user |
| 10:54 | <annevk> | you can use !important in style |
| 10:54 | <annevk> | e.g. ele.style.setProperty("background", "blue", "!important") |
| 10:54 | <annevk> | e.g. ele.style.setProperty("background", "blue", "!important") |
| 10:54 | <annevk> | it's just not clear whether it should also work from .style.background = "blue !important" because arguably that string only sets the second argument |
| 10:55 | <annevk> | it's just not clear whether it should also work from .style.background = "blue !important" because arguably that string only sets the second argument |
| 10:55 | <Workmon> | I don't see why .style.background = x should be different from style="background: x" |
| 10:55 | <Workmon> | I don't see why .style.background = x should be different from style="background: x" |
| 10:56 | <annevk> | because it maps to setProperty and that API does not map straight |
| 10:57 | <annevk> | because it maps to setProperty and that API does not map straight |
| 10:58 | <Workmon> | Is that an already implemented mapping, or just the proposed way to specify it? |
| 10:59 | <Workmon> | Is that an already implemented mapping, or just the proposed way to specify it? |
| 11:00 | <annevk> | that's been the API since before my time |
| 11:00 | <annevk> | that's been the API since before my time |
| 11:08 | <annevk> | ooh |
| 11:08 | <annevk> | geo v2 is stopped |
| 11:09 | <annevk> | ooh |
| 11:09 | <annevk> | geo v2 is stopped |
| 11:09 | <annevk> | reverse geocoding as browser utility was kind of weird I guess |
| 11:09 | <annevk> | reverse geocoding as browser utility was kind of weird I guess |
| 11:10 | <mattwest> | yer I guess but would of come in handy sometimes |
| 11:10 | <mattwest> | yer I guess but would of come in handy sometimes |
| 11:10 | <annevk> | there's web services that do it |
| 11:11 | <annevk> | there's web services that do it |
| 11:11 | <mattwest> | yerp you're right |
| 11:12 | <mattwest> | yerp you're right |
| 11:13 | <mattwest> | annevk: who would have been responsible for mainatining the db? |
| 11:14 | <mattwest> | annevk: who would have been responsible for mainatining the db? |
| 11:14 | <annevk> | presumably a third-party or the vendor itself |
| 11:15 | <annevk> | presumably a third-party or the vendor itself |
| 11:16 | <annevk> | that's already the case for geolocation; most browsers use a web service to transmit e.g. SSIDs towards to get the location |
| 11:16 | <mattwest> | ah I see |
| 11:16 | <mattwest> | ah I see |
| 11:17 | <mattwest> | thanks |
| 11:17 | <mattwest> | thanks |
| 11:19 | <annevk> | zewt: wouldn't it be nicer to have some kind of decoder/encoder object that you pass strings/bytes as you get them in over the network and it gives the decoded stuff thus far |
| 11:20 | <annevk> | zewt: wouldn't it be nicer to have some kind of decoder/encoder object that you pass strings/bytes as you get them in over the network and it gives the decoded stuff thus far |
| 11:20 | <annevk> | and you don't have to worry about state at all |
| 11:20 | <annevk> | and you don't have to worry about state at all |
| 11:21 | <annevk> | e.g. enc = new Decoder("utf-8"); string = enc.feed(bytes) |
| 11:21 | <annevk> | and you just invoke .feed() each time you get some more stuff from the network and the object keeps track of state etc. |
| 11:21 | <annevk> | e.g. enc = new Decoder("utf-8"); string = enc.feed(bytes) |
| 11:21 | <annevk> | and you just invoke .feed() each time you get some more stuff from the network and the object keeps track of state etc. |
| 11:48 | <hsivonen> | what's the deal with canvas path objects having an a11y impact now? |
| 11:49 | <hsivonen> | what's the deal with canvas path objects having an a11y impact now? |
| 11:51 | <annevk> | they're tauted as a better way of doing the hit testing stuff Microsoft proposed, except the relevant bit (I think Hixie calls it hit regions) is not in the spec yet |
| 11:51 | <annevk> | they're tauted as a better way of doing the hit testing stuff Microsoft proposed, except the relevant bit (I think Hixie calls it hit regions) is not in the spec yet |
| 11:53 | <hsivonen> | annevk: so one would keep path objects around for later hit testing? |
| 11:54 | <hsivonen> | annevk: so one would keep path objects around for later hit testing? |
| 11:55 | <annevk> | http://wiki.whatwg.org/wiki/Canvas#Regions |
| 11:55 | <annevk> | http://wiki.whatwg.org/wiki/Canvas#Regions |
| 12:10 | <annevk> | do people still want new Range(); new NodeIterator() and such? |
| 12:10 | <annevk> | do people still want new Range(); new NodeIterator() and such? |
| 12:11 | <annevk> | they're quite easy to add, but I wonder if there's no demand or if people simply gave up trying to get those kind of improvements |
