| 00:13 | <Hixie> | i don't understand what IE does |
| 00:14 | <Hixie> | "ht\tp:dem\o⊙ec" is treated teh same as "ht/tp:demo/o⊙ec" but "mailto:dem\o⊙ec" is treated literally |
| 00:15 | <Hixie> | how does it know that "ht\tp:" isn't a scheme? |
| 00:16 | <Hixie> | i guess \ can't be in a cheme |
| 00:16 | <Hixie> | scheme |
| 00:30 | <Hixie> | hsivonen, Philip`: do you want spaces in URLs to be allowed? Even though URI/IRI disallow it? |
| 00:37 | <Philip`> | Hixie: No, since it's too dangerous - if someone writes <a href="foo bar"> and then tells their text editor to wrap lines, the link might break without anyone noticing |
| 00:38 | <Philip`> | and if someone writes <a href="foo bar"> then it won't do what they expect |
| 00:39 | <Philip`> | Oops, yes it will, I don't know what I'm talking about, please ignore me |
| 00:41 | <Philip`> | but the first thing still seems like a reason to discourage spaces in URLs |
| 00:48 | <Philip`> | Also it would be bad for <head profile> which uses space-separated URIs |
| 00:49 | <Hixie> | fair enough |
| 01:01 | <Philip`> | Hixie: "This should not be pretty well-defined." - did you mean s/not/now/? |
| 01:04 | <Hixie> | er yes |
| 01:04 | <Hixie> | now |
| 06:01 | <mcarter> | Hixie, don't suppose you're around? |
| 07:46 | <Hixie> | mcarter: here |
| 07:51 | <annevk> | Maybe "Web URL" if we really need a separate name... "HTML URL" doesn't really scale well as it's likely we want to use this everywhere just to kep surprises for authors to a minimum |
| 07:56 | <annevk> | Hixie, your "Re: How Open is Open?" never went to the public list... |
| 07:56 | <Hixie> | oh? |
| 07:56 | <annevk> | It seems you changed the name of the list, but not the actual e-mail address :) |
| 07:57 | <Hixie> | woops |
| 07:57 | Hixie | goes to fix |
| 08:12 | <mcarter> | Hixie, when you have some time i was hoping you could take a look at this article i wrote and let me know if i'm getting anything wrong. also i put a couple questions that i wanted to print your responses to |
| 08:13 | <mcarter> | Hixie, http://orbited.org/svn/orbited/branches/0.5/articles/mcarter/0.5_world_announcement |
| 08:13 | <mcarter> | Hixie, the point is to explain websocket and also to shed some light on what we're doing with Orbited |
| 08:14 | <mcarter> | Hixie, the plan is to publish it after we've worked out the WebSocket stuff |
| 08:19 | <Hixie> | looks good to me |
| 08:20 | <Hixie> | for your first two questions -- I don't think it'll change the face of the Web per se, it's just a new tool in the growing toolbox that Web developers have at their disposal, one less reason for developers to use a proprietary platform over the Web. |
| 08:21 | <Hixie> | Having said that, it will make a lot of interactive applications like mail, IM, remote control tools, and the like, much more powerful |
| 08:21 | <Hixie> | and much easier to write. |
| 08:22 | <Hixie> | for your third question -- There's a plan to have a plan! |
| 08:22 | <mcarter> | cool |
| 08:22 | <Hixie> | As with everything in HTML5, we want to start small, and see where it goes before adding too much. |
| 08:23 | <mcarter> | ok, i'll add these answer to the article |
| 08:25 | <mcarter> | I'll probably have to change the tone of the article slightly -- I did put a lot of hype in there |
| 08:27 | <mcarter> | one thing to remember is that I'm planning on publishing this on cometdaily.com, and the point of it is to get the readership to stop thinking what they want is "comet" and really its a socket |
| 08:27 | <mcarter> | thats why i talk about comet at all, if you were wondering |
| 08:31 | <Hixie> | cool |
| 08:31 | <Hixie> | time for bed |
| 08:31 | <Hixie> | nn |
| 08:32 | <annevk> | WebSocket is not actually defined yet, right? |
