| 00:09 | <AryehGregor> | So it seems like in Firefox, Selections aren't actually associated with a document, but rather with a window. |
| 00:09 | <AryehGregor> | This actually makes sense. |
| 00:10 | <dglazkov> | yeeehaa |
| 00:10 | <dglazkov> | sorry, wrong window. |
| 00:10 | <dglazkov> | but somehow still appropriate :) |
| 00:11 | <AryehGregor> | "yeeehaa" is always appropriate. |
| 00:13 | <dglazkov> | or Leeeroooy |
| 01:21 | <AryehGregor> | What unsung genius decided that Python needed its own (broken and poorly-designed) package management system instead of integrating with the standard ones everyone uses? |
| 01:21 | <AryehGregor> | sudo easy_install lxml has now not only failed repeatedly because non-Python dependencies weren't present, it then spent several minutes compiling something or other from source. |
| 01:21 | <AryehGregor> | Now I just hope it works. |
| 01:24 | <AryehGregor> | Wait a sec, does the DOM Range spec only compile using Ms2ger's forked version of Anolis? |
| 01:24 | <AryehGregor> | . . . |
| 01:24 | <AryehGregor> | Okay, I guess I'll ask him when he shows up. I'll call it a day and start writing tests if I don't figure out how to get the spec to work. |
| 01:25 | <Hixie> | can't go wrong writing tests ;-) |
| 01:25 | Hixie | has just tripped over a part of the ISSUE-129 change proposal that has _so many_ errors he doesn't know what to do about it |
| 01:26 | <AryehGregor> | Hixie, make your counter-proposal end something like this? http://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags/1732454#1732454 |
| 01:26 | <Hixie> | (e.g. accordngy to the a11y tf, apparently aria-checked="true" must be set on any checkbox whose _default_ state is checked!) |
| 01:27 | <Hixie> | AryehGregor: very tempted. Though the spec already has a zalgo reference. ;-) |
| 01:27 | <TabAtkins> | wut |
| 01:27 | <Hixie> | also they don't say it should be =true, they should say it should "match the state", without saying what that means |
| 01:28 | <AryehGregor> | Does anyone else here only reliably notice it's their birthday when all the vBulletin forums they're a member of send them a happy birthday e-mail? |
| 01:28 | <AryehGregor> | Because I just realized that's the case for me, which is sad. |
| 01:29 | <TabAtkins> | Where are you seeing that, Hixie? |
| 01:29 | <Philip`> | AryehGregor: Happy birthday |
| 01:29 | <Hixie> | also according to this CP you are _not allowed_ to ever set aria-multiselectable="false" on a <select> even if the <select> has no multiple="" attribute |
| 01:29 | Philip` | falls for AryehGregor's subtle ploy |
| 01:29 | <Hixie> | TabAtkins: the trove of errors that is the "WAI-ARIA to HTML attribute mapping" table |
| 01:29 | <AryehGregor> | :P |
| 01:30 | <TabAtkins> | Hixie: Yeah, I'm looking at the input[type=checkbox] section of the table right now. |
| 01:30 | <AryehGregor> | If I don't even remember it's my own birthday, I'm hardly going to fish for happy birthdays. I was aware it was sometime around here, just didn't think about the exact day . . . |
| 01:30 | <AryehGregor> | :P |
| 01:30 | <Hixie> | pretty much every row of that table has at least one mistake |
| 01:30 | <Hixie> | i'm not gonna list them all in the CCP, i'd be here forever |
| 01:31 | <TabAtkins> | Hixie: Are you looking at http://www.paciellogroup.com/blog/misc/HTML5/aria-html5-proposal.html ? |
| 01:32 | <TabAtkins> | Oh, I see, you're reading the lower table. |
| 01:33 | <Hixie> | yeah, sorry for being unclear |
| 01:33 | <TabAtkins> | Right, that lower table has several mistakes. |
| 01:34 | <AryehGregor> | I'd offer to help, but I know absolutely nothing about ARIA. |
| 01:34 | <AryehGregor> | Also, I have to go to bed and should stop procrastinating about that, so good night. |
| 01:50 | <jamesr_> | Hixie: you should reply that the change proposal is not accessible to people with brains |
| 01:57 | <TabAtkins> | Btw, I'll go ahead and write the Issue-130 CP next week or so. |
| 02:08 | <Hixie> | TabAtkins: cool. let me know if you need help, happy to help with that one. |
| 02:11 | <wirepair> | 2/win 3 |
| 02:11 | <wirepair> | ! |
| 02:11 | <TabAtkins> | No, I think it'd be pretty simple. (1) Tables for layout are still bad for many reasons [list reasons]. |
| 02:15 | <othermaciej> | note that "X is bad" is not necessarily sufficient to prove "it's better for X to be noncomforming" |
| 02:16 | <TabAtkins> | True, but it may be sufficient for "This one 'improvement' isn't enough to justify making them conforming." |
| 02:23 | <othermaciej> | you'd have to translate the "X is bad" into what effects occur when X is conforming vs. nonconforming |
| 02:23 | <othermaciej> | presumably there is some marginal effect on willingness of authors to use X, at least for authors who are aware of and care about non-machine-checkable criteria |
| 03:24 | karlcow | discovering rowspan="0" http://reference.sitepoint.com/html/th/rowspan |
| 03:24 | <karlcow> | weird |
| 03:25 | <karlcow> | "There is also a special value of "0", which should tell the browser to span the cell to the end of all rows inside the current rowgroup (that is, the thead, tfoot, or tbody)." |
| 03:25 | <karlcow> | "Browser support for rowspan is generally good, with one exception—rendering rowspan="0". Firefox and Opera are the only browsers tested that correctly span the cell to which this attribute is applied across all subsequent rows; the other browsers tested render the cell in the first row only." |
| 05:26 | <Hixie> | TabAtkins: the table thing has at least two other pretty important points to include: first, that no version of HTML has ever allowed layout tables in the first place, so it's not like we're making old valid code invalid, and second, that making tables-for-layout ok would undo literally a decade or more of advocacy, which would seriously harm the credibility of the effort with early adopters and advocates |
| 05:39 | <hsivonen> | Hixie: the TF hasn't yet quite figured out what problem it is trying to solve, but it is trying to come up with a list of use cases to consider |
| 05:41 | <hsivonen> | GPH-Zeke: the validator has been updated piecemeal and is always at or near the tip of the tree |
| 05:42 | <hsivonen> | MikeSmith: window.parseHtmlDocument(inputText, documentObject, callbackWhenDone, nullPlaceHolderForErrorCallback); |
| 05:43 | <MikeSmith> | hsivonen: thanks |
| 05:43 | <GPHemsley> | hsivonen: K, thanks. |
| 06:14 | <Hixie> | hsivonen: interesting |
| 06:14 | <Hixie> | hsivonen: is there a draft of those use cases anywhere? |
| 06:20 | <MikeSmith> | hsivonen: I've checked in patches for all the remaining validator bugs/enhancements I had open for now |
| 06:21 | <MikeSmith> | if you could please redeploy today or soon, I will also redeploy the HTML5 backends on all the W3C validator hosts so we can have those synced up |
| 06:33 | <hsivonen> | MikeSmith: I'll redeploy today |
| 06:33 | <MikeSmith> | cool |
| 06:33 | <MikeSmith> | thanks |
| 06:33 | <hsivonen> | Hixie: http://lists.w3.org/Archives/Public/public-html-xml/2010Dec/0064.html |
| 06:34 | <othermaciej> | I've been skimming public-html-xml archives today |
| 06:35 | <othermaciej> | it is hard to tell what the goal is there |
| 06:41 | <Hixie> | as far as i can tell, all the use cases listed in that e-mail have well-understood, well-established, and already-deployed solutions |
| 06:46 | <Hixie> | 1: what normal describes (there's no namespace problem); 2: use an XML parser (browsers have been doing this for years); 3: as he says, XHTML, or as Atom does, a CDATA block; 4: <script> is designed for this purpose now; 5: just a combination of 3 and 4, nothing special beyond that is needed. |
| 06:54 | <Hixie> | case 2 was explained further in http://lists.w3.org/Archives/Public/public-html-xml/2011Jan/0045.html; the solution is to transform the content into HTML using an XML parser, and HTML serialiser, and some logic |
| 07:28 | <othermaciej> | hsivonen: if you think the polyglot spec should have dire warnings, I suggest filing a bug to that effect (re http://lists.w3.org/Archives/Public/public-html-xml/2011Jan/0138.html ) |
| 07:29 | <Hixie> | only 14 bits remaining to write on http://wiki.whatwg.org/wiki/Change_Proposal_for_ISSUE-129 ... in case anyone is bored and looking for something to do :-D |
| 07:42 | <hsivonen> | othermaciej: OK |
| 08:33 | <hsivonen> | who do I need to login again and again on every login-requiring action on the W3C Bugzilla? |
| 08:33 | <hsivonen> | I have allowed cookies from w3.org and www.w3.org |
| 08:33 | <hsivonen> | doesn't happen on other bugzillas |
| 09:07 | <hsivonen> | othermaciej: I filed a couple of bugs about Polyglot Markup |
| 09:08 | <annevk> | AryehGregor, yeah, you need his version of Anolis |
| 09:09 | <annevk> | AryehGregor, the way I solved this was by just changing the source and let Ms2ger do the rest :) |
| 09:11 | <hsivonen> | hmm. 134 "hard" blockers for Firefox 4 still |
| 09:12 | <annevk> | I read something about a release end of February? |