| 12:11 | <annevk> | they're quite easy to add, but I wonder if there's no demand or if people simply gave up trying to get those kind of improvements |
| 13:19 | <jkl444> | sup |
| 13:19 | <jkl444> | I need your help guys |
| 13:19 | <jkl444> | looking for the html5 developers guide on a single html file |
| 13:19 | <jkl444> | to import to my kindle |
| 13:19 | <jkl444> | sup |
| 13:19 | <jkl444> | I need your help guys |
| 13:19 | <jkl444> | looking for the html5 developers guide on a single html file |
| 13:19 | <jkl444> | to import to my kindle |
| 13:21 | <jkl444> | Hi Evanescence, espadrine |
| 13:21 | <jkl444> | Can you perhaps help me? |
| 13:22 | <jkl444> | Hi Evanescence, espadrine |
| 13:22 | <jkl444> | Can you perhaps help me? |
| 13:25 | <zcorpan> | what's the html5 developers guide? |
| 13:25 | <espadrine> | jkl444: what's your issue? |
| 13:26 | <zcorpan> | what's the html5 developers guide? |
| 13:26 | <espadrine> | jkl444: what's your issue? |
| 13:27 | <jkl444> | I want to convert the guide to my Kindle as one continous document |
| 13:27 | <jkl444> | I want to convert the guide to my Kindle as one continous document |
| 13:27 | <jkl444> | At the moment I'll have to convert each section |
| 13:28 | <jkl444> | At the moment I'll have to convert each section |
| 13:28 | <jkl444> | Is there a single unified HTML of the developers guide? |
| 13:28 | <zewt> | annevk: i thought about that; if your input buffer is much larger than your output buffer, that means the decoder object wants to make a copy of most of the data, though |
| 13:28 | <jkl444> | Is there a single unified HTML of the developers guide? |
| 13:28 | <zewt> | annevk: i thought about that; if your input buffer is much larger than your output buffer, that means the decoder object wants to make a copy of most of the data, though |
| 13:28 | <jkl444> | zcorpan: http://developers.whatwg.org/ |
| 13:29 | <jkl444> | zcorpan: http://developers.whatwg.org/ |
| 13:30 | <hsivonen> | jkl444: I suggest downloading the chapters of http://diveintohtml5.info/ and copying and pasting them together if you need everything in one file |
| 13:30 | <hsivonen> | jkl444: I suggest downloading the chapters of http://diveintohtml5.info/ and copying and pasting them together if you need everything in one file |
| 13:30 | <annevk> | zewt: hmm yeah, if you have fixed-size buffers the API will need to be different |
| 13:30 | <espadrine> | jkl444: I know the tools used to generate that website are at https://github.com/benschwarz/developers.whatwg.org |
| 13:30 | <annevk> | zewt: hmm yeah, if you have fixed-size buffers the API will need to be different |
| 13:30 | <espadrine> | jkl444: I know the tools used to generate that website are at https://github.com/benschwarz/developers.whatwg.org |
| 13:30 | <jkl444> | hsivonen: alright I'll try that. espadrine: I'll need to learn how to use python for that :) rather merge some HTML's |
| 13:31 | <jkl444> | hsivonen: alright I'll try that. espadrine: I'll need to learn how to use python for that :) rather merge some HTML's |
| 13:32 | <zewt> | annevk: fixed-size buffers are probably the main use case for streaming |
| 13:32 | <zewt> | eg. a 10MB input string and 16k output packets |
| 13:32 | <zewt> | annevk: fixed-size buffers are probably the main use case for streaming |
| 13:32 | <zewt> | eg. a 10MB input string and 16k output packets |
| 13:32 | <zewt> | though most of the time strings aren't gigantic, so it's okay to convert the 10MB input string to a 10MB-or-so output ArrayBuffer, then just split the result |
| 13:33 | <zewt> | though most of the time strings aren't gigantic, so it's okay to convert the 10MB input string to a 10MB-or-so output ArrayBuffer, then just split the result |
| 13:41 | <annevk> | zewt: for streaming I was mostly thinking bytes as input and strings as output and fixed-size input does not really matter |
| 13:41 | <annevk> | zewt: for streaming I was mostly thinking bytes as input and strings as output and fixed-size input does not really matter |
| 13:43 | <zewt> | the idea of supporting all legacy encodings here is insane and I can't believe anyone is even suggesting it |
| 13:43 | <zewt> | the idea of supporting all legacy encodings here is insane and I can't believe anyone is even suggesting it |
| 13:44 | <zewt> | "let's get rid of legacy encodings! and let's add new support for them while we're at it, just to make that even harder" |
| 13:44 | <zewt> | "let's get rid of legacy encodings! and let's add new support for them while we're at it, just to make that even harder" |
| 13:45 | <annevk> | we're not going to get rid of legacy encodings |
| 13:46 | <annevk> | we're not going to get rid of legacy encodings |
| 13:46 | <zewt> | doesn't mean we should make it easier to use them |