| 08:32 | <mcarter> | goodnight |
| 08:32 | <mcarter> | annevk, no, not yet |
| 09:21 | <takkaria> | gah, 47 mails since yesterday? I thought public-html had become sensible |
| 09:26 | <annevk> | only when you don't look |
| 10:07 | <Dashiva> | Is there any point in the discussion at uri@? |
| 10:08 | <takkaria> | not really |
| 12:04 | <Lachy> | http://standardssuck.org/w3c-digging-the-xml-grave |
| 12:05 | annevk | goes in to fix typo and adds some tags |
| 12:05 | <annevk> | thanks Lachy |
| 12:06 | <Lachy> | what typo? |
| 12:06 | <annevk> | "insight into the its failure." |
| 12:06 | <Lachy> | oh |
| 12:08 | <annevk> | also, seems our titles are not capitalized so I removed that too |
| 12:17 | <takkaria> | if they weren't leading questions I don't know what are |
| 12:18 | <annevk> | hm? |
| 13:28 | <hsivonen> | Hixie: I'm not sure what I want Re: spaces in URLs |
| 13:28 | <hsivonen> | Hixie: I can see reasonable arguments both ways |
| 13:29 | <gsnedders> | Hixie: I don't want spaces in URLs to be conforming. |
| 13:29 | <hsivonen> | Hixie: ATM, I'm leaning on the side of not allowing spaces as conforming |
| 13:29 | <gsnedders> | Hixie: I'd argue that any URL must be a valid IRI |
| 13:30 | <hsivonen> | fwiw, I was trying to watch the latest standards suck episode with the software that Nokia ships with S60r3, but there was no way to navigate to the .3gp version of YouTube |
| 13:31 | <hsivonen> | (I don't know if YouTube chooses to encode the standards suck episodes as .3gp anyway) |
| 14:38 | <Lachy> | hsivonen, YouTube uses h.264 encoded video in a Flash video container format. |
| 14:39 | <hsivonen> | Lachy: interesting. I thought they hadn't made the transition to H.264 yet |
| 14:40 | <hsivonen> | Lachy: (also, putting h.264 in the .flv container seems weird. Do you have a reference?) |
| 14:40 | <hsivonen> | Lachy: YouTube also has a partial mobile version. the .3gp videos there play on RealPlayer shipped by Nokia. |
| 14:41 | <Lachy> | well, I just downloaded the video using Miro, opened it up in QuickTime and had a look at what codec it used. |
| 14:41 | <Lachy> | So unless Miro transcoded it, then they use h.264 |
| 14:41 | <hsivonen> | Lachy: they do use h.264 for AppleTV and iPhone |
| 14:41 | <hsivonen> | perhaps Miro pulls in those files? |
| 14:42 | <Lachy> | though, it may have done so, cause Miro put it into a .mp4 container format too. |
| 14:43 | <Lachy> | I'll download the flv with another tool and see what it has. |
| 14:43 | <hsivonen> | (it seems to me that putting h.264 into .flv wouldn't make sense, as versions of Flash plug-ins that support h.264 support it in .mp4) |
| 14:45 | <Lachy> | yeah, the video that miro downloaded is different from the flv. |
| 14:46 | <Lachy> | the FLV has h.263 with MP3 audio. |
| 14:46 | <Lachy> | and is lower quality than the MP4 that Miro downloaded |
| 14:47 | <hsivonen> | can standards suck be subscribed to with iTunes? |
| 14:47 | <Lachy> | I wonder if Miro somehow got the original video that Marcos had uploaded |
| 14:47 | <Lachy> | not yet |
| 14:51 | <hsivonen> | Are there Web formats that use namespaced attributes an the namespace isn't the XML namespace, the XMLNS namespace or the XLink namespace? |
| 14:51 | <Lachy> | oh, btw, I realised the background noise you can hear in the videos is actually the noisy air conditioner in the office |
| 14:52 | <Lachy> | I'm not sure how to effectively filter it out though |
| 14:55 | <Philip`> | hsivonen: Does <license rdf:resource="http://creativecommons.org/licenses/by-nc/1.0/"; /> etc count as "Web formats", given that it's used on the web a lot? |