| 09:14 | <hsivonen> | annevk: this http://groups.google.com/group/mozilla.dev.planning/msg/cb1c625584526e97 ? |
| 09:14 | <annevk> | I think so, it was a Dutch news article that linked to some newsgroup post |
| 09:21 | <zcorpan> | hsivonen: any parser bugs that are blocking? |
| 09:22 | <hsivonen> | zcorpan: the <figure> ISSUE is a "soft" blocker |
| 09:22 | <hsivonen> | zcorpan: no hard ones |
| 10:08 | <matjas> | are you kidding me? @gruber didn’t know YouTube has been serving WebM for months? |
| 10:10 | <hsivonen> | matjas: awesome, isn't it? |
| 10:11 | <matjas> | makes me a sad, sad panda |
| 10:13 | <Rik`_> | this really shows how careful he was before bahsing |
| 10:16 | <annevk> | fuck |
| 10:16 | <annevk> | my dreamhost account was hacked |
| 10:16 | <hsivonen> | annevk: when? |
| 10:16 | <Lachy> | annevk, what happened to it? |
| 10:17 | <annevk> | I'm not sure when it happened, but they modified a file and redirected people coming from search engines to some scam site |
| 10:17 | <Lachy> | check the modification date on that file |
| 10:17 | <annevk> | and they also fucked up my yearly archive file encoding somehow |
| 10:18 | <annevk> | Lachy, sigh, I modified it already :/ |
| 10:18 | <Lachy> | ok |
| 10:18 | <Lachy> | did they edit .htaccess, or one of your PHP or python scripts? |
| 10:18 | <annevk> | PHP |
| 10:19 | <annevk> | but I have not checked everything yet |
| 10:19 | <annevk> | I have changed my dreamhost account and database password |
| 10:21 | <gsnedders> | Do you use any non-custom stuff on it? I remember WP having some bug that was exploited like that… |
| 10:22 | <hsivonen> | WP seriously needs an action for "Delete this post, its author and all other posts from the same author" in the "Posts" admin view |
| 10:22 | <annevk> | yes, in various places I do |
| 10:22 | <gsnedders> | annevk: I would suspect them more than any password. |
| 10:23 | <Lachy> | hsivonen, it's easier if you delete the author instead, and then you take out all the posts along with them |
| 10:25 | <annevk> | gsnedders, but the hack was crafted specifically for my site |
| 10:25 | <hsivonen> | Lachy: I mean the Posts view seriously needs a shortcut for that |
| 10:25 | <gsnedders> | annevk: In what way? |
| 10:25 | <Lachy> | hsivonen, yeah, I know |
| 10:25 | <hsivonen> | Lachy: I think copying and pasting usernames between the two views isn't a useful exercise |
| 10:26 | <Lachy> | no, it's annoying, but it's the most effective way I found to do it, given the limitations |
| 10:26 | <annevk> | gsnedders, they figured out how my site was structured, modified one file and added another one to aid and that was it |
| 10:26 | <hsivonen> | I'm rather nervous about having Perl and Python on my server |
| 10:26 | <hsivonen> | but buying a separate VM for Bugzilla and test cases seems like an overkill |
| 10:26 | <hsivonen> | s/Python/PHP/ |
| 10:31 | Philip` | wonders what's wrong with having Perl on one's server |
| 10:33 | <hsivonen> | Philip`: I see security patches for the Perl packages too often |
| 10:34 | <annevk> | oh I can still check modified dates |
| 10:36 | Philip` | likes that Perl has taint mode, since that makes things feel a bit safer |
| 10:36 | <annevk> | "2011-01-09 10:22" |
| 10:36 | <annevk> | five fricking days |
| 10:37 | <hsivonen> | annevk: did you already end up in some kind of badware filter? |
| 10:37 | <annevk> | I found out via a Google search |
| 10:37 | <hsivonen> | annevk: how? did Google flag your site as dangerous? |
| 10:38 | <annevk> | I was wondering what http://i.imgur.com/KpI5n.jpg would do for my own site |
| 10:38 | <annevk> | and when I ran that I got results for Viagra |
| 10:38 | <annevk> | which made me go o_O |
| 10:40 | <MikeSmith> | hsivonen: I don't have that problem with W3C bugzilla |
| 10:40 | <annevk> | but they have fiddled with my .htaccess as well |
| 10:41 | <MikeSmith> | but I can help try to figure out why it's happening |
| 10:41 | <annevk> | but they made a few mistakes which screwed up the encoding and made the 2011 archive stop working |
| 10:41 | <annevk> | that is only why I noticed that |
| 10:43 | <Philip`> | Do you have a known-good backup of the site you can diff against? |
| 10:43 | <annevk> | I wish |
| 10:51 | <annevk> | ls -lha is nice |
| 10:53 | <MikeSmith> | hsivonen: if people want to help test Firefox releases or nightlies, is there a good page to point them to for details? |
| 10:55 | <MikeSmith> | other than just http://www.mozilla.com/en-US/firefox/all-beta.html I mean |
| 10:57 | <MikeSmith> | (minus en-US) |
| 11:02 | <MikeSmith> | http://www.mozilla.org/contribute/ maybe |
| 11:03 | <Rik`> | MikeSmith: maybe http://quality.mozilla.org/docs/misc/how-can-i-help-test/ ? |
| 11:03 | <MikeSmith> | Rik`: yes! |
| 11:03 | <MikeSmith> | thanks |
| 11:04 | <Rik`> | btw, there is a sync testday today http://quality.mozilla.org/events/2011/01/11/firefox-sync-testday-friday-january14th/ |
| 11:13 | <Rik`> | MikeSmith: the getting started section on http://quality.mozilla.org/docs/ seems better |
| 11:13 | MikeSmith | takes a look |
| 11:25 | <jgraham> | AryehGregor: Pretty much all programming languages have their own distribution systems. Relying on apt or whatever doesn't really work because the timescales for updates are so different\ |
| 11:26 | <jgraham> | AryehGregor: With python the One True Way is virtualenv (always) + pip |
| 11:31 | <annevk> | I think I cleared it up |
| 11:31 | <hsivonen> | MikeSmith: http://quality.mozilla.org/docs/misc/how-can-i-help-test/ that Rik` pointed to looks like a good reference |
| 11:32 | <annevk> | I wonder how they got in |
| 11:32 | <hsivonen> | annevk: you could let Dreamhost know and see if they find something interesting in their logs |
| 11:32 | <annevk> | it seems like they got hold of the ssh login somehow as the changes are pretty local |
| 11:32 | <MikeSmith> | hsivonen: thanks |
| 11:34 | <MikeSmith> | so Google says WebVTT needs "a structure to add header-style metadata" |
| 11:35 | <MikeSmith> | the concern about non-browser applications is also interesting |
| 11:38 | <MikeSmith> | http://edocket.access.gpo.gov/cfr_2007/octqtr/pdf/47cfr15.119.pdf and https://docs.google.com/viewer?a=v&q=cache:UKnzJubrIh8J:tech.ebu.ch/docs/tech/tech3264.pdf seem worth perusing |
| 11:38 | <MikeSmith> | EBU STL and CEA-608/708 captions |
| 11:38 | <hsivonen> | MikeSmith: is that the floppy disk and MS DOS-oriented EBU spec? |
| 11:38 | <MikeSmith> | dunno |
| 11:38 | <MikeSmith> | that's the first i ever learned of it |
| 11:39 | <MikeSmith> | I guess Silvia would know |
| 11:39 | <hsivonen> | MikeSmith: oh yes. see section 2. Medium for exchange for the lolz |
| 11:40 | <MikeSmith> | heh |
| 11:40 | <MikeSmith> | beautiful |
| 11:40 | <hsivonen> | I read that spec as part of my job in 2002 |
| 11:40 | <MikeSmith> | I guess this says something about the persistence of decently specced formats |
| 11:41 | <MikeSmith> | hsivonen: note also: |
| 11:41 | <MikeSmith> | * underline: EBU STL, CEA-608 and CEA-708 support underlining of |
| 11:41 | <MikeSmith> | characters. The underline character is also particularly important for |
| 11:41 | <MikeSmith> | some Asian languages. Please make it possible to provide text |
| 11:41 | <MikeSmith> | underlines without the use of CSS in WebVTT. |
| 11:43 | <MikeSmith> | the stuff about how default-track choice works in YouTube is interesting |
| 11:44 | <MikeSmith> | hmm, "Ability to move captions out of the way" |
| 11:44 | <hsivonen> | MikeSmith: where are you quoting from? list email that hasn't reached me yet? |
| 11:44 | <MikeSmith> | seems like a UI feature |
| 11:44 | <MikeSmith> | http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2011-January/029859.html |
| 11:45 | <MikeSmith> | message from Silvia |
| 11:52 | <hsivonen> | Firefox 4 say "probably" for "video/webm". Chrome, Chromium, Opera and Midori say "maybe" |
| 11:53 | <hsivonen> | all of them say "probably" with codecs="vp8, vorbis" |
| 11:54 | <zcorpan> | "Generally, a user agent should never return "probably" for a type that allows the codecs parameter if that parameter is not present." |
| 11:55 | <hsivonen> | I wonder if we have a bug for that |
| 12:00 | <annevk> | hsivonen, DreamHost said it did not have the resources. I nonetheless asked again because of the circumstances. This was quite a specific attack. |
| 12:00 | <annevk> | And I only ever use SSH so how they got my password... |
| 12:03 | <Philip`> | annevk: Could it have been via a PHP script vulnerability? |
| 12:07 | <annevk> | Philip`, I do not really see how |
| 12:07 | <Evet> | is amplesdk+xul best for desktop-like RIAs? |
| 12:08 | <annevk> | Google Webmaster Tools did not detect malware |
| 12:08 | <annevk> | weird |
| 12:09 | Philip` | wonders why he's only ever heard of this kind of attack on Dreamhost sites |
| 12:09 | <Philip`> | (It sounds similar to http://philip.html5.org/misc/spammy-sites.txt) |
| 12:09 | <Philip`> | (Maybe it's just because almost everybody uses Dreamhost for their sites) |