| 13:46 | <zewt> | especially for the uglier encodings that actively restrict API options (ISO-2022) |
| 13:46 | <zewt> | doesn't mean we should make it easier to use them |
| 13:46 | <zewt> | especially for the uglier encodings that actively restrict API options (ISO-2022) |
| 13:46 | <annevk> | if we want encodings to work consistently it seems better to expose them consistently too |
| 13:47 | <annevk> | if we want encodings to work consistently it seems better to expose them consistently too |
| 13:47 | <zewt> | blocking the nastier encodings from new features is pretty much the same as blocking new features from sync XHR |
| 13:47 | <jgraham> | If there are some legacy encodings that won't work with the API we want that might be a good reason not to have them |
| 13:47 | <zewt> | blocking the nastier encodings from new features is pretty much the same as blocking new features from sync XHR |
| 13:47 | <jgraham> | If there are some legacy encodings that won't work with the API we want that might be a good reason not to have them |
| 13:47 | <annevk> | yeah sure |
| 13:48 | <annevk> | yeah sure |
| 13:48 | <jgraham> | But in general it seems like there should just be a fixed list of encodings for the platform |
| 13:48 | <jgraham> | But in general it seems like there should just be a fixed list of encodings for the platform |
| 13:50 | <annevk> | letting authors deal with leftover bytes seems wrong though, that should be the decoders job |
| 13:50 | <zewt> | let it be a fixed list without things like ISO-2022 (and other gross ones, like that ibm-whatever encoding that remaps something in the ASCII range), and let those be exceptional addititions where they're needed |
| 13:50 | <annevk> | and once it handles leftover bytes |
| 13:50 | <annevk> | it can handle the other stuff too |
| 13:50 | <annevk> | there's nothing in the encoding standard that remaps ASCII afaik |
| 13:50 | <annevk> | letting authors deal with leftover bytes seems wrong though, that should be the decoders job |
| 13:50 | <zewt> | let it be a fixed list without things like ISO-2022 (and other gross ones, like that ibm-whatever encoding that remaps something in the ASCII range), and let those be exceptional addititions where they're needed |
| 13:50 | <annevk> | and once it handles leftover bytes |
| 13:50 | <annevk> | it can handle the other stuff too |
| 13:50 | <annevk> | there's nothing in the encoding standard that remaps ASCII afaik |
| 13:50 | <zewt> | there was at some point |
| 13:51 | <zewt> | there was at some point |
| 13:51 | <annevk> | well, ESC I guess is still special in some encodings |
| 13:51 | <annevk> | well, ESC I guess is still special in some encodings |
| 13:52 | <annevk> | btw, iso-2022-jp only needs a single additional flag, iso-2022-kr has no additional state |
| 13:52 | <annevk> | per spec |
| 13:52 | <annevk> | well it has a state of course doh |
| 13:52 | <annevk> | btw, iso-2022-jp only needs a single additional flag, iso-2022-kr has no additional state |
| 13:52 | <annevk> | per spec |
| 13:52 | <annevk> | well it has a state of course doh |
| 13:52 | <annevk> | but not an encoding switch |
| 13:53 | <annevk> | but not an encoding switch |
| 13:54 | <zewt> | ibm864 had weirdness, at least in some browser, i think |
| 13:54 | <zewt> | anyway need to go to work |
| 13:55 | <zewt> | ibm864 had weirdness, at least in some browser, i think |
| 13:55 | <zewt> | anyway need to go to work |
| 13:55 | <annevk> | ah yeah Gecko has something weird there |
| 13:55 | <annevk> | ah yeah Gecko has something weird there |
| 15:45 | <aklein> | annevk: good morning. re your question yesterday about transient observers, they do need to be added everywhere you remove the children of the DocumentFragment (to each child of the fragment if the fragment has subtree observers) |
| 15:49 | <smaug____> | aklein: IIRC DOM4 has also a bit wrong definition for transient observers |
| 15:49 | <smaug____> | the subtree part is missing... |
| 15:50 | <smaug____> | could be also that I just haven't found that in the spec |
| 15:51 | <aklein> | smaug____: not sure what you mean, http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#concept-node-remove step 7 only adds a transient observer if the original one is a subtree observer |
| 15:51 | <smaug____> | ah, ok it is there |
| 15:51 | <smaug____> | I had missed that |
| 15:52 | <aklein> | it would be nice if those DocumentFragment special cases could be condensed somehow |
| 15:52 | <aklein> | but I suppose that can be left for later, once we've got the correct semantics in place |
| 15:56 | <nesta_> | :) |
| 15:58 | <smaug____> | aklein: we should create a testsuite for mutationobservers |
| 15:59 | <smaug____> | take rafaelw_'s awesome tests, and de-googlefy them, add also my tests (and de-mochitestify them) and write some more |