| 14:59 | <hsivonen> | Philip`: it counts if it is used as an actual element in SVG metadata. It doesn't count if it is use inside comments in HTML. |
| 15:00 | <hsivonen> | Philip`: but I guess I should add 'rdf' to my hardwired attribute prefixes |
| 15:02 | <hsivonen> | (I finally decided to write an XML serializer so that I don't need to depend on someone else's bugs and design decisions) |
| 15:02 | <Philip`> | There's lots of VML stuff like v:ext |
| 15:02 | <hsivonen> | using the Xalan serializer seems to take more time than it saves |
| 15:03 | <hsivonen> | Philip`: v:ext isn't for real XML content, though. |
| 15:07 | <Philip`> | Some collection of real XML SVG uses rdf: and inkscape: and sodipodi: and i: (Illustrator) and x: (XMP Toolkit) |
| 15:08 | <hsivonen> | I also need a list of element namespaces that are canonically prefixed. rdf is one of those |
| 15:08 | <hsivonen> | Philip`: do you have a URL to a sample document? |
| 15:09 | <Philip`> | hsivonen: A sample of what in particular? |
| 15:09 | <hsivonen> | Philip`: an SVG file where I can see abundant real-world prefixes. |
| 15:10 | <hsivonen> | Philip`: those were all for attributes, right? |
| 15:10 | <hsivonen> | Philip`: what's the Illustrator ns URI? |
| 15:11 | <hsivonen> | and what's the XMP URI? |
| 15:12 | <Philip`> | http://upload.wikimedia.org/wikipedia/en/e/e4/Tetrahedral_group.svg has some i: and a: and rdf: |
| 15:12 | <hsivonen> | Philip`: thanks |
| 15:12 | <Philip`> | and it has x: |
| 15:13 | <hsivonen> | whee! adobe-proprietary URI scheme! |
| 15:13 | <hsivonen> | that document has some bad NS and metadata craziness going |
| 15:13 | <hsivonen> | so sad |
| 15:14 | Philip` | goes away |
| 15:14 | hsivonen | adds the URIs to Validator.nu |
| 15:16 | <hsivonen> | I think ns.adobe.com doesn't adhere to the TAG's advice |
| 15:18 | <hsivonen> | Soon I'll have more Adobe namespaces than W3C namespaces on file |
| 15:19 | hsivonen | lols at GenericCustomNamespace |
| 19:04 | <tusho> | What should I use to mark up some text like 'Published 2008-06-28'? |
| 19:04 | <tusho> | I guess just a <p>... |
| 19:07 | <Philip`> | You could use a <time> around the date, if you want to add markup for its own sake (or for the sake of tools that don't exist yet) |
| 19:09 | <tusho> | Right. But I have subtitles too. |
| 19:09 | <tusho> | So: |
| 19:09 | <tusho> | <h1>My Trip Out</h1><h2>or, a very boring subtitle</h2><p>Published 2008-06-28.</p> |
| 19:09 | <tusho> | However, CSS-wise, the subtitle and the published line will be styled the same and on the same line, like 'or, a very boring subtitle. Published 2008-06-28.' |
| 19:10 | <tusho> | and it just kinda feels wrong to mix a p in with that. |
| 19:10 | <tusho> | *shrug* Probably just me being islly. |
| 19:10 | Philip` | isn't any good at working out which elements to use in any given case, so he usually just picks whichever one gets rendered in the desired way :-) |
| 19:10 | <tusho> | *g* |
| 19:10 | <tusho> | I could just drop subtitles, who reads them anyway? |
| 19:10 | <tusho> | But otherwise my titles will be too long for my headings. :( |
| 19:14 | <tusho> | :\ |
| 19:26 | <tusho> | Any opinions? |
| 19:32 | <Philip`> | tusho: I don't see what the problem would be with the markup you suggested |
| 19:32 | <tusho> | True. |
| 19:32 | <tusho> | I'll go for it, then. |
| 19:32 | <Philip`> | although in the spirit of promotion HTML5 elements even though I don't know what good they would do, maybe the whole thing could be wrapped in a <header> |
| 19:32 | <tusho> | yes |
| 19:32 | <tusho> | it will be |
| 19:32 | <tusho> | :) |
| 19:33 | <Philip`> | s/promotion/promoting/ |
| 19:33 | <tusho> | I saw http://intertwingly.net/blog/2008/06/24/Minimalist-Markup and thought "well, why does it have to be minimal