| 01:28 | <[bjoern]> | http://www.websitedev.de/temp/canvas-heatmap-2.html -- Should this be part of a circle with a solid blue color in the upper left corner (QtWebkit), a colorful circle somewhere on the canvas (Firefox nightly), or a large colorful rectangle (Opera)? |
| 02:54 | <MikeSmith> | [bjoern]: Safari and Chrome on OSX vote for making it the large colorful rectangle |
| 02:56 | <[bjoern]> | That suits me, I'll claim that's the correct behavior then. Thanks. |
| 03:32 | <MikeSmith> | http://twitter.com/azaaza/status/15383718153 |
| 03:32 | <MikeSmith> | "Did you know Apple has a Web Apps directory? You have to be an ADC Premier member to submit. http://www.apple.com/webapps/"; |
| 03:33 | <paul_irish> | it's the web app store that gets no traffic! |
| 03:33 | <paul_irish> | \o/ |
| 03:35 | <MikeSmith> | :) |
| 03:36 | <MikeSmith> | http://www.apple.com/webapps/utilities/html5reference.html |
| 03:39 | <miketaylr> | link to: http://www.biocourseware.com/iphone/html5/index.htm#1 ? |
| 03:40 | <miketaylr> | bogus, totally isn't working in any browsers on this mac :/ |
| 03:41 | miketaylr | failed to realize its an iphone webapp :( |
| 03:41 | miketaylr | goes away |
| 04:03 | <MikeSmithX> | http://twitter.com/emilbjorklund/status/15328003664 |
| 04:03 | <MikeSmithX> | "Second annoyance of the day: the html5 spec links you get from the validator crashes firefox." |
| 04:03 | <MikeSmithX> | should have v.nu link to the multipage version of the spec |
| 06:51 | <zcorpan_> | awesome how http://www.apple.com/html5/ refuses to show the demos in any other browser than safari |
| 06:52 | <zcorpan_> | including chrome |
| 06:56 | <Peter`> | A few months from now they'll have a fair reason to do that with <video>, being the only major browser which doesn't support WebM ;-) |
| 06:56 | <zcorpan_> | i assume chrome is still going to support h.264 |
| 06:57 | <Peter`> | As will IE 9 :) Loading the full URL of a test in Chrome doesn't work either ( http://www.apple.com/html5/showcase/typography/ ), redirects to the overview page |
| 06:57 | <Peter`> | Ironically the sliders are broken in Safari |
| 07:48 | <MikeSmith> | wtf |
| 07:48 | <MikeSmith> | http://www.w3.org/TR/MathML/chapter3.html#id.3.5.1.2 |
| 07:48 | <MikeSmith> | allowed values are (top | bottom | center | baseline | axis) [ rownumber ] |
| 07:49 | <MikeSmith> | where [ rownumber ] is an optional number that can be appended to "top", "bottom", etc.? |
| 07:49 | <MikeSmith> | e.g., top3 ? |
| 07:49 | <MikeSmith> | can that really be? |
| 07:50 | <MikeSmith> | attribute align {xsd:string { |
| 07:50 | <MikeSmith> | pattern ='\s*(top|bottom|center|baseline|axis)\s*[0-9]*'}}? |
| 07:51 | <MikeSmith> | wow |
| 08:03 | <boblet> | sounds funky |
| 08:03 | <boblet> | gsnedders’ outliner doesn’t show implicit sectioning does it |
| 08:06 | <boblet> | hrm… still not sure if I understand implied sections :/ |
| 08:07 | <zcorpan_> | <h1>foo</h1><h2>hi, i'm in an implied section</h1> |
| 08:08 | <boblet> | d’oh! gsnedders’ outliner does show implied sections, but only those without typos |
| 08:08 | <boblet> | obviously the mind-reading update hasn’t happened yet |
| 08:09 | <boblet> | zcorpan_: I kinda understand the theory, but I’m not *feelin’ it* ye |
| 08:09 | <boblet> | yet |
| 08:10 | <boblet> | it’s basically an error correction mechanism, right? for UGE |
| 08:10 | <boblet> | (user-generated errors) |
| 08:13 | <hsivonen> | MikeSmith: curious that they allow whitespace before but not after |