| 16:00 | <aklein> | smaug____: and there's also WebKit tests that can be de-webkitified |
| 16:00 | <smaug____> | testsuite should probably live somewhere under w3.org |
| 16:00 | <Ms2ger> | Go ahead and put them on dvcs.w3.org/hg/webapps/DOMCore |
| 16:00 | <zcorpan> | w3c-test.org ? |
| 16:02 | <aklein> | smaug____: are you thinking of being systematic about it, or just throwing everything we have up there? |
| 16:02 | <Ms2ger> | Let's start with throwing up what you have :) |
| 16:03 | <smaug____> | yeah, that would be good start :) |
| 16:03 | <aklein> | heh, ok. I only ask because I was fighting with some hilariously broken W3C dom/xhtml tests yesterday. |
| 16:03 | <aklein> | (unrelatedly) |
| 16:04 | <smaug____> | oh, yes, tests are often broken, like everything else in the web :p |
| 16:05 | <Ms2ger> | Which? |
| 16:05 | <smaug____> | Ms2ger: tests may test specs, which are wrong :) |
| 16:06 | <Ms2ger> | Definitely |
| 16:06 | <aklein> | Ms2ger: can't figure out where the originals are now, but the one's I was fighting with are now stored here: http://trac.webkit.org/browser/trunk/LayoutTests/dom/xhtml/level3/core/ |
| 16:07 | <Ms2ger> | Oh, I think I remember those |
| 16:07 | <aklein> | most of them actually work |
| 16:08 | <Ms2ger> | Now, it would be nice if you could get them into the DOM4 suite :) |
| 16:08 | <aklein> | but at least 20 of them fail to run the JS containing their test code! |
| 16:08 | <aklein> | hehe |
| 16:09 | <Ms2ger> | If you ever can't sleep and want to do something boring, yet useful... ;) |
| 16:10 | <smaug____> | writing tests for DOM4 can be quite effective sleeping pill |
| 16:11 | <smaug____> | hmm, who can change the topic of this channel |
| 16:12 | <Ms2ger> | You |
| 16:12 | <smaug____> | Ms2ger: um, how did you do that |
| 16:12 | <smaug____> | anyway, does anyone object if I add http://logbot.glob.com.au/?c=freenode%23whatwg to the topic |
| 16:13 | <Ms2ger> | Asked chanserv nicely |
| 16:13 | <smaug____> | as an alternative log |
| 16:13 | <Ms2ger> | Go for it |
| 16:13 | <jgraham> | I don't object, but why? |
| 16:14 | <jgraham> | Also, dammit we had had that /topic for like 2 years or something |
| 16:14 | <Ms2ger> | Mon Oct 19 23:03:06 2009 |
| 16:14 | <smaug____> | jgraham: krijn mentioned at some point that he may not be able to keep up logging |
| 16:15 | <jgraham> | smaug____: I think that was sorted out |
| 16:15 | <smaug____> | oh, that information never reached me |
| 16:18 | <Velmont> | I just write the tests testharness.js from the start, -- so there should be no need to deoperatize it. |
| 16:19 | <Ms2ger> | I should look at my review comments for making that possible here |
| 16:19 | <Velmont> | Meh, too little honey in this hot milk. -- Had too much last time, so guess I overcompensated. |
| 16:21 | <Ms2ger> | Got a cold? |
| 16:21 | <Velmont> | Yup. Or something similar. Don't feel too well at least :-) |
| 16:22 | <Velmont> | Good that there's lots of free milk and honey in the cantina. Don't have that at home, so better at work :] |
| 16:22 | <jgraham> | Hmm, I wonder if we have honey here... that would be a really good idea |
| 16:22 | <Velmont> | jgraham: You should move office to Oslo, just sayin'. :] |
| 16:23 | <jgraham> | BTW I have a feature request for allowing "manual" testing from testharness.js. Would basically be for tests that have a few interaction steps but can still decide if they pass or not |
| 16:23 | <jgraham> | automatically |
| 16:23 | <Ms2ger> | I've got that supported with an annotation in the manifest :) |
| 16:23 | <jgraham> | I was thinking of something like setup({"requires_interaction":true}) that would disable the timeout |
| 16:23 | <jgraham> | Dunno if that's a good idea |
| 16:24 | <Velmont> | jgraham: I've also done interactive testharness.js test. -- I needed to upload a picture for it to work. Was maybe not the best fit, but ohwell. If all you know is a hammmer and all :P |
| 16:25 | <jgraham> | Velmont: Umm, no. Apart from the possible lack of honey, the linköping office is nicer. Also Oslo makes Sweden look cheap. And I think it's even colder there! |
| 16:25 | <Velmont> | jgraham: Actually, as I said before, I would not be against spending some time in swedish Opera offices. |
| 16:25 | <Velmont> | http://w3c-test.org/webappsec/tests/cors/submitted/opera/interactive/file-preflight.htm << there, interactive. |
| 16:26 | <Velmont> | Doesn't actually work because the PHP files is not put out on w3c-test.org though. |
| 16:27 | <jgraham> | (don't get me wrong the Oslo office is nice, it's just that this one is nicer :) |
| 16:27 | <jgraham> | +, |
| 17:03 | <Ms2ger> | "editing: with great responsibility comes no power" |