| 12:23 | <Ms2ger> | AryehGregor, yes, DOM Range uses the extensions I added to anolis for it ;) |
| 12:34 | <MikeSmith> | why not use refspec? |
| 12:34 | <MikeSmith> | or respec I guess it is |
| 12:46 | <Ms2ger> | Why would I? |
| 12:46 | <Ms2ger> | And does it have cross-spec cross-references? |
| 12:47 | <annevk> | Philip`, would be nice to find out |
| 15:15 | <AryehGregor> | jgraham, there's nothing to stop Python repos from at least integrating sanely with the native package management system. Like automatically installing dependencies from the local package manager. Or, I don't know, shipping binaries instead of just source code. |
| 15:15 | <AryehGregor> | Anyway, timescales are irrelevant. You can have an APT repository that's updated every two minutes. |
| 15:16 | <AryehGregor> | I didn't say it had to be the distro's repo, just that it should use the same system, or one with comparably many features. |
| 15:16 | AryehGregor | grumbles |
| 15:16 | <jgraham> | You are proposing that PyPI provides its own repositories for linux systems? |
| 15:16 | <jgraham> | That seems complex and not that useful |
| 15:16 | <AryehGregor> | Why not? Plenty of places do it. Opera, for instance. |
| 15:17 | <AryehGregor> | Just write something that will convert eggs into .deb's. |
| 15:17 | <AryehGregor> | And allow authors to specify per-distro dependencies. |
| 15:17 | <AryehGregor> | Ms2ger, so do you have instructions somewhere for how to compile your version of Anolis, or do I just have to submit patches blindly and hope they work? Because there's a bunch more stuff I'm slated to do with Selections. |
| 15:17 | <Philip`> | .deb files aren't especially useful for distros that don't use .deb files |
| 15:17 | <jgraham> | Authors don't want to specify per-distro anything since they are likely not using linux at all |
| 15:17 | <AryehGregor> | So also have something that converts them to RPM files. |
| 15:17 | <Philip`> | Some don't use RPM files either |
| 15:18 | <jgraham> | But mostly you don't want to use system-wide packages |
| 15:18 | <AryehGregor> | easy_install doesn't use system-wide packages? |
| 15:18 | <AryehGregor> | Philip`, they can suffer for their obstinacy. |
| 15:18 | <jgraham> | The only packages I would consider installing system-wide are things like NumPy |
| 15:18 | <jgraham> | AryehGregor: It does, by default |
| 15:18 | <jgraham> | That is a bug |
| 15:18 | <AryehGregor> | I see. |
| 15:18 | <jgraham> | Like I said the One True Way is to use Virtualenv + pip |
| 15:19 | <AryehGregor> | I stand by my position: this is a mess. |
| 15:19 | <jgraham> | It's not great, but no worse than ruby + gems or java + (whatever it is they use) |
| 15:19 | <jgraham> | maven? |
| 15:19 | <Philip`> | AryehGregor: I don't want to suffer :-( |
| 15:20 | <AryehGregor> | Or PHP + PEAR, or Perl + CPAN, I guess. |
| 15:20 | <AryehGregor> | Philip`, then use a distro that doesn't behave radically differently from the distros normal people use. |
| 15:21 | <Philip`> | Some relatively normal people use Gentoo/Arch/Pardus/etc |
| 15:22 | <AryehGregor> | I dispute that characterization. |
| 15:23 | <Philip`> | Really normal people just use Windows, so it's only a difference of degree |
| 15:24 | Philip` | installs most Perl/Python packages on Gentoo via the standard Portage package system, since it's got most of the common ones that he wants |
| 15:26 | AryehGregor | should have checked that before using easy_install on all of these |
| 15:26 | <AryehGregor> | Is there a way to uninstall stuff installed via easy_install? |
| 15:26 | <AryehGregor> | Wow, they even have html5lib. |
| 15:28 | AryehGregor | finds docs, although it has no man page . . . |
| 15:28 | <karlcow> | maybe uneasy_install ;) |
| 15:30 | <jgraham> | AryehGregor: Not really |
| 15:30 | <Philip`> | Clearly they couldn't be missing a fundamental feature like that, so just run easy_install --help and look for the uninstall option, which is, uh, ... not there? |
| 15:31 | <jgraham> | In general the install script can do anything so it's not really trivial to implement |
| 15:31 | <AryehGregor> | I did easy_install -mxN followed by rm -rf on the .egg directory it mentioned when I did that, will that horribly break everything? |
| 15:31 | <AryehGregor> | Yeah, I actually remember that this is what I was told one time before when I used easy_install. |
| 15:31 | <jgraham> | No. |
| 15:31 | <jgraham> | It shouldn't horribly break everything |
| 15:31 | <jgraham> | pip has uninstall |
| 15:32 | <AryehGregor> | I guess I must have optimistically hoped that whoever was responsible for that decision must have been clubbed to death by an angry mob and replaced by someone not entirely incompetent in the interim. |
| 15:32 | <jgraham> | Did I mention that virtualenv + pip is the way forward? |
| 15:32 | <AryehGregor> | Yes, but the stuff I was pointed to by docs said easy_install. |
| 15:32 | <jgraham> | AryehGregor: That is not quite unlike the truth |
| 15:32 | <Philip`> | Python library install scripts surely can't be more complex than arbitrary Linux package install scripts, and Linux package managers provide uninstall, so that doesn't seem like a very good excuse |
| 15:32 | <AryehGregor> | Like this: http://code.google.com/p/html5lib/wiki/UserDocumentation |
| 15:32 | <AryehGregor> | Yeah, presumably you just have to write a matching uninstall script. |
| 15:32 | <MikeSmith> | what Philip` said |
| 15:32 | <jgraham> | AryehGregor: pip will install anything easy)install will install |
| 15:33 | <MikeSmith> | I guess the easy_ part means "simple minded" |
| 15:33 | <MikeSmith> | a package-management system that doesn't provide an uninstall mechanism is nuts |
| 15:33 | <jgraham> | (they didn't club him to death, but they did treat him as damage to be routed around) |
| 15:34 | Philip` | has never installed html5lib, he always just downloads from SVN/Hg and points PYTHONPATH at it, because that seems the least confusing option |
| 15:34 | <jgraham> | (by writing better replacements for his tools) |
| 15:34 | jgraham | generally clones from hg and does python setup.py develop inside a virtualenv |
| 15:37 | <AryehGregor> | What exactly is virtualenv? |
| 15:38 | <jgraham> | It is a way of creating a python environment in a given directory |
| 15:38 | <jgraham> | usually the root directory of whatever you are writing |
| 15:39 | <jgraham> | so if you are creating a spam project, you do |
| 15:39 | <jgraham> | vitrualenv spam |
| 15:39 | <jgraham> | and it wil create ./spam |
| 15:39 | <jgraham> | then cd spam; source bin/activate |
| 15:40 | <jgraham> | and all python packages you install whilst the environment is activated will go in the spam directory |
| 15:40 | <AryehGregor> | Does anyone know how to get blame info for Mozilla stuff from the pre-hg era? I can't quickly find instructions on Google. |
| 15:40 | <jgraham> | and that will be first on your python path, so everything will be loaded from there too |
| 15:40 | <jgraham> | (--no-site-packages means that it will never look outside the virtualenv) |
| 15:41 | <jgraham> | http://iamzed.com/2009/05/07/a-primer-on-virtualenv/ |
| 15:41 | <Philip`> | AryehGregor: http://mxr.mozilla.org/mozilla/source/ ? |
| 15:41 | <Philip`> | then select file then select CVS Blame |
| 15:42 | <AryehGregor> | I don't see a CVS blame option here: http://mxr.mozilla.org/mozilla-central/source/layout/generic/nsSelection.cpp |
| 15:42 | <Philip`> | http://mxr.mozilla.org/mozilla/source/layout/generic/nsSelection.cpp |
| 15:42 | <Philip`> | (No -central) |
| 15:43 | <AryehGregor> | Ah. |
| 15:48 | <AryehGregor> | Thanks. |
| 15:59 | <Ms2ger> | AryehGregor, here now |
| 15:59 | <AryehGregor> | Ms2ger, so how do I get your version of anolis to work so I can test out the spec? |
| 16:03 | <AryehGregor> | :( |
| 16:06 | <annevk> | we should really find a way to get the work from Ms2ger integrated in pimpmyspec.net |
| 16:06 | <annevk> | then HTML5 can use it too |
| 16:07 | <annevk> | and nobody has to deal with pesky Python dependencies |
| 16:07 | <annevk> | except for jgraham |
| 16:07 | <jgraham> | "A good way" === me pulling his repository |
| 16:07 | <jgraham> | and maybe making UI changes |
| 16:07 | <jgraham> | I don't really want to make the UI changes :) |
| 16:08 | <jgraham> | What extra options does it add? |
| 16:08 | <annevk> | his version also takes more input |
| 16:08 | <annevk> | cross-spec cross-references |
| 16:08 | <annevk> | supercool |
| 16:08 | <jgraham> | How do you provide those? |
| 16:08 | <a-ja> | re: google comments -- version numbers in webvvt files? seriously? |
| 16:08 | <a-ja> | deja vu |
| 16:09 | <annevk> | the generic metadata thing also seems kind of dodgy |
| 16:09 | <annevk> | jgraham, euhm let me check |
| 16:10 | <annevk> | jgraham, does https://bitbucket.org/ms2ger/dom-range/src/170a27569ef0/Makefile help? |
| 16:11 | <jgraham> | annevk: a little. I'm not sure what xrefs.json is for |
| 16:11 | <jgraham> | I expected it to be input, but it seems to be output |