| 08:13 | <hsivonen> | not after the number |
| 08:13 | <hsivonen> | but before the text |
| 08:15 | <MikeSmith> | hsivonen: the more I see of that vocabulary the more it scares me |
| 08:16 | <boblet> | another outlining q: using <body><article><h1>title</h1>… gives a root “Untitled section”. wassup with that? we’re meant to add a heading element outside sectioning content? I don’t get where it’s coming from |
| 08:16 | <MikeSmith> | anyway, it seems clear there are some more deficiencies in that MathML 2.0 rng schema |
| 08:17 | <MikeSmith> | I'm wondering if it would be easier to just start with the MathML 3.0 one and remove stuff |
| 08:17 | <MikeSmith> | maybe I should ping David Carlisle and ask |
| 08:17 | <jgraham> | boblet: Implied sections are basically HTML4 |
| 08:18 | <jgraham> | boblet: The <body> element is also a sectioning element |
| 08:18 | <boblet> | jgraham: aah! I forgot about sectioning root. doh |
| 08:18 | <boblet> | thanks |
| 08:18 | <jgraham> | So if you have <body><h1>My funky site</h1><article><h1>My article everything works |
| 08:19 | <boblet> | outlining algorithm needs to be communicated by mind-meld. explaining it in writing is brutal |
| 08:20 | <hsivonen> | MikeSmith: have you filed the scary stuff as spec bugs against MathML 3? |
| 08:20 | <MikeSmith> | no |
| 08:20 | <MikeSmith> | I guess I should |
| 08:20 | <MikeSmith> | I just wonder if any of this stuff is actually supported |
| 08:21 | <MikeSmith> | I mean, what does align=baseline3 even mean? |
| 08:21 | hsivonen | has no idea |
| 08:25 | <MikeSmith> | hsivonen: btw, you got my question from the other day about hooking v.nu HTML5 parser into other java apps? |
| 08:25 | <MikeSmith> | http://krijnhoetmer.nl/irc-logs/whatwg/20100602#l-181 |
| 08:26 | <hsivonen> | MikeSmith: I've missed that bit in the logs |
| 08:27 | <hsivonen> | MikeSmith: you need a DocumentBuilderFactory that instantiates HtmlDocumentBuilder |
| 08:27 | <MikeSmith> | OK |
| 08:27 | <hsivonen> | MikeSmith: I haven't supplied one, because I wasn't aware of people actually needing one |
| 08:28 | <MikeSmith> | hsivonen: I was thinking it would be useful for anybody who wants to just use the HTML5 parser in place of whatever existing XML one |
| 08:28 | <MikeSmith> | e.g., with Saxon |
| 08:28 | <hsivonen> | I thought they'd just do new HtmlDocumentBuilder |
| 08:28 | <MikeSmith> | or jing or whatever |
| 08:28 | <hsivonen> | I didn't realize people actually want the factory stuff |
| 08:28 | <MikeSmith> | I mean using it from the command-line |
| 08:28 | <hsivonen> | but yeah, that would be doable |
| 08:29 | <MikeSmith> | OK |
| 08:29 | <MikeSmith> | should I open a bug for it? |
| 08:29 | <hsivonen> | MikeSmith: yes, please |
| 08:29 | <MikeSmith> | ok |
| 08:31 | <jgraham> | TabAtkins: Surely there is a combining overbar character somewhere in unicode? |
| 08:33 | <zcorpan_> | jgraham: with multiple overbars? |
| 08:34 | <MikeSmith> | hsivonen: what sax, what would the corresponding SAXParserFactory need to instantiate? |
| 08:35 | <MikeSmith> | sax/HtmlParser ? |
| 08:40 | <hsivonen> | yeah |
| 08:45 | <MikeSmith> | ok |
| 08:47 | <boblet> | does “sectioning element” mean anything specific in the spec? I’d like to use it as a reference for sectioning content and sectioning root elements, but wonder if it’s already being used to mean something else… |
| 08:49 | <jgraham> | zcorpan_: Maybe not |
| 08:49 | <jgraham> | (but you would have thought there would be if you need it for roman numerals) |