| 17:15 | <Velmont> | Ms2ger: Any idea how to detect that browser is doing new style DOMExceptions? |
| 17:15 | <Velmont> | *best detect |
| 17:17 | <Ms2ger> | try { document.body.removeChild(document) } catch (e) { // e.name }? |
| 17:18 | <Velmont> | Ms2ger: Okay, -- was thinking about doing something that throws for it, -- but found it a bit heavy handed. But I guess it's the safest way to do it :-) |
| 17:18 | <Ms2ger> | Yeah |
| 17:19 | <Velmont> | hrfm, -- there's that opera bug yeah, -- name = DOMException, message = NOT_FOUND_ERR |
| 17:21 | <jgraham> | (ugh) |
| 17:21 | <jgraham> | I sort of object to making the test harness have that in |
| 17:22 | <annevk> | whoa topic change? |
| 17:22 | <annevk> | radical |
| 17:22 | <Velmont> | jgraham: OK. -- I have problems actually writing that at all, in a consice and nice name. |
| 17:24 | <annevk> | aklein: yeah, the DocumentFragment stuff could be factored out |
| 17:24 | <Velmont> | annevk: I'm trying to have testharness.js get the DOM4 DOMExceptions. |
| 17:24 | <Velmont> | annevk: https://github.com/jgraham/testharness.js/pull/1 < you might have seen that. |
| 17:24 | <annevk> | Velmont: yay you |
| 17:24 | <Velmont> | jgraham: I could need a tips for how you'd like to see it done, if you know. :-) |
| 17:24 | <jgraham> | I don't really |
| 17:25 | <jgraham> | But I would prefer that we avoid having the correct behaviour of one function depend on correctness in an unrelated part of DOM |
| 17:26 | <jgraham> | i.e. assert_throws shouldn't suddenly have funky behaviour if you break document.body.removeChild(document) somehow |
| 17:26 | <annevk> | aklein: as for the other thing, I guess I'll order some food and then study transient observers for a bit |
| 17:26 | <Velmont> | jgraham: Hm. Maybe test for new-style first. Then ALSO do a test for the old one if first one fails? |
| 17:26 | <Velmont> | And I need a mapping (NETWORK_ERR, NetworkError, 17) tuple for that I guess. |
| 17:27 | <jgraham> | You mean on the actual exception that is thrown by the code under test? |
| 17:27 | <jgraham> | That sounds more reasonable |
| 17:27 | <Velmont> | Okay. |
| 17:28 | <Ms2ger> | That would allow throwing both old-style and new-style in the same implementation? |
| 17:28 | <jgraham> | Another feature testharness.js doesn't have is boolean logic for exceptions. That makes me sad |
| 17:28 | <jgraham> | Ms2ger: yeah, but I doubt anyone will actually do that |
| 17:28 | <Ms2ger> | I guess |
| 17:28 | <jgraham> | and it's not really clear why it's worse than some implementations doing one thing and some doing another |
| 17:29 | <Velmont> | Well, -- it's possible for us to do it, -- and depending on a few things that might happen a short while. -- Haven't really talked/thinked much about it yet. |
| 17:30 | <jgraham> | For who to do what? |
| 17:30 | <Velmont> | jgraham: Opera |
| 17:30 | <jgraham> | To do DOM4 style exceptions? |
| 17:30 | <annevk> | I think the API should just be assert_throws("SyntaxError", ...) |
| 17:30 | <Velmont> | jgraham: Throw newstyle exceptions for e.g. idb, and oldstyle for e.g. xhr. |
| 17:30 | <annevk> | and then testharness.js knows the .code for SyntaxError |
| 17:30 | Ms2ger | doesn't want to change all the tests |
| 17:31 | <Velmont> | annevk: So Fix all the tests(!!11)? :| |
| 17:31 | <annevk> | we can keep the old API around too, and support the reverse mapping for those |
| 17:31 | <jgraham> | Velmont: When I looked at the code it seemed easier to do it all at once. I could be wtong. I Am Not A C++ Expert, This Is Not Coding Advice |
| 17:31 | <gsnedders> | Velmont: Really you want to be in the Göteborg office: as nice as the Linköping office, but a less dull city. |
| 17:31 | <Ms2ger> | I'd look at the code for a second opinion, but you won't let me :( |
| 17:32 | <Velmont> | gsnedders: Hehe, okay :] Might do both though. It's all in the same country anyways. |
| 17:33 | <gsnedders> | (That said, I'm in Oslo for the first half of next week) |
| 17:33 | <jgraham> | (me too :() |
| 17:34 | <jgraham> | (not because of Oslo, but because of the train journey) |
| 17:34 | gsnedders | is flying via LHR, which will be fun |
| 17:34 | <jgraham> | Nothing wrong with that |
| 17:34 | <jgraham> | I am going via Katrineholm, which I claim is worse |
| 17:34 | <gsnedders> | Eh, I have too many experiences of losing luggage doing transfers at LHR. |
| 17:35 | <gsnedders> | At least changing at Katrineholm you shouldn't lose a suitcase. |
| 17:35 | <Velmont> | Oh, nice. -- What's the occasion? Some meeting I can crash? :P |
| 17:35 | <jgraham> | Well that could happen |
| 17:35 | <jgraham> | But at least you aren't likely to die of boredom |
| 17:35 | <gsnedders> | (Literally the only place I have ever lost a suitcase is going through LHR… and it's happened twice.) |