| 16:12 | <annevk> | https://bitbucket.org/ms2ger/anolis/src has a README |
| 16:14 | <jgraham> | Oh, that is a bit helpful |
| 16:15 | <jgraham> | I'm not sure I support references at all |
| 16:15 | <annevk> | section 4.4 and 4.5 |
| 16:15 | <annevk> | yeah, references is another new feature |
| 16:15 | <jgraham> | Yeah |
| 16:17 | <annevk> | I think ideally trusted people can update the references so they do not need to be provided separately |
| 16:17 | <annevk> | same maybe for cross-spec cross-ref |
| 16:17 | <jgraham> | I think they need to be provided |
| 16:17 | <jgraham> | But I think it should be OK to provide a URL |
| 16:18 | <jgraham> | Which I guess doesn't work right now |
| 16:18 | <annevk> | you mean to override? |
| 16:18 | <annevk> | I guess that's fine, I hope I never need it though :) |
| 16:18 | <jgraham> | No, I mean that you have to provide a json file with paths to the differnt xspec-xrefs |
| 16:19 | <jgraham> | instead you should be able to provide URLs to the different xspec-xrefs |
| 16:19 | <jgraham> | (or rather to the json files that provide them) |
| 16:19 | <annevk> | I see... can't pimpmyspec simply keep them around? |
| 16:20 | <jgraham> | Not really. There is no unique name or anything |
| 16:20 | <annevk> | shortname of specs ought to be unique |
| 16:20 | <jgraham> | Yeah, but afaict nothing enforces that in the current design |
| 16:21 | <jgraham> | and you need to deal with updates and such |
| 16:21 | <annevk> | if all specs are generated through pimpmyspec.net you should always have uptodate info :) |
| 16:21 | <jgraham> | Anyway, this is all possible but will probably require some minor backend changes |
| 16:36 | <AryehGregor> | I think I'm writing too many tests. |
| 16:36 | <AryehGregor> | Well, we'll see. |
| 16:37 | <Philip`> | How is it possible to have too many tests? |
| 16:37 | <Ms2ger> | If they're manual ones, not too hard |
| 16:37 | <AryehGregor> | These are automated. |
| 16:37 | <AryehGregor> | But they still might run way too long. |
| 16:37 | <AryehGregor> | I dunno, I haven't finished writing them yet. |
| 16:38 | <Ms2ger> | Computer time is cheap |
| 16:38 | <AryehGregor> | Tell that to Mozilla, they removed one of my textarea.maxLength tests because it was taking several seconds to run. :( |
| 16:38 | <Philip`> | Developer time waiting for tests to complete running is expensive |
| 16:38 | <AryehGregor> | (Apparently synthesizing keystrokes is fairly expensive?) |
| 16:39 | <Ms2ger> | That's possible |
| 16:39 | <AryehGregor> | Hmm, I can't really give all my tests meaningful distinct names. Oh well, I'll just number them. |
| 16:40 | <AryehGregor> | It's annoying that in jgraham's framework you can't name asserts. |
| 16:40 | <AryehGregor> | But I guess it's not a huge deal. |
| 16:40 | <jgraham> | AryehGregor: You can, no? |
| 16:40 | <AryehGregor> | Can you? |
| 16:40 | <AryehGregor> | I'll have to start doing that, then. |
| 16:40 | <jgraham> | assert_true(true, "True is true") |
| 16:41 | <AryehGregor> | Oh, so you can. |
| 16:41 | <AryehGregor> | Look at that. |
| 16:41 | <AryehGregor> | I'll have to update my tests now. |
| 16:41 | <Ms2ger> | Are you on bitbucket, btw? |
| 16:41 | jgraham | is happy to have brought a little joy to your life |
| 16:41 | <jgraham> | Ms2ger: Who AryehGregor or me? |
| 16:41 | <Ms2ger> | Aryeh |
| 16:41 | <AryehGregor> | No. |
| 16:43 | <Ms2ger> | Then I can't give you write access, apparently |
| 16:44 | <AryehGregor> | So much for distributed. |
| 16:45 | <AryehGregor> | I'll make an account later, I guess. |
| 16:55 | <AryehGregor> | What's the JavaScript way to check if an element is in an array? .indexOf != -1, no cute "in" stuff like in Python? |
| 16:57 | <Ms2ger> | How well supported is indexOf? |
| 16:57 | <AryehGregor> | I don't know. What's the alternative? |
| 16:59 | <Ms2ger> | for (var i = 0; i < haystack.length; ++i) { if (haystack[i] === needle) { win(); } } lose() |
| 17:00 | <Philip`> | http://www.w3schools.com/jsref/jsref_obj_array.asp doesn't mention Array.indexOf |
| 17:00 | <Philip`> | so I guess it didn't exist when they wrote that site in the 90s |
| 17:00 | <nimbupani> | http://w3fools.com/ :)) |
| 17:10 | <annevk> | nimbupani, nice |
| 17:16 | <AryehGregor> | nimbupani, should I contact them about the inaccuracies I've found in their corrections? :) |
| 17:17 | <AryehGregor> | There are actually quite a lot. |
| 17:17 | <AryehGregor> | Although not nearly as bad as w3schools. |
| 17:17 | <nimbupani> | them = me paul_irish and a bunch of others who are also lurking here |
| 17:18 | <nimbupani> | so please say it here |
| 17:18 | <nimbupani> | so I can correct em on the fly :) |
| 17:18 | <AryehGregor> | Ah. |
| 17:18 | <AryehGregor> | "Again, frames are considered among the very worst of practices in modern Web development. In fact, they are considered so bad, they have been removed completely from the HTML5 spec." There are still implementation conformance requirements for <frame> in HTML5. |
| 17:19 | <AryehGregor> | '"Correctly" does not mean "try to guess and maybe get it right some of the time", which is what actually happens, usually in different ways between different browsers.' The last part is wrong in common cases (end-tag substitution is often pretty uniform), and will be totally wrong once all browsers are using HTML5 parsers. |
| 17:19 | <AryehGregor> | "In HTML-dialect HTML5, boolean attributes like hidden do not have a value." They do have a value. The value is just irrelevant to how the attribute is processed. |
| 17:20 | <AryehGregor> | "Furthermore, <br /> isn't semantic and probably should not be mentioned at all." <br> is semantic in some cases, such as poems or addresses. |
| 17:20 | <AryehGregor> | "This is wrong. All of these elements were deprecated in HTML4." But <i>, <b>, and <small> are no longer deprecated as of HTML5. |
| 17:21 | <AryehGregor> | "This code is wrong. Non-block-level elements (such as <input> or CDATA) are not valid directly inside <form> tags until HTML5." Thus it's right, as of HTML5 . . . |
| 17:21 | <nimbupani> | got it. |
| 17:22 | <AryehGregor> | "This is false. URL encodings are not ISO-8859-1. RFC 3986 is the standard defining URL encoding, and specifies that textual characters should, in fact, be mapped to UTF-8 octets." But in practice they're often generated and processed using various random non-UTF-8 encodings (AFAIK). |
| 17:22 | <Philip`> | "XHTML has been discontinued" - not really; XHTML5 is still being developed |
| 17:23 | <karlcow> | maybe the good sentence is "XHTML 2.0 has been discontinued" |
| 17:24 | <Philip`> | That would be an irrelevant sentence |
| 17:24 | <annevk> | AryehGregor, actually no |
| 17:24 | <AryehGregor> | No to which? |
| 17:24 | <Ms2ger> | All of them ;) |
| 17:24 | <annevk> | AryehGregor, only the query string can end up as non-UTF-8 |
| 17:24 | <AryehGregor> | Well, that's part of the URL, isn't it? |
| 17:24 | <karlcow> | Philip`: would you prefer? "The work on XHTML 2.0 has been discontinued and published as a W3C note. XHTML5 is the new XML serialization of HTML5" |
| 17:24 | <annevk> | AryehGregor, yes, but the path is always encoded as UTF-8 |
| 17:25 | <AryehGregor> | Well, okay. |
| 17:25 | <AryehGregor> | The point stands. |
| 17:25 | <Ms2ger> | I believe the point was that any reference to XHTML2 is irrelevant |
| 17:25 | <annevk> | AryehGregor, well, it's not random and it's also not various |
| 17:26 | <Philip`> | "Oh, and by the way, user-agent sniffing is a very bad thing, because is easily spoofable" - that's not why it's a very bad thing |
| 17:26 | <annevk> | AryehGregor, would be a pretty poor substitute :) |
| 17:26 | <AryehGregor> | "The language was eventually standardized under the creative name ECMAScript by the EMCA international standards organization" Typo, EMCA instead of ECMA. |
| 17:26 | <karlcow> | hehe |
| 17:26 | <AryehGregor> | annevk, all right, all right, but the original was still wrong. |
| 17:27 | <karlcow> | nimbupani, paul_irish, miketaylr: I think you need a way to collect issues ;) |
| 17:27 | <nimbupani> | oops thanks AryehGregor |
| 17:27 | <nimbupani> | we have a github :) |
| 17:27 | <TabAtkins> | "Also, the note about the variable argument sounds like it's half-BS to me." Complete BS, right? |
| 17:27 | <nimbupani> | http://github.com/paulirish/w3fools |
| 17:27 | <miketaylr> | yeah, maybe tweet the github link, nimbupani? |
| 17:27 | <AryehGregor> | "On that note, W3C doesn't mention variable scoping at all…" Presumably you mean W3Schools here? |
| 17:27 | <paul_irish> | karlcow: i think *they* need a way to collect issues :) |
| 17:27 | <nimbupani> | OH SHIT AryehGregor |
| 17:27 | <nimbupani> | damn it |
| 17:28 | <karlcow> | hehe |
| 17:28 | <karlcow> | You gotta love irc :) |
| 17:30 | <AryehGregor> | "Furthermore, addressing has nothing to do with TCP. It happens on the OSI network layer; this information is very specific to IPv4 and is going to be increasingly obsolete as time goes on." To be extremely pedantic, TCP does actually know about IP addresses, in that it uses the source and destination IP addresses to compute the checksum it puts in the segment header. |