| 08:49 | <boblet> | can’t find it in the spec, so maybe I’m safe |
| 08:50 | <boblet> | roman numerals? |
| 08:51 | <boblet> | oh, you mean I should have said “as shorthand for…”? |
| 08:52 | <jgraham> | boblet: I was replying to zcorpan_ |
| 08:52 | <boblet> | jgraham: haha, d’oh |
| 08:53 | <boblet> | orz |
| 10:56 | <zcorpan_> | 'interact with web standards' just arrived to my door |
| 11:13 | <sid123> | hi I wanted to know about the storage limit of web data bases... There is a size parameter that is provided to the Database initilaization... But is that followed strictly or are there limits on it in any way? |
| 11:32 | <Lachy> | wow. What a great way to promote HTML5 and web standards, by requiring the use of Safari to view any of their demos http://www.apple.com/html5/ |
| 11:32 | <Lachy> | </sarcasm> |
| 11:32 | <zcorpan_> | old news |
| 11:32 | <Lachy> | well, new to me. I only just read my news feeds |
| 11:33 | <boblet> | adactio: thanks for the email (was starting to wonder if I had the right addy :P ) |
| 11:33 | <adactio> | boblet: yeah, really sorry it took me so long to respond. :-( |
| 11:33 | <boblet> | heh |
| 11:33 | <boblet> | adactio: not like you’ve been up to much either :| |
| 11:34 | <adactio> | boblet: ;-) |
| 11:34 | <boblet> | thanks for the feedback tho, really appreciated (just finished a big screed on sectioning root today) |
| 11:37 | <hsivonen> | Lachy: I'm rather amazed that whoever put the browser sniffing in place didn't notice the mismatch with the message and readjust |
| 11:38 | <hsivonen> | or then, the message isn't really about standards but about Safari vs. Flash |
| 11:38 | <hsivonen> | not even about WebKit vs. Flash |
| 11:38 | hsivonen | hasn't yet bothered to examine the demos in Safari to see how the VR demo works |
| 11:42 | <hsivonen> | ah. CSS 3D Transforms |
| 11:43 | <hsivonen> | at least their developer site has the decency to call them "Safari Technology Demos" |
| 11:43 | <Rik`> | in the video example, they're using a CSS mask. It's not yet in any draft, is it ? |
| 11:44 | <zcorpan_> | it's web standards!1 |
| 11:45 | <boblet> | bbl |
| 11:46 | <hsivonen> | more clever than dealing with a spherical mapping |
| 12:22 | <hsivonen> | where can I find a signed Chrome extension? |
| 12:23 | <Peter`> | Signed how? |
| 12:24 | <hsivonen> | Peter`: that's what I want to find out |
| 12:24 | <Peter`> | they're just zip files with a certificate to generate an unique hash |
| 12:24 | <hsivonen> | Peter`: where's the cert? |
| 12:25 | <Peter`> | hsivonen: On the computer of the person who packed it, as a .pem file |
| 12:25 | <Peter`> | You can rename .crx to .zip and it'll work as well |
| 12:25 | <hsivonen> | Peter`: where's the signature? |
| 12:26 | <Peter`> | I think it's only used to generate the extension hash, used for auto-updating and such. It's not in the archive either |
| 12:27 | <hsivonen> | Peter`: ah. that doesn't match what the intro video on youtube said |
| 12:28 | <Peter`> | I could be wrong, I'm not a Chrome developer, just somewhat familiar with the code. Best place to ask probably would be #chromium |
| 13:00 | <hsivonen> | well, I just asked on the whatwg list |
| 13:25 | <boblet> | hey all, more questions on outlining algorithm — what’s it’s raison d'être? does it assist UAs to provide direct access? is it to assist browsers in DOM reflow? something else? |
| 13:25 | <boblet> | I feel like I understand it without really getting it |