| 17:35 | <gsnedders> | (So I'm kinda bitter.) |
| 17:36 | <gsnedders> | Velmont: Easter holidays from uni is my excuse :P |
| 17:36 | <Ms2ger> | A good time to prepare for your exams? |
| 17:36 | <annevk> | jgraham: what's next week? |
| 17:36 | <annevk> | jgraham: did I miss something? |
| 17:36 | <gsnedders> | Ms2ger: Indeed. |
| 17:36 | <Ms2ger> | That, or ES QA |
| 17:37 | <gsnedders> | Ms2ger: Got another three weeks before exams start, though, even if the two hardest ones are first. |
| 17:37 | <jgraham> | annevk: No, nothing special, just talking about our test system |
| 17:37 | <gsnedders> | (The fact that myself and jgraham coincide in Oslo is purely coincidence) |
| 17:38 | <annevk> | ah kk |
| 17:42 | <Velmont> | jgraham, Ms2ger: So... assert_throws, -- it does now (with the newstyle domexceptions) more or less check the incoming code against ex.type. That actually resembles normal exceptions quite well. |
| 17:43 | <Velmont> | jgraham: But is it just stupid widening it to other exceptions as well? Instead of requiring assert_throws({ type: "TypeError" }) |
| 17:43 | <Velmont> | Or I guess that'd be mixing stuff, so not good. |
| 17:44 | <jgraham> | Can't you do something like assert_throws(TypeError()m func)? |
| 17:44 | <jgraham> | s/m,/ |
| 17:44 | <jgraham> | oh, something |
| 18:00 | <annevk> | aklein: sorry got distracted by "improving the W3C process" email |
| 18:00 | <annevk> | aklein: might defer it to tomorrow |
| 18:26 | <Velmont> | OK, jgraham, Ms2ger, second try: https://github.com/jgraham/testharness.js/pull/1 |
| 18:33 | <rafaelw_> | smaug____: you mentioned you found the bug in the fuzzer yesterday? |
| 18:34 | <smaug____> | rafaelw_: oh, not in the fuffer |
| 18:34 | <smaug____> | er |
| 18:34 | <smaug____> | fuzzer |
| 18:34 | <smaug____> | in my implementation |
| 18:34 | <rafaelw_> | ah. that's cool. what was the bug? |
| 18:35 | <smaug____> | I wasn't using the "listenens all attribute changes" flag always in transient observers |
| 18:36 | <smaug____> | rafaelw_: now my implementation passes both test.html (when change to handle attributes case-sensitively) and testTreeMirror.html |
| 18:36 | <rafaelw_> | excellent! |
| 18:37 | <rafaelw_> | i'm excited to get two compatible implementations out there. |
| 18:37 | <rafaelw_> | i'm going to tear out case-insensitivity today in webkit. |
| 18:37 | <smaug____> | rafaelw_: did you see those two problems I found in webkit |
| 18:37 | <smaug____> | 'this' handling and documentfragment handling |
| 18:37 | rafaelw_ | looks at his email |
| 18:38 | <rafaelw_> | adam is gonna handle 'this'. |
| 18:38 | <smaug____> | I would assume those are easy to fix |
| 18:39 | <smaug____> | rafaelw_: so I suggested earlier today that we should put our tests to some generic testsuite |
| 18:39 | <smaug____> | w3c testsuite |
| 18:39 | <rafaelw_> | sounds like a good idea to me. |
| 18:39 | <smaug____> | rafaelw_: is it hard to make your stuff to not depend on google script libraries |
| 18:39 | <Ms2ger> | You're all in the WebApps WG? |
| 18:40 | <rafaelw_> | you mean on the closure test harness? |
| 18:40 | <smaug____> | yeah |
| 18:40 | <rafaelw_> | do w3c testsuites use a standard test library? |
| 18:40 | <smaug____> | Ms2ger: do you know about testharness.js or whatever it is called? |
| 18:41 | <Ms2ger> | Yes |
| 18:41 | <Ms2ger> | Someone landed it in WebKit, IIRC |
| 18:41 | <Ms2ger> | http://w3c-test.org/resources/testharness.js |
| 18:43 | <aklein> | rniwa probably knows something about this, he's been poking at that stuff recently |
| 18:44 | <Hixie> | blimey, the number of possible conditions that addHitRegion() can throw an exception for is quite ludicrous |
| 18:44 | <rniwa> | aklein: right, we have testharness.js checked in |
| 18:44 | <rniwa> | aklein: in LayoutTests/resources |
| 18:45 | <Hixie> | CSS people, what do i reference for 'cursor'? CSS3UI or CSS2.1 or other? |
| 18:45 | <Ms2ger> | 2UI |
| 18:45 | <Ms2ger> | 3UI, even |
| 18:46 | <Hixie> | roger, thanks |
| 18:46 | <Ms2ger> | Np |
| 18:46 | <rafaelw_> | smaug___: I see the document fragment bug. Yes. this should be easy to fix. Good catch. |
| 18:47 | <Hixie> | anyone know off-hand if ARIA's role="" is case sensitive? |
| 18:49 | <Hixie> | lol wut. http://www.w3.org/TR/wai-aria/roles references the alt text section in HTML then says "We've asked the HTML5 WG to remove or reduce this section, so we may remove the reference to it from ARIA." |
| 18:50 | <Hixie> | but it references both "HTML" (meaning HTML4) and "HTML5"'s sections... |
| 18:50 | <Hixie> | anywho |
| 18:51 | <rafaelw_> | smaug___: aklein points out to me the documentfragment issue is different than i was thinking. he has a patch for it, but its not 'easy'. |