| 17:31 | <nimbupani> | AryehGregor: our audience is mostly people who think w3c and w3schools are the same :/ |
| 17:32 | <AryehGregor> | "Actually, correct HTML is <br>, correct XHTML is <br />" <br /> is correct HTML5, as well as <br>. |
| 17:32 | <AryehGregor> | (likewise in at least one other case) |
| 17:32 | <nimbupani> | yes agreed. |
| 17:33 | <Philip`> | W3Schools says "TCP/IP" anyway, not "TCP" |
| 17:33 | <Philip`> | so complaining about TCP is irrelevant |
| 17:33 | <annevk> | "We think the resources we've recommended are superior for a variety of reasons." -- either say why or omit "variety of reasons" |
| 17:34 | <AryehGregor> | Okay, that's enough nitpicking. |
| 17:34 | <AryehGregor> | Back to work. |
| 17:34 | <AryehGregor> | Yay tests! |
| 17:34 | <Philip`> | Also addressing doesn't just happen on the network layer, it happens all over the place (like in DNS and FTP) |
| 17:35 | <nimbupani> | Thanks AryehGregor appreciate it quite a bit. |
| 17:38 | <annevk> | "Question 6 Bold? B? Really!?" and "Question 7 Italic? I? Really!?" is actually correct |
| 17:38 | <annevk> | better that people use <b> and <i> than something uninformed |
| 17:39 | <TabAtkins> | <b> is certainly better than <span style="font-weight:bold;"> |
| 17:39 | <TabAtkins> | And in many cases better than <strong>. |
| 17:39 | <AryehGregor> | Yeah, I wasn't even getting into the stuff that was dodgy but not actually incorrect. |
| 17:41 | <karlcow> | <b>behaviors</b> and <strong>opinions</strong> on IRC |
| 17:42 | <TabAtkins> | Oh jeez, I just frightened myself by looking at a bill, until I realized the "balance" was just my mortgage. Whew. |
| 17:42 | <miketaylr> | heh |
| 17:51 | <AryehGregor> | Okay, I apparently hit a bug in jgraham's test harness: http://aryeh.name/tests-root/tests/submission/AryehGregor/selection-extend.html |
| 17:51 | <AryehGregor> | Error: invalid array length |
| 17:51 | <AryehGregor> | Source File: http://aryeh.name/tests-root/tests/resources/testharness.js |
| 17:51 | <AryehGregor> | Line: 944 |
| 17:51 | <AryehGregor> | This is why I should start by writing smaller tests first. |
| 17:56 | <MikeSmith> | othermaciej: fyi - |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-2dcontext-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-html-alt-techniques-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-html-polyglot-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-html5-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-html5-diff-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-microdata-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-rdfa-in-html-20110113/ |
| 17:56 | <MikeSmith> | http://www.w3.org/TR/2011/WD-html-markup-20110113/ |
| 17:57 | <othermaciej> | MikeSmith: cool |
| 17:57 | <MikeSmith> | still need webmaster to set up the TR symlinks |
| 17:57 | <MikeSmith> | but that should be done today |
| 17:57 | <MikeSmith> | pending any more pubrules or broken-link problems that might remain for me to fix |
| 17:58 | <MikeSmith> | annevk: btw, you still got a lot of borken fragment refs in the diffs doc |
| 17:58 | <AryehGregor> | This call stack is like 30 deep of maps and substitutes. Somehow a 0 got passed as the array argument to map(), but I have no idea how. |
| 17:58 | <MikeSmith> | annevk: due to stuff ending up in different split-out files than it did before |
| 17:59 | <MikeSmith> | in your editor's draft I mean |
| 17:59 | <nimbupani> | some bloopers fixed http://w3fools.com/ |
| 17:59 | <MikeSmith> | http://validator.w3.org/checklink/checklink?uri=http%3A%2F%2Fdev.w3.org%2Fhtml5%2Fhtml4-differences%2F&hide_type=all&depth=&check=Check |
| 18:00 | <annevk> | MikeSmith, bah |
| 18:00 | <annevk> | MikeSmith, hmm, so you fixed them or something? |
| 18:00 | <MikeSmith> | yeah |
| 18:00 | <MikeSmith> | sorta |
| 18:00 | <MikeSmith> | i dealt with it |
| 18:01 | <annevk> | maybe I should stop linking directly to HTML5 |
| 18:01 | <MikeSmith> | if anybody notices any problems with any of the other WD- above, please let me know so I can fix it before the hammer comes down on me |
| 18:01 | <MikeSmith> | links are good |
| 18:01 | <annevk> | feel free to patch the source file |
| 18:01 | <annevk> | i.e. http://dev.w3.org/cvsweb/html5/html4-differences/Overview.src.html |
| 18:01 | <MikeSmith> | the mood does not strike me |
| 18:02 | karlcow | takes his red pen and starts reading :p |
| 18:02 | <annevk> | okay |
| 18:02 | <MikeSmith> | I welcome others to feel free |
| 18:03 | <MikeSmith> | If pubrules were a person I would karate chop him in the throat |
| 18:03 | <AryehGregor> | Ah, it's working now. |
| 18:03 | <AryehGregor> | Is 4400 too many tests for extend()? |
| 18:03 | <AryehGregor> | There are still a whole bunch of things I haven't tested. |
| 18:04 | <AryehGregor> | But first let me check that my existing tests are sane . . . |
| 18:04 | <AryehGregor> | Actually, it's 4401, sorry. |
| 18:05 | <myakura> | don't we need to update copyright notice? (year++) |
| 18:06 | <MikeSmith> | myakura: <sigh> yeah, I expect so |
| 18:06 | <MikeSmith> | thanks for catching that |
| 18:06 | <Ms2ger> | Anolis supports doing that automatically ;) |
| 18:07 | MikeSmith | heads off to do some find. | xargs perl -pi -e |
| 18:07 | <MikeSmith> | Anolis is broken by design |
| 18:07 | <gsnedders> | Meow :'( |
| 18:07 | <MikeSmith> | it is a work-alike for something else that was broken by design |
| 18:07 | <gsnedders> | I is cute and cuddly! |
| 18:07 | <MikeSmith> | kind of like SVN was meant to be a workalike for CVS |
| 18:08 | <MikeSmith> | and we all know what a brilliant design goal that turned out to be |
| 18:08 | <Ms2ger> | What, HTML? |
| 18:09 | <MikeSmith> | I meant the cleverly named "CSS3 Module Postprocessor" |
| 18:10 | <karlcow> | MikeSmith: which document had broken fragments? |
| 18:10 | <MikeSmith> | none document |
| 18:10 | <MikeSmith> | now |
| 18:10 | <MikeSmith> | as far as WDs |
| 18:10 | <karlcow> | ah ok |
| 18:10 | <MikeSmith> | most all of the editor's drafts have broken fragments |
| 18:10 | <MikeSmith> | except for Hixie 's |
| 18:11 | <karlcow> | ok ok. had misunderstood |
| 18:15 | <AryehGregor> | Ms2ger, you previously mentioned that when the test harness fails an expected exception with the reason "threw with code INDEX_SIZE_ERR (1) expected INDEX_SIZE_ERR (1)" on Firefox, it's because of a known Firefox bug. What bug is that? |
| 18:15 | <AryehGregor> | It's really annoying. |
| 18:15 | <AryehGregor> | Because it causes every single expected exception assert to fail, which seriously clutters up the test results. |
| 18:15 | <Ms2ger> | Hmm, I guess that would've just been known to me |
| 18:15 | <Ms2ger> | I'll file |
| 18:19 | <AryehGregor> | Thanks. |
| 18:22 | <MikeSmith> | ah geez |
| 18:22 | <MikeSmith> | http://pitchfork.com/news/41200-rip-broadcasts-trish-keenan/ |
| 18:23 | <MikeSmith> | Trish Keenan from Broadcast |
| 18:28 | <Ms2ger> | Filed, cc:d you |
| 18:28 | <AryehGregor> | Thanks. |
| 18:31 | <AryehGregor> | Hixie, first pass at extend() tests (subject to change as I add more in real time): http://aryeh.name/tests-root/tests/submission/AryehGregor/selection-extend.html |
| 18:32 | <AryehGregor> | Awesome, Opera only passes like 20 of them. |
| 18:32 | <AryehGregor> | gsnedders was right when he said it wasn't web-compatible. |
| 18:33 | <gsnedders> | It works for all but a few sites! :P |
| 18:35 | <Ms2ger> | WebKit-based browsers also support setBaseAndExtent,... |
| 18:35 | <Ms2ger> | AryehGregor, seems like they have more work for you ;) |
| 18:35 | <AryehGregor> | I believe modify() is next on my list. |
| 18:36 | <AryehGregor> | By the way, you never answered me about how I'm supposed to compile your spec. |
| 18:36 | <Ms2ger> | I did, but my net didn't like you |
| 18:37 | <AryehGregor> | Could you do it again and hope that your net is feeling more friendly now? |
| 18:37 | <AryehGregor> | Perhaps you could feed it or stroke it gently to improve its disposition. |
| 18:37 | <Ms2ger> | Heh |
| 18:37 | <Ms2ger> | Do |hg clone ssh://hg⊙bo/ms2ger/specification-data data| inside your dom-range dir |
| 18:38 | <AryehGregor> | Although I admit to uncertainty as to how a net could have anything to do with it. Do you keep your computer inside a net or something? |
| 18:38 | <AryehGregor> | Ah, that was the "missing data" issue. |
| 18:38 | <Ms2ger> | Yeah |
| 18:39 | <AryehGregor> | You could add a make rule for that, couldn't you? |
| 18:39 | <AryehGregor> | How do I get your anolis instead of the official one? |
| 18:39 | <AryehGregor> | I don't have a public key that allows me to check out from there. |
| 18:39 | <Ms2ger> | hg clone ssh://hg⊙bo/ms2ger/anolis && cd anolis && sudo python setup.py install |
| 18:39 | <AryehGregor> | Maybe there's an HTTP URL? |
| 18:39 | <Ms2ger> | There is |