| 13:27 | <hsivonen> | boblet: the primary point is allowing browsers to provide UI for navigating directly a section or getting an overview of teh document structure |
| 13:27 | <hsivonen> | boblet: mainly for screen readers |
| 13:27 | <hsivonen> | boblet: the secondary point is styling based on outline depth |
| 13:28 | <hsivonen> | boblet: which depends on whether browser vendors and the CSS WG add a way of selecting on outline depth |
| 13:28 | <hsivonen> | if such a selector proves to be infeasible, well... |
| 13:41 | <boblet> | hsivonen: sorry — story time. thanks for your explanation |
| 13:43 | <boblet> | so for sectioning root elements like figure and blockquote, is it expected they’ll contain large chunks of content with headers, or that UAs will provide access to sectioning root outlines? |
| 13:44 | <boblet> | I’m not sure why sectioning root outlines don’t contribute to the main outline — perhaps because they normally lack a heading? |
| 13:49 | <hsivonen> | boblet: the idea is to hide quoted stuff from the main outline |
| 13:52 | <boblet> | hsivonen: quoted stuff (blockquote/figure) and generally non-outline stuff (details, fieldset, td) that sometimes contains headings, right? |
| 13:55 | <hsivonen> | boblet: right |
| 13:58 | <boblet> | hsivonen: ok, thanks for your explanation. I was worried there might have been something else that I wasn’t getting, but it appears not |
| 15:58 | TabAtkins | is sad whenever he sees attributes which accept true/false. |
| 15:58 | <TabAtkins> | Rather than either being binary or giving the values better names for future extensibility. |
| 16:00 | <Philip`> | true/false allows plenty of room for extensibility |
| 16:00 | <Philip`> | http://thedailywtf.com/Articles/What_Is_Truth_0x3f_.aspx |
| 16:00 | <TabAtkins> | true/false/file-not-found |
| 17:14 | <asmodai> | lol |
| 17:14 | <asmodai> | http://i.imgur.com/cT08B.png |
| 17:18 | <AryehGregor> | Is html5test.com actually a competent testing service? |
| 17:18 | <AryehGregor> | Some of the tests look bogus to me. |
| 17:19 | <AryehGregor> | Like testing for specific audio and video codecs. |
| 17:20 | <AryehGregor> | And it reports that Chrome dev channel doesn't support the autocomplete attribute, which is nonsense -- browsers have all supported that for years, haven't they? |
| 17:20 | AryehGregor | doesn't trust it |
| 17:20 | <AryehGregor> | Safari and Chrome have practically identical HTML5 conformance if you compare comparable versions, because they both use WebKit. |
| 17:20 | <TabAtkins> | We do different things on top of that base, though. |
| 17:21 | <TabAtkins> | Some things show up in Chrome before we push them all the way back to Webkit, frex. |
| 17:21 | <AryehGregor> | Not ones that are relevant to almost any of the features that html5test.com supposedly test. |
| 17:21 | <TabAtkins> | And presumably the same sometimes happens with Safari. |
| 17:21 | <AryehGregor> | A few, yes, but not many. |
| 17:21 | <TabAtkins> | Yeah, wouldn't know about that. |
| 17:37 | <Philip`> | It seems the popularity of a test suite is very highly correlated with how easy it is for a random user to run it in their own browser and see a big clear interestingly-designed page indicating the results |
| 17:38 | <Philip`> | and very loosely correlated with the quality or range of its test cases |
| 17:38 | <TabAtkins> | That makes sense. Random users have no idea how to evaluate the quality of tests or their coverage. |