| 18:52 | <Hixie> | can anyone point me to a user agent conformance criteria for aria-controls? |
| 18:52 | <Hixie> | i'm having trouble working out what some of these states and properties actually do in practice |
| 18:54 | <aklein> | maybe medium :) |
| 19:00 | <annevk> | is the atomic operation thing the hard part or something related to observers? |
| 19:01 | <annevk> | hmm I guess those are sort of intertwined :) |
| 19:05 | <jgraham> | Hixie: Is that "anywho" as in "Abandon all hope anywho enter here"? |
| 19:09 | <Ms2ger> | Hixie, so |
| 19:09 | <Ms2ger> | http://www.whatwg.org/specs/web-apps/current-work/multipage/webstorage.html#the-storage-event |
| 19:09 | <rafaelw_> | cd .. |
| 19:09 | <Ms2ger> | For clear(), it wants key to be null, but it isn't nullable |
| 19:16 | <aklein> | annevk: the atomic operation thing |
| 19:16 | <Hixie> | jgraham: basically |
| 19:16 | Hixie | sends in a CCP for this disaster |
| 19:16 | <Ms2ger> | Hixie, filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=16469 |
| 19:17 | <aklein> | annevk: tricky to change without breaking assumptions elsewhere in WebKit |
| 19:22 | <Hixie> | Ms2ger: k |
| 19:23 | <Ms2ger> | Do you have an ETA for finishing the canvas stuff? :) |
| 19:23 | <Hixie> | no |
| 19:29 | <TabAtkins> | Workmon: From an author style sheet? No, the only way for an author to override an author-level !important is with a more specific author-level !important. |
| 19:29 | <TabAtkins> | Workmon: But user-level !important is higher than author-level !important. |
| 19:32 | <TabAtkins> | annevk: The sqoot thing was that in their advertisement for a code jam, they listed "women bringing you beer" as one of the perks. |
| 19:33 | <Workmon> | TabAtkins: The use case is someone with a fairly high specificity !important rule that they want to override for a single element |
| 19:33 | <Workmon> | But I guess calling setProperty is the way to go |
| 19:33 | <TabAtkins> | Workmon: The proper solution is "dont' use !important, you dummy". |
| 19:35 | <annevk> | aklein: I guess the good thing is that only concept-node-insert actually removes children from DocumentFragment |
| 19:36 | <annevk> | aklein: we could use concept-node-insert for the other stuff too potentially |
| 19:36 | <annevk> | aklein: the only problem might be MutationRecord order |
| 19:36 | <aklein> | annevk: sorta? then we end up in a world where MutationEvents can cause MutationObservers to just be wrong. but that's only a problem for implementors, not spec writers :) |
| 19:36 | <aklein> | yeah, I'm not sure about ordering |
| 19:37 | <aklein> | this stuff can get complicated |
| 19:38 | <annevk> | so I guess we can queue after the actual remove/insert operations |
| 19:38 | <annevk> | e.g. in http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#concept-node-replace-all |
| 19:39 | <annevk> | we drop 2-3 and let concept-node-insert handle that again |
| 19:39 | <annevk> | to make that work however, we need to move 4 to be after 6 |
| 19:39 | <annevk> | that also works for concept-node-replace |
| 19:39 | <annevk> | can you see anything wrong with that aklein? |
| 19:42 | <TabAtkins> | annevk: This Selectors bug. Shouldn't it be brought up in www-style? |
| 19:42 | <annevk> | TabAtkins: it's out of scope of Selectors; Selectors says it is up to HTML (well, "host language"), which in turn says it's ASCII case-insensitive |
| 19:43 | <aklein> | annevk: don't we still need 2 to get the list of nodes? |
| 19:43 | <TabAtkins> | Ah, kk. |
| 19:43 | <TabAtkins> | annevk: Any idea what other browsers do? |
| 19:43 | <annevk> | aklein: yes |
| 19:43 | <annevk> | TabAtkins: Gecko ASCII, everyone else Unicode |
| 19:43 | <TabAtkins> | annevk: Ok. |
| 19:44 | <annevk> | TabAtkins: I filed bugs on Opera too |
| 19:44 | <TabAtkins> | I'll argue against AP. |
| 19:44 | <annevk> | oh god |
| 19:44 | <annevk> | not sure I want to read that bugmail |
| 19:44 | <TabAtkins> | It's just him saying "this doesnt' make sense" |
| 19:45 | <annevk> | oh, I guess he's not familiar with how HTML works then |
| 19:45 | <TabAtkins> | Correct. |
| 19:45 | <annevk> | or at least its parser |
| 19:46 | <annevk> | "Event if WebKit disagrees with every other browser out there, it's still not clear whether we should change this." |
| 19:46 | <annevk> | kind of sad that this (or if spec says so, we should just violate the spec) seems to be his go to feedback these days |
| 19:46 | <aklein> | annevk: hmm, seems like you'd also have to make concept-node-insert do something funny for the suppress observers flag, since you now need to have it take care of queueing for the DocumentFragment removal |
| 19:46 | <annevk> | he used to make WebKit match the XHR spec |
| 19:47 | <annevk> | aklein: yeah it would simply not listen to the flag for that operation |