| 18:40 | AryehGregor | correctly guesses it |
| 18:43 | <Ms2ger> | Please pull anolis again, I just pushed a patch I had locally |
| 18:46 | <AryehGregor> | Has anyone ever remarked that anolis is extremely verbose and painful to use? |
| 18:47 | <AryehGregor> | SyntaxError: Term not defined: root-element in domcore. |
| 18:47 | <AryehGregor> | root-element seems to be defined there. What am I missing? data needs an update? |
| 18:53 | <Ms2ger> | You want "root element", I think |
| 18:54 | <AryehGregor> | Convenient. |
| 18:54 | <annevk> | with a dash is the ID |
| 18:55 | <AryehGregor> | Given how anolis works, aren't all the id's in DOM Range kind of poorly chosen? Like, why couldn't you call anchorNode just "anchorNode" instead of "dom-selection-anchorNode"? Wouldn't that let you leave off all those titles? |
| 18:56 | <Ms2ger> | Following the style in HTML |
| 18:56 | <AryehGregor> | Hurrah, success. |
| 18:56 | <AryehGregor> | http://aryeh.name/tmp/dom-range.html#dom-selection-extend |
| 18:57 | <Ms2ger> | Yay |
| 19:00 | <hsivonen> | annevk: https://twitter.com/#!/kroc_camen 's latest tweets might be of interest |
| 19:00 | <hsivonen> | site hack via AWStats |
| 19:15 | <MikeSmith> | Hixie: anubis.dkuug.dk appears to be no more |
| 19:16 | <MikeSmith> | reference that points to in the spec |
| 19:16 | <MikeSmith> | for ISO885911 |
| 19:22 | <AryehGregor> | Okay, I think I've written enough tests for now. I'm up to 16,525 and it takes like ten seconds to run them all on Chrome. |
| 19:22 | <AryehGregor> | Oh, wait, some are broken. |
| 19:22 | <oojacoboo> | if you were marking up a privacy policy or terms and conditions or something, would you use <article> ? |
| 19:23 | <oojacoboo> | the spec seems to suggest that this is more blog related, as most of the spec does... thinking everything is a blog |
| 19:23 | <AryehGregor> | Sure, <article> is fine for that. |
| 19:24 | <Ms2ger> | Has the [ISO8601] ref in HTML always needed a password? |
| 19:25 | <oojacoboo> | I wish someone would rewrite the spec without such a heavy slant towards blogs |
| 19:26 | <oojacoboo> | it's very frustrating and does not provide a very comprehensive outlook into use case scenarios for the elements |
| 19:26 | <AryehGregor> | If it makes you feel better, every previous version of the spec was slanted even more heavily toward static documents. |
| 19:26 | <oojacoboo> | AryehGregor: static? |
| 19:26 | <AryehGregor> | You know. |
| 19:27 | <AryehGregor> | Like .html files that you wrote by hand and uploaded to your server by FTP. |
| 19:27 | <AryehGregor> | In 1996. |
| 19:27 | <Ms2ger> | oojacoboo, suggest good examples and they will very probably be added |
| 19:27 | <AryehGregor> | But yeah, feel free to suggest further examples for everything. |
| 19:27 | <AryehGregor> | You can use the comment form. |
| 19:27 | <AryehGregor> | I'm pretty sure the editor thinks there aren't enough examples. |
| 19:27 | <oojacoboo> | Ms2ger: how about thinking about it like you are developing an ecommerce site, and then make the changes |
| 19:27 | <oojacoboo> | you will see then that it needs SERIOUS edits |
| 19:27 | <Ms2ger> | Patches welcome |
| 19:27 | <AryehGregor> | Well, none of us are actually developing an ecommerce site, so probably you'd have more insight into that than us. |
| 19:28 | <AryehGregor> | If you want to give specific suggestions, they'd most likely be welcomed. |
| 19:28 | <oojacoboo> | AryehGregor: fair enough |
| 19:28 | <oojacoboo> | AryehGregor: the problem is that I have a hard enough time interpreting the intended use for certain elements |
| 19:29 | <AryehGregor> | Frankly, the elements like <article> are mostly interchangeable with <div>s in practice, and will probably remain so forever. |
| 19:29 | <oojacoboo> | just yesterday, I struggled with this <address> element... it's incredibly slanted towards blogs, assuming I want to use it to wrap referenced links |
| 19:29 | <AryehGregor> | So if you misuse them, it's not the end of the world. |
| 19:29 | <AryehGregor> | <address> is old, the definition dates back years. |
| 19:29 | <oojacoboo> | AryehGregor: well, it's still very lightly used and needs more clarification |
| 19:29 | <oojacoboo> | I used it for our company address on the /contact page |
| 19:30 | <oojacoboo> | maybe not intended, but that makes perfect sense to me |
| 19:30 | <AryehGregor> | That's correct usage. |
| 19:30 | <AryehGregor> | Okay, I really need to get rid of some of these tests. |
| 19:30 | <oojacoboo> | AryehGregor: the spec wouldn't lead you to believe that |
| 19:30 | <oojacoboo> | infact it says that a postal address is generally incorrect |
| 19:31 | <Ms2ger> | Generally |
| 19:31 | <oojacoboo> | the best thing for this spec would be for one of the authors to pick an ecommerce site, and rewrite it, and learn the issues, and doc them |
| 19:31 | <Ms2ger> | But not if it's contact information for the site author |
| 19:32 | <oojacoboo> | I just say ecommerce, b/c it's much different from a blog |
| 19:32 | <AryehGregor> | oojacoboo, realistically, the author of the spec has lots and lots of other more pressing things to do, unfortunately. |
| 19:32 | <oojacoboo> | AryehGregor: well, the people that are implementing this are going to fuck it up otherwise |
| 19:32 | <oojacoboo> | so, if that isn't important... |
| 19:32 | <AryehGregor> | Practically no authors read the spec anyway. |
| 19:33 | <oojacoboo> | mmm, it's the best reference I can find :/ |
| 19:33 | <AryehGregor> | the non-machine-checkable authoring conformance requirements are going to be almost totally ignored no matter what, even in very high-profile "standards-compliant" software. |
| 19:33 | <AryehGregor> | s/^t/T/ |
| 19:33 | <annevk> | hsivonen, thanks, but DreamHost is not using them |
| 19:34 | <oojacoboo> | AryehGregor: are we not banking on UAs taking advantage of these semantic elements for better interaction with webpages? |
| 19:34 | <AryehGregor> | oojacoboo, there are no differences in behavior defined for them, so I'm banking on no, they won't. |
| 19:34 | <oojacoboo> | for instance, allowing <address> to be clickable to open the address in google maps, or something |
| 19:34 | <AryehGregor> | <address> has existed since the 1990s, and no UA has done that. |
| 19:34 | <AryehGregor> | So there's your answer. |
| 19:35 | <oojacoboo> | AryehGregor: but mobile is just taking off, so I disagree |
| 19:35 | <AryehGregor> | I don't see what that has to do with it. |
| 19:36 | <oojacoboo> | AryehGregor: with the limited screen real estate and the ability to interpret the content of a page, mobile browsers could customize the viewing experience dramatically |
| 19:36 | AryehGregor | pares it down to 10837 tests |
| 19:36 | <oojacoboo> | take the idea that you could choose to do something different with <aside> as it isn't related to the page |
| 19:37 | <oojacoboo> | or ping your <nav> into the UA |
| 19:37 | <oojacoboo> | pin* |
| 19:37 | <oojacoboo> | etc |
| 19:37 | <AryehGregor> | oojacoboo, could, yes. I've thought of that too. But it's not specced, so UAs won't bother doing it now because not enough authors use the elements, so authors will use the elements randomly, so UAs won't do it because it would mess up pages. |
| 19:37 | <AryehGregor> | That's my prediction. |
| 19:37 | <AryehGregor> | It would be cool, though, yes. |
| 19:37 | <oojacoboo> | I see that as the future, but only when people code their markup properly |
| 19:37 | <oojacoboo> | and currently thats damn near impossible to understand |
| 19:38 | <oojacoboo> | the day where we have to rewrite a page multiple times for different UAs could be much closer! |
| 19:38 | <oojacoboo> | esp with you look at mobile/tablet/desktop |
| 19:39 | <oojacoboo> | meh, maybe I care too much... |
| 19:40 | <AryehGregor> | Oh, you can't actually make a doctype the boundary point of a Range. How sad. |
| 19:40 | <Ms2ger> | The question is, can you use the text node child of an attr node |
| 19:41 | <AryehGregor> | It's not a matter of caring too much or too little. It's a matter of separating what you'd ideally like to see, from what's actually attainable in real life given the forces involved. |
| 19:41 | <AryehGregor> | Ms2ger, attr nodes don't exist in Web DOM. |
| 19:41 | <Ms2ger> | Fortunately |
| 19:41 | <oojacoboo> | one other question before I get back to work... when it comes to heading elements |
| 19:41 | <oojacoboo> | why does the spec suggest nearly everything be an <h1> ? |
| 19:42 | <oojacoboo> | am I not to assume that this would throw bots for some loops |
| 19:42 | <Ms2ger> | Because it's easier to move content around like that |
| 19:42 | <oojacoboo> | Ms2ger: why not kill the h2-h6 then |
| 19:43 | <Ms2ger> | Because people use the |
| 19:43 | <Ms2ger> | them* |
| 19:43 | <oojacoboo> | or better yet, start to deprecate it by introducing a new heading element |
| 19:43 | <Ms2ger> | That's not very backwards-compatible |
| 19:43 | <oojacoboo> | people use <center> as well, but it's been deprecated |