| 17:39 | <TabAtkins> | They *can* evaluate whether or not it runs, and how pretty a results page is. |
| 17:42 | <Philip`> | (Popularity isn't a particularly important metric for a test suite, though) |
| 18:00 | <boblet> | is there a section on design principles in WHATWG spec? similar to http://www.w3.org/TR/html-design-principles/ |
| 18:08 | <Philip`> | boblet: Those principles were mostly just writing down the principles that had been used in the WHATWG process |
| 18:08 | <Philip`> | so there's no difference |
| 18:09 | <boblet> | Philip`: thanks |
| 18:09 | <Philip`> | (They weren't written down explicitly before then) |
| 18:09 | <Philip`> | (mostly since they were already obvious to most WHATWG members) |
| 18:34 | <dglazkov> | TabAtkins: btw, we don't have a fork of WebKit. All development for Chromium features happens on WebKit ToT. No pushing is involved :) |
| 18:37 | <TabAtkins> | Ah, I am mistaken then. |
| 18:42 | <boblet> | regarding the accessibility design principle, would you say that “Design features to be accessible to users with disabilities” could be accurately summarised in “accessibility by default”? |
| 18:43 | <Dashiva> | I would say so, but it's not entirely uncontroversial |
| 18:44 | <boblet> | Dashiva: nothing in a11y ever is :) |
| 18:44 | <TabAtkins> | boblet: Subtly different. |
| 18:44 | <Dashiva> | There's the "bolt-on accessibility is a sign of design failure" and "bolt-on accessibility is the way to great accessibility" |
| 18:45 | <TabAtkins> | The former could mean doing something that *can* be accessible if authors take some effort. The latter would avoid doing that in favor of something that is, well, accessible by default. |
| 18:45 | <boblet> | TabAtkins: oh? uhoh |
| 18:46 | <TabAtkins> | Ideally we want to avoid requiring any author interaction whatsoever to acheive good accessibility. |
| 18:46 | <TabAtkins> | That isn't quite possible, so instead we just minimize the amount of author interaction necessary. |
| 18:46 | <boblet> | well I sense a trend towards things being more accessible by default, eg moving attribute-based meta info into main content. but I don’t see anything so nebulous in the design principles :) |
| 18:46 | <boblet> | aah |
| 18:47 | <Dashiva> | Where possible, trick authors into writing accessible content without realizing it |
| 18:47 | <TabAtkins> | Got it, Dashiva. ^_^ |
| 18:47 | <boblet> | :/ |
| 18:48 | <boblet> | actually… Dashiva, can I quote you on that? |
| 18:48 | <boblet> | ;) |
| 18:48 | <TabAtkins> | Simple, widely-attested truth - the amount of accessible content decreases exponentially with the amount of effort required. |
| 18:48 | <TabAtkins> | (Well, maybe not exponentially - I don't have any numbers to attest that. But super-linear at least.) |
| 18:49 | <boblet> | no doubt |
| 18:49 | <Dashiva> | Re: moving attribute-based meta info into main content, that's not primarily for accessibility, that's because hidden metadata has problems in itself |
| 18:50 | <Dashiva> | But accessibility-targeted hidden metadata is has even more problems |
| 18:50 | <Dashiva> | -is |
| 18:50 | <boblet> | I was referring to previously a11y-focussed metadata like whatever-it-is on table |
| 18:50 | <Dashiva> | Since it's not only hidden, but it's basically non-existent to users who don't need the extra accessibility |
| 18:52 | <TabAtkins> | Also: people don't know how to write good @summary text. And auto-generated @summary is useless - if it can be autogenerated by the producer, it can be auto-generated by the consumer equally well. |