| 20:34 | <annevk> | AryehGregor: sounds good btw, though maybe add a comment in the source documenting the selected text thingie |
| 20:34 | <annevk> | AryehGregor: re https://www.w3.org/Bugs/Public/show_bug.cgi?id=13843 |
| 20:34 | <AryehGregor> | annevk, you mean a comment in the source giving my test-case? |
| 20:35 | <annevk> | yeah something like that or a pointer to the bug report |
| 20:35 | <AryehGregor> | Sure. |
| 20:36 | <annevk> | i'm off, mutation changes will likely happen tomorrow |
| 20:37 | <Ms2ger> | Night |
| 20:37 | <annevk> | nn |
| 20:46 | <hober> | yo. |
| 20:46 | <hober> | could some of you reply to http://lists.w3.org/Archives/Public/public-html/2012Feb/0275.html today? |
| 20:47 | <Ms2ger> | No |
| 20:48 | <hober> | Ms2ger: :) |
| 20:49 | <Ms2ger> | I'm not yet mad enough to write to public-html |
| 20:51 | <aklein> | annevk: for the this argument of MutationCallbacks, does the spec actually say that "this" is the MutationObserver? |
| 20:51 | <aklein> | oh |
| 20:51 | <aklein> | nm |
| 20:51 | <aklein> | annevk: misread the sentence |
| 20:52 | <jgraham> | hober: What Ms2ger said. Pretty sure we aren't falling over ourselves to implement that, but I guess you really want an answer from annevk |
| 20:52 | <hober> | Ms2ger: you are a smarter person than I am, it would seem |
| 20:52 | <hober> | jgraham: "a solicitation of UA implementor interest has gone unanswered for a month" is a good argument. "3 vendors claim they won't implement" is a better one. :) |
| 20:53 | <Ms2ger> | The chairs consider arguments now? |
| 20:53 | <hober> | heh |
| 20:54 | <Ms2ger> | Last time I checked, it was only FORMAL OBJECTIONs |
| 20:54 | <othermaciej> | actually it i more the opposite |
| 20:54 | <othermaciej> | so far, the only thing we have done with formal objections is record them |
| 20:55 | <jgraham> | hober: Yeah and I am very happy for annevk to provide a claim one way or the other :) |
| 20:56 | <othermaciej> | here is the list of all Formal Objections we ever got: <http://dev.w3.org/html5/status/formal-objection-status.html>; |
| 20:56 | <hober> | jgraham: me too! :) |
| 20:56 | <Ms2ger> | Anyway, I've given up on the HTMLWG a long time ago, so consider whatever you like :) |
| 22:03 | <TabAtkins> | I shouldn't be amused by my response to https://bugs.webkit.org/show_bug.cgi?id=81752 , but I am. |
| 22:03 | <Ms2ger> | SQUIRREL |
| 22:21 | <Hixie> | hey maths people. what do you call a (cartesian) shape that has no holes, and where at any point outside the shape you can go in two orthgonal directions without approaching the shape? |
| 22:21 | <Hixie> | i.e. a shape like a filled O or P, or like a T, but unlike an S or F |
| 22:22 | <Hixie> | is there some special term that describes that kind of shape? |
| 22:26 | <Hixie> | (i guess i mean Euclidean, not Cartesian) |
| 22:29 | <TabAtkins> | Um, convex? |
| 22:29 | <Philip`> | P isn't convex |
| 22:29 | <TabAtkins> | True. |
| 22:30 | <TabAtkins> | I doubt there's a word for this. |
| 22:30 | <TabAtkins> | Why would you possibly need it? |
| 22:30 | <TabAtkins> | For describing stroke or something? |
| 22:30 | <Hixie> | describing the characteristic of a bounding circumference |
| 22:30 | <Hixie> | that doesn't have to be rectangular |
| 22:32 | <Philip`> | Why would it apply to T differently to how it applies to F? |
| 22:32 | <Hixie> | because of the bit between the bars of the F |
| 22:33 | <Philip`> | How's that different to the bit between the horizontal and vertical bars of the T? |
| 22:33 | <Hixie> | well it's different in the way i described earlier... not sure what you mean beyond that |
| 22:34 | <Philip`> | I think I mean, what algorithm cares about the difference you described earlier? |
| 22:34 | <Philip`> | (rather than just caring about e.g. convex vs non-convex) |
| 22:35 | <Philip`> | (or holey vs unholey) |
| 22:35 | <Hixie> | yeah i dunno |
| 22:35 | <Hixie> | maybe i don't need this |
| 22:36 | <jgraham> | TabAtkins: It might be funny, but there are also some reasonable arguments for not requiring unicode case insensitivity in general e.g. it introduces locale-dependence |
| 22:36 | <hober> | any thoughts on this (zero edit) CP before I submit it? http://www.w3.org/html/wg/wiki/User:Eoconnor/ISSUE-195 |
| 22:36 | <zewt> | jgraham: no, case-insensitivity is normally locale-independent ("case folding") |
| 22:37 | <zewt> | case-insensitivity in unicode isn't the same as converting both strings to lowercase before comparing |
| 22:37 | <TabAtkins> | hober: Looks good. |
| 22:39 | <jgraham> | zewt: Hmm, OK |
| 22:40 | <hober> | TabAtkins: I want to include http://www.dreamwidth.org/userpic/1073989/1201616 in it but can't quite figure out how. :) |
| 22:41 | <TabAtkins> | heh |