| 19:44 | <Ms2ger> | Adding new elements is expensive |
| 19:44 | <Ms2ger> | Multimillion dollar-expensive, if I recall Hixie's calculation correctly |
| 19:44 | <AryehGregor> | That calculation is just silly, though. |
| 19:44 | <oojacoboo> | that's understandable |
| 19:44 | <Ms2ger> | Somewhat |
| 19:45 | <oojacoboo> | I still don't think that keeping elements that are confusing as shit, is smart |
| 19:45 | <AryehGregor> | Adding an element isn't really much more expensive than many other spec changes. |
| 19:45 | <oojacoboo> | infact, I just straight up disagree with using the <h1> all over the site, refuse to code like that |
| 19:45 | <AryehGregor> | Feel free not to, the spec allows you not to. |
| 19:45 | <AryehGregor> | Ooh, I think my tests found a stupid corner case where my spec doesn't match reality! |
| 19:46 | <oojacoboo> | well, if my comments are useless, I'll leave it be then |
| 19:46 | <AryehGregor> | Your comments are potentially useful, it's just that we've heard some of them before and have discussed them at length. |
| 19:46 | <oojacoboo> | might I suggest pulling in some folks that code things other than blogs on a regular basis though |
| 19:46 | <AryehGregor> | Like the using <h1> everything thing, it's been discussed a lot. |
| 19:47 | <AryehGregor> | oojacoboo, are you volunteering? If not, where do you want us to find such people? |
| 19:47 | <oojacoboo> | AryehGregor: maybe you have discussed them with the wrong people |
| 19:47 | <AryehGregor> | (FWIW, I code wikis, not blogs) |
| 19:47 | <AryehGregor> | If so, only because the right people didn't choose to participate in the discussions. |
| 19:47 | <AryehGregor> | Feel free to hang out here, subscribe to the whatwg mailing list, read it all and chip in. |
| 19:47 | <AryehGregor> | You'll make a difference if you do that, I assure you. |
| 19:48 | <oojacoboo> | AryehGregor: one easy solution is to just familiarize yourselves with different layouts |
| 19:48 | <Ms2ger> | That's not an easy solution |
| 19:48 | <AryehGregor> | Well, yes, but who exactly is going to be doing this? We all have other things to do. |
| 19:48 | <AryehGregor> | This kind of thing takes time. |
| 19:48 | <oojacoboo> | AryehGregor: fair enough, I can def provide feedback on things from an application development standpoint |
| 19:48 | <AryehGregor> | For instance, right now I'm speccing and testing a feature that's been implemented since about 1998 but never actually specified, so everyone does it differently if they do it at all. |
| 19:49 | <erlehmann> | oojacoboo, i like the h1 pattern, since i can reuse parts of sites without nesting troubles. what exactly is wrong about it? |
| 19:49 | <oojacoboo> | erlehmann: google doesn't |
| 19:49 | <AryehGregor> | Who says Google doesn't? |
| 19:49 | <erlehmann> | oojacoboo, so. |
| 19:49 | <AryehGregor> | The person who wrote the spec works for Google. |
| 19:49 | <oojacoboo> | AryehGregor: google places priority on the content for heading elements |
| 19:50 | <AryehGregor> | <h1> is a heading element. |
| 19:50 | <Ms2ger> | Google has said on the record that that pattern is fine |
| 19:50 | <oojacoboo> | if you have 20 <h1> tags on a page, and don't have the rest of the markup coded perfectly according to their alog, you won't be optimized very well |
| 19:50 | <erlehmann> | oojacoboo, googles algorithms evolve. they employ smart people. also, outline algorithm, do you speak it? |
| 19:51 | <Ms2ger> | oojacoboo, Google says that's not true |
| 19:51 | <oojacoboo> | outline algo? |
| 19:51 | <erlehmann> | oojacoboo, if gsnedders can do this, google can too: http://gsnedders.html5.org/outliner/process.py?url=http%3A%2F%2Fblog.dieweltistgarnichtso.net |
| 19:51 | <oojacoboo> | what do you mean erlehmann |
| 19:51 | <erlehmann> | see the outline of my blog. |
| 19:51 | <erlehmann> | sectioning is a bitch and the outline algo tames it. |
| 19:52 | <oojacoboo> | erlehmann: I have seen the outliner, yes |
| 19:52 | <erlehmann> | apparently my outline still is wrong in some cases. |
| 19:52 | <oojacoboo> | if this is the idea, then what becomes the use of <h2> - <h6> ? |
| 19:53 | <annevk> | oh hey |
| 19:53 | <annevk> | Google posted an update |
| 19:53 | <annevk> | http://blog.chromium.org/2011/01/more-about-chrome-html-video-codec.html |
| 19:53 | <oojacoboo> | oh god, don't get me started on that |
| 19:53 | <Peter`> | they're going to publish plugisn for Safari and IE9 |
| 19:53 | <Peter`> | adding support for WebM |
| 19:53 | <annevk> | and Peter` put it on twitter already |
| 19:53 | <Peter`> | the rest we already know :p |
| 19:53 | <AryehGregor> | annevk, in the sense of <http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#root-element>;, what's the root element of a Document node? |
| 19:53 | <erlehmann> | oojacoboo, implied sectioning. |
| 19:53 | <AryehGregor> | The Document itself, or its element child (if any)? |
| 19:54 | <AryehGregor> | I'd say the Document itself, but it's not clear to me. |
| 19:54 | <annevk> | the Document itself as I read it |
| 19:54 | <annevk> | Ms2ger wrote that part |
| 19:54 | <oojacoboo> | erlehmann: I thought that was the role of tiered elements? |
| 19:54 | <oojacoboo> | containers* |
| 19:54 | <annevk> | or Hixie did probably and Ms2ger copied it :) |
| 19:54 | <AryehGregor> | Good, that's how I read it too. |
| 19:55 | <AryehGregor> | It's confusing for "root element" to mean two things. |
| 19:55 | <Ms2ger> | Hixie did |
| 19:56 | <erlehmann> | oojacoboo, containers are *explicit* |
| 19:57 | <oojacoboo> | erlehmann: yea, but implicit doesn't do any good for the markup |
| 19:58 | <oojacoboo> | if the spec is stating that you are to use <h1> and I use <h2> or implicit reasons, then the markup isn't being interpreted correctly |
| 19:58 | <oojacoboo> | no? |
| 19:58 | <Ms2ger> | Either is fine |
| 19:58 | <erlehmann> | oojacoboo, next you will be asking us to markup <head> and <body> |
| 19:59 | <oojacoboo> | erlehmann: don't be melodramatic |
| 19:59 | <erlehmann> | oojacoboo, html is not python. there are several ways to do something wrong. |
| 20:00 | <oojacoboo> | erlehmann: well, maybe that's half the problem |
| 20:00 | <MikeSmith> | "only related to the <video> tag, which is part of the emerging HTML platform" |
| 20:00 | <oojacoboo> | maybe there should be more structure to it |
| 20:00 | <erlehmann> | oojacoboo, feel free to spec oojacobooML. |
| 20:00 | <erlehmann> | or use XHTML. |
| 20:00 | <oojacoboo> | erlehmann: I do use xhtml in come cases |
| 20:01 | <oojacoboo> | some even |
| 20:02 | <oojacoboo> | my main concern is that with the proliferation of mobile devices in the next few years, the requirement to code a site out on a platform that will generate different markup based on the device, will be necessary. However, if there was more structure to the html, it'd be possible for UAs to display, at the request of the site with a meta tag, a different version, based on the semantic markup |
| 20:03 | <Hixie> | i made "root element" mean two things so that when i defined things in terms of the root element, it would just work for fragments without having to explicitly mention fragments (which would make the spec even more verbose otherwise) |
| 20:03 | <oojacoboo> | this would allow us to code a site once, and possible provide different stylesheets, and improve the mobile browsing experience drastically |
| 20:04 | <Hixie> | this way it's complicated for people who are able to understand why it's complicated and it doesn't complicate the spec for people who don't care about fragments |
| 20:07 | gsnedders | whisles, "anything I can do Google can too" |
| 20:12 | Ms2ger | pats gsnedders on the back |
| 20:12 | <Ms2ger> | How did the programming project turn out? |
| 20:13 | <gsnedders> | Um, well… I realized it counted for nothing (apart from having to do 7/10 exercises), I had practically ran out of time, so I gave up and didn't do anything for it. |
| 20:13 | gsnedders | is a good university student |
| 20:13 | <Ms2ger> | Hah |
| 20:13 | <gsnedders> | Speaking of which, time for me to decide what to wear to go out tonight. |
| 20:13 | <gsnedders> | (I'm not a teenage girl, honest.) |
| 20:15 | <Ms2ger> | I hadn't doubted that before, but now you mention it... |
| 20:15 | <gsnedders> | (Time to prove this point by deciding to wear a girl's top :P) |
| 20:15 | <MikeSmith> | gsnedders: chaps |
| 20:16 | <MikeSmith> | plus, your cane |
| 20:16 | <Ms2ger> | Also, don't listen to Mike |
| 20:16 | <MikeSmith> | the one that has the glove with the live goldfish inside |
| 20:16 | <AryehGregor> | Hixie, I don't see how that concern is relevant to your definition. The "root element of a document" is its first element child, while the "root element of a node" when that node happens to be a document is the document itself. |
| 20:16 | <AryehGregor> | What does that have to do with fragments? |
| 20:18 | <gsnedders> | Ms2ger: Life is more interesting when you listen to MikeSmith |
| 20:18 | <Ms2ger> | I won't argue with that |