| 18:52 | <TabAtkins> | And then the consumer *knows* that it's auto-generated and can treat it differently than custom-crafted text. |
| 18:53 | <boblet> | yeah, I completely agree with the idea |
| 18:53 | <boblet> | did the issue about that get resolved? I should check… |
| 18:56 | <boblet> | review huh |
| 18:58 | <AryehGregor> | http://beta.html5test.com/ actually looks pretty good. |
| 18:59 | <AryehGregor> | Blech, it tests <device>, is that even specced yet? |
| 18:59 | <TabAtkins> | It *exists* in a spec. |
| 18:59 | <boblet> | I hope someone starts graphing score over time in one of these sometime |
| 18:59 | <TabAtkins> | It's not "specced" though. |
| 19:16 | <AryehGregor> | http://damowmow.com/playground/htmlwg/chart.html "Last updated NaN days ago." |
| 19:21 | <TabAtkins> | var updateStr = ' Last updated ' + (age < 1 ? 'today' : age.toFixed(0) + ' day' + (age.toFixed(0) == 1 ? '' : 's') + ' ago') + '.'; |
| 19:23 | <Dashiva> | Good-ole' NaN |
| 19:29 | <TabAtkins> | The problem is the toFixed(0) call. It appears to give "NaN" as the result for any number. |
| 19:30 | <TabAtkins> | At least, here in Chrome. |
| 19:31 | <TabAtkins> | Firefox does it right. |
| 19:31 | <TabAtkins> | So, looks like a bug on our end. |
| 19:35 | <AryehGregor> | The display doesn't work for me either. |
| 19:36 | <TabAtkins> | Yeah, me neither. |
| 19:36 | <TabAtkins> | What do you want to bet it's toFixed again? |
| 19:36 | <Dashiva> | It errors in Opera on a lineTo call |
| 19:37 | <AryehGregor> | Is this test completely broken, or is it just me? http://samples.msdn.microsoft.com/ietestcenter/html5/selection/select.htm |
| 19:38 | <Dashiva> | There's a NaN object in there |
| 19:38 | <Dashiva> | Hum |
| 19:39 | <AryehGregor> | Looking at the source code, it looks as though it selects the contents of the input, then checks whether the contents are *not* selected, or something similarly crazy. |
| 19:41 | <boblet> | I’m wondering what the big a11y changes in HTML5 are. I know alt text, longdesc, table summary, plus the current video/canvas and interactive element (progress, details etc) issues. anything else I’m missing? |
| 19:41 | <Dashiva> | It looks fine to me, where in the code is the problem? |
| 19:42 | <AryehGregor> | Dashiva, it does input1.select(), then checkDefaultSelectionAttributes(). The latter checks things like null != selection.anchorNode. Why should selection.anchorNode be null if you just selected the input? |
| 19:45 | <TabAtkins> | boblet: Not an expert, but I think that's it. |
| 19:45 | <boblet> | TabAtkins: thanks. more sanity-checking than anything :) |
| 19:47 | <Dashiva> | AryehGregor: Those don't apply to selection in input fields |
| 20:13 | <AryehGregor> | Dashiva, where does the spec say that? It says that text fields have their own selections, but not that selecting text in text fields doesn't also select it in the browsing context. Am I missing something? |
| 20:14 | <AryehGregor> | It just says that the UA should allow the user to change the selection. |
| 20:14 | <AryehGregor> | "This one selection must be shared by all the content of the browsing context (though not by nested browsing contexts), including any editing hosts in the document."\ |
| 20:14 | <AryehGregor> | s/\\// |
| 20:15 | <AryehGregor> | Firefox and WebKit seem to have the more plausible interpretation to me. |
| 20:27 | <Dashiva> | AryehGregor: Firefox seems weird to me |
| 20:34 | <AryehGregor> | Ugh, faked browser UI in Chrome: http://www.huffingtonpost.com/stewart-abercrombie-baker/smallpox-in-the-garage_b_596340.html |
| 20:34 | <AryehGregor> | Evil. |
| 20:34 | <AryehGregor> | They could have done better by using position: fixed. |
| 20:34 | <Dashiva> | AryehGregor: Looks to me like Firefox claims there's a zero-length selection at a random location in the document... |
| 20:35 | <Dashiva> | When there's no selection at all |
| 20:35 | <AryehGregor> | Dashiva, well, the spec says that the UA should provide a means for the user to change the selection, not that it has to make sense. Clearly, this is merely a quality-of-implementation issue, and everyone passes. |
| 20:36 | <Dashiva> | The spec also says "Returns null if there's no selection." |
| 20:36 | <Dashiva> | Which firefox does not |
| 20:37 | <AryehGregor> | Well, perhaps Firefox has simply chosen to interpret page load as a user choice of selection. |
| 20:37 | <AryehGregor> | This is what happens when you leave UI totally undefined, am I right? :) |
| 20:38 | <Dashiva> | If at any point you (or a script) have selected anything, it never forgets |
| 20:38 | <Dashiva> | E.g. select some random text, remove the selection, and it claims there's a zero-length selection at the start of the ex-selection |
| 20:39 | <Dashiva> | If there is no ex-selection (such as in that test), it picks body for some reason |
| 20:40 | <AryehGregor> | Whee. |
| 20:40 | <Dashiva> | Never mind |
| 20:40 | <Dashiva> | It picks the parent of the input the script selected |
| 20:41 | <Dashiva> | But the offsets are 1, not the offset of the input |
| 20:42 | <Dashiva> | http://dashiva.net/test/selection.html |
| 21:21 | <dglazkov> | nobody wants to stay in touch on LinkedIn. Well said. |
| 21:23 | <TabAtkins> | Haha, yes. |
| 21:28 | <AryehGregor> | What? |
| 21:28 | <TabAtkins> | On the whatwg list. |
| 21:28 | <AryehGregor> | Ohl |
| 21:28 | <AryehGregor> | Oh. |
| 21:59 | <TabAtkins> | Oh man, using unicode box-drawing characters when doing ascii illustrations is 1000% sexier than using +-|. |
| 22:00 | <TabAtkins> | Though I suppose it no longer qualifies as "ascii". ^_^ |
| 22:00 | <Dashiva> | Real men use 1x1 pixel divs |
| 22:01 | <TabAtkins> | Real men use tables with background-colors on the cells. |
| 22:01 | <Dashiva> | But that isn't _semantic_ |
| 22:01 | <TabAtkins> | Box-drawing characters are the answer, then. Just wrap it in a <pre> and you're golden. |
| 22:02 | <TabAtkins> | (Or just send plain-text emails like I do and hope everyone uses a fixed-width font to view emails.) |
| 22:02 | <Philip`> | TabAtkins: Tables like in http://blogs.msdn.com/b/oldnewthing/archive/2010/05/26/10015253.aspx ? |
| 22:02 | <TabAtkins> | Yes, exactly like that. |
| 22:02 | <Philip`> | Gmail ought to automatically detect ASCII art and switch to a fixed-width font |
| 22:04 | <TabAtkins> | That wouldn't be overly hard, I bet. |
| 22:05 | <Dashiva> | Your mission, should you choose to accept it, is to produce a non-edge case that would trigger a false positive |
| 22:05 | <Philip`> | TabAtkins: I will rely on you and await the feature eagerly |
| 22:05 | <TabAtkins> | What's the easiest way to add new key combos on linux? I know on Windows you can just produce a new keyboard layout file. |
| 22:05 | <TabAtkins> | Philip`: I'll send the suggestion upstream. |
| 22:05 | TabAtkins | wants to put the box-drawing characters on his keyboard. |