| 20:19 | <erlehmann> | hahahaha http://my.opera.com/haavard/blog/2011/01/13/openness#comment52158072 |
| 20:19 | <erlehmann> | >maybe Google should offer a free version of Chrome with WebM+Theora and a paid one with +H.264 |
| 20:21 | <TabAtkins> | oojacoboo: What you described (different sites for mobile) *was* the case in the past. These days, it's thankfully rarely necessary. It definitely won't get worse again. |
| 20:22 | <AryehGregor> | If it's rarely necessary, why does everyone and their dog do it? |
| 20:22 | <oojacoboo> | TabAtkins: current sites suck |
| 20:22 | <AryehGregor> | Including, e.g., lots of Google sites? |
| 20:22 | <oojacoboo> | unless they are custom developeed |
| 20:22 | <Ms2ger> | Why does everyone and their dog use <center>? |
| 20:22 | <Ms2ger> | Including, e.g., lots of Google sites? |
| 20:22 | <oojacoboo> | Ms2ger: they don't know better! |
| 20:22 | <TabAtkins> | AryehGregor: Dont' hyperbolize. The vast majority of sites do not have a mobile version. |
| 20:23 | <oojacoboo> | my business partner used it the other day, backend guy Ms2ger |
| 20:23 | <oojacoboo> | !! |
| 20:23 | <erlehmann> | easy solution: kill your business partner. |
| 20:23 | <AryehGregor> | TabAtkins, yes, and they all look horrible on mobile. |
| 20:23 | <AryehGregor> | I was talking about the ones that actually work decently on mobile. |
| 20:23 | <AryehGregor> | And use nontrivial layout. |
| 20:24 | <oojacoboo> | erlehmann: very good at what he does, just not markup, I handle all that |
| 20:24 | <AryehGregor> | (aryeh.name works pretty well on mobile, no kidding.) |
| 20:24 | <TabAtkins> | AryehGregor: I find that my mobile browser handles most sites great. |
| 20:24 | <erlehmann> | i use mobile versions of okcupid and twitter because they work faster and aren't that bloated. |
| 20:24 | <AryehGregor> | TabAtkins, I find that they often waste tons of space and I have to either zoom in to cut out the garbage around the sides, or scroll way too much, or (occasionally) just suffer. |
| 20:24 | <oojacoboo> | erlehmann: yes, the experience is hands down better, and takes us one step farther from getting off the app bandwagon! |
| 20:24 | <oojacoboo> | that's why apps are so popular |
| 20:24 | <AryehGregor> | This is on the Android browser. |
| 20:25 | <oojacoboo> | but, if you could allow UAs to use parts of your site in the UA, you could really start doing some things |
| 20:25 | <TabAtkins> | AryehGregor: Get a better browser. I can just double-tap what I want to read and it auto-zooms and changes the width to work. |
| 20:25 | <oojacoboo> | the <nav> is an amazing example of this |
| 20:25 | <oojacoboo> | <menu> is another great example |
| 20:26 | <oojacoboo> | it's a shame that THE best thing about html5 is being overlooked! |
| 20:26 | <oojacoboo> | otherwise, it's nothing more than new element names, who gives a flying **#*@(! |
| 20:26 | <AryehGregor> | TabAtkins, what browser are you using? |
| 20:26 | AryehGregor | hasn't tried double-tapping except by accident |
| 20:26 | <TabAtkins> | Android. |
| 20:26 | <TabAtkins> | On the most recent OS. |
| 20:26 | <AryehGregor> | oojacoboo, . . . you don't actually realize how many things are part of HTML5, I think, even narrowly construed. |
| 20:27 | <erlehmann> | oojacoboo, html5 is new element names, haahaha. |
| 20:27 | <oojacoboo> | TabAtkins: you are of the camp that believes in mediocrity |
| 20:27 | AryehGregor | will try double-tapping next time. |
| 20:27 | <TabAtkins> | oojacoboo: Um, what? |
| 20:27 | <AryehGregor> | oojacoboo, HTML5 is a multi-hundred-page spec that took years to write and is still being painstakingly refined. The new semantic elements are a more or less trivial part of it. |
| 20:27 | <erlehmann> | oojacoboo, you are of the camp that believes in trolliocrity |
| 20:27 | <oojacoboo> | AryehGregor: I do, but they don't have the ability to shape the way we use the internet like mobile does |
| 20:28 | <AryehGregor> | Well, actually, yes they do. |
| 20:28 | <oojacoboo> | AryehGregor: they are trivial b/c you make them as such |
| 20:28 | <oojacoboo> | AryehGregor: 5 years from now, I bet 60% of internet traffic is mobile |
| 20:29 | <erlehmann> | 5 years from now, I bet 60% of porn is WebM. |
| 20:29 | <oojacoboo> | you know why iOS app store is exploding.. b/c of that realization and the UA experience isn't good eough |
| 20:29 | <oojacoboo> | html5 has the potential to fix this issue! |
| 20:29 | <oojacoboo> | and it's huge |
| 20:29 | <erlehmann> | yay html5! |
| 20:29 | <erlehmann> | html5 is huge! |
| 20:30 | <oojacoboo> | yea, well, the only way it will happen is if there is a movement |
| 20:30 | <erlehmann> | (hundreds of pages that took hundreds of programmer slaves hundreds of man-years to write) |
| 20:30 | <oojacoboo> | it won't just happen on it's own... what will happen are these javascript frameworks :/ |
| 20:30 | <erlehmann> | join the html5 movement! |
| 20:30 | <erlehmann> | hey, i use jquery, you insensitive clod! |
| 20:31 | <oojacoboo> | so do I ;) |
| 20:31 | <erlehmann> | also, i wrote a canvas animation library for 2d gravity simulations. |
| 20:32 | <oojacoboo> | anyway, I am around, I have input, but no need to blast my take on things now/here |
| 20:32 | <erlehmann> | i don't expect anyone to use canvas primitives. a fine framework is the love of other programmers frozen in code. |
| 20:44 | <MikeSmith> | othermaciej: fyi http://www.w3.org/News/2011#entry-8990 |
| 20:45 | <othermaciej> | MikeSmith: neat |
| 20:45 | MikeSmith | heads off to get some breakfast |
| 20:55 | TabAtkins | still wonders how he's in the mediocre camp. |
| 20:59 | <othermaciej> | mediocre camp? |
| 21:01 | <TabAtkins> | oojacoboo said I believe in mediocrity. I'm not sure what it was in reference to, though, so I'm not sure how much I should be offended. |
| 21:04 | <Hixie> | AryehGregor: the root element is always an element node |
| 21:04 | <AryehGregor> | Hixie, "The term root element, when not explicitly qualified as referring to the document's root element, means the furthest ancestor element node of whatever node is being discussed, or the node itself if it has no ancestors." |
| 21:04 | <AryehGregor> | That tells me that if it's not an element node and has no ancestors, it's its own (non-element) root element. |
| 21:05 | <AryehGregor> | What's the root element of document.createTextNode("")? |
| 21:05 | <Hixie> | oh, i see what you're saying |
| 21:05 | <Hixie> | does anything refer to the root element of a Document node? |
| 21:06 | <AryehGregor> | Well, my extend() spec does, for one. |
| 21:06 | <Hixie> | ah :-) |
| 21:06 | <AryehGregor> | Behavior differs if the current selection's root element isn't the same as the target's. |
| 21:06 | <Hixie> | well i agree that the root element of a Document should be its documentElement and not itself |
| 21:06 | <Hixie> | it's possible i misuse the term |
| 21:07 | <AryehGregor> | Other parts of DOM Range also refer to the root element of an arbitrary Node. |
| 21:07 | <Hixie> | k, let's update the definition |
| 21:07 | <Hixie> | one sec |
| 21:09 | <mven> | extenze() |
| 21:09 | <Hixie> | AryehGregor: ok regenning with a fix |
| 21:10 | <AryehGregor> | Hixie, you're regenning the Web DOM Core spec, or some other spec that also contains that definition? |
| 21:10 | <Hixie> | the html spec |
| 21:10 | <Hixie> | i don't edit dom core :-) |
| 21:10 | <AryehGregor> | So they have different definitions of the same thing? One should reference the other. |
| 21:10 | <AryehGregor> | Where's the definition in HTML5? |
| 21:11 | <Hixie> | Terminology > DOM Trees |
| 21:12 | <AryehGregor> | I think what I actually want for my definition is "root node", not "root element". |
| 21:12 | <AryehGregor> | Maybe I'll make up a definition. |
| 21:21 | <jgraham> | AryehGregor: It is not impossible that the bug you have seen is the one where some argument to a template is not explictly converted to a string |
| 21:21 | <jgraham> | This is because the template system is utterly braindead |
| 21:22 | <jgraham> | Which is entirely my fault |
| 21:22 | <jgraham> | (it is also kind of hard to debug, as you noticed) |
| 21:22 | <jgraham> | Of course it might not be that at all |
| 21:23 | <TabAtkins> | Yay, I think I've got a pretty good outline going on my blog. |
| 21:23 | <annevk> | AryehGregor, Hixie, DOM Core editors expect HTML5 to eventually reference DOM Core for this definition |
| 21:26 | <jgraham> | gsnedders: You really should go out with a top hat and cane. Then you can recreate scenes like http://questionablecontent.net/view.php?comic=1425 |
| 22:44 | <dglazkov> | Hixie: have you visited http://wiki.whatwg.org/wiki/Component_Model_Use_Cases lately? It's all grown up! |
| 22:44 | <dglazkov> | Still lots of work to do, but should be more useful than before. |
| 23:50 | <Hixie> | dglazkov: interesting |
| 23:51 | <Hixie> | in other news, http://wiki.whatwg.org/wiki/Change_Proposal_for_ISSUE-129 is down to 11 sections to do remaining |