| 22:05 | <Dashiva> | Probably depends on the WM |
| 22:06 | <Dashiva> | There are like four layers of stuff between the app and the keyboard |
| 22:06 | <TabAtkins> | That's dumb. |
| 22:06 | <TabAtkins> | The WM shouldn't be taking part in this. >_< |
| 22:06 | <Dashiva> | I'm trying to stay blissfully unaware, but I stumbled across it via a discussion about some laptop key used to switch displays or something |
| 22:07 | <TabAtkins> | Maybe XKB will work? |
| 22:09 | AryehGregor | finally gets the LinkedIn joke |
| 22:14 | <AryehGregor> | It's probably not the WM, it's probably the desktop environment. |
| 22:14 | <AryehGregor> | Which certainly has to take part in things like shortcut keys, although it'd be nice if that sort of thing were standardized. |
| 22:17 | <jgraham> | Philip`: Popularity is an important metric for a testsuite if it causes pressure on implementors to fix specific issues |
| 22:17 | <jgraham> | e.g. Acid 3 |
| 22:17 | <jgraham> | (for better or worse) |
| 22:25 | <Philip`> | jgraham: True |
| 22:31 | <jcranmer> | TabAtkins: xkb; if you're willing to do compose, a .XCompose file lets you add even more compose keys |
| 22:36 | <TabAtkins> | Ooh, interesting. |
| 23:09 | <AryehGregor> | "Interesting" is an excellent description for how Linux works in so many cases. :) |
| 23:13 | <tabatkins> | Given <label><input type=checkbox><a href=#>foo</a></label>, I *think* that the 3.2.5.1 algo says to activate the link (and not toggle the checkbox) when you click the link. |
| 23:13 | <tabatkins> | Yes/no? |
| 23:14 | <Dashiva> | 3.2.5.1 Kinds of content? |
| 23:14 | <TabAtkins> | Sorry, 3.2.5.1.7 |
| 23:14 | <Dashiva> | ah |
| 23:15 | <TabAtkins> | It looks like you click on the <a>, the <a> is the nearest activatable element, then you run the click steps on it. |
| 23:15 | <Dashiva> | Yeah |
| 23:33 | <TabAtkins> | AryehGregor: Any way to trigger a monospace font in mediawiki? |
| 23:33 | <AryehGregor> | <span style="font-family:monospace">, <tt>, <code>, . . . |
| 23:33 | <AryehGregor> | You can also begin a line with a space. |
| 23:33 | <TabAtkins> | Danke. |
| 23:34 | <TabAtkins> | Ah, cool. |
| 23:34 | <TabAtkins> | Someone stupid did the examples for the box-drawing characters using a table for some reason. |
| 23:35 | <TabAtkins> | Apparently it's for some sort of compat? Whatever. <pre><code> will work as expected? |
| 23:36 | <AryehGregor> | <pre> is magic in MediaWiki. |
| 23:36 | <AryehGregor> | It means the same as <pre><nowiki>, so the contents are escaped. |
| 23:37 | <AryehGregor> | So don't use that. An initial space on a line will create a <pre>. |
| 23:37 | <TabAtkins> | kk |
| 23:37 | <AryehGregor> | That's the only way to create a normal <pre>, annoyingly. |
| 23:48 | <TabAtkins> | AryehGregor: You considered putting in a few -moz-linear-gradient() rules in the places where mediawiki uses gradients? |
| 23:48 | <AryehGregor> | TabAtkins, I didn't know it does use gradients. :) |
| 23:49 | <TabAtkins> | Well, like, the bar on top of each page uses gradients for each item. |
| 23:49 | <AryehGregor> | In the new skin (Vector) or the old one (Monobook)? |
| 23:49 | <TabAtkins> | Of course, it's just a 1x100 px png, but still, one less network request on ff. |
| 23:49 | <TabAtkins> | new one |
| 23:51 | <AryehGregor> | Well, got to go. |
| 23:52 | <AryehGregor> | Maybe I'll do it later if I remember, or you can remind me if you like. |
| 23:53 | <TabAtkins> | kk |
| 23:53 | <TabAtkins> | + |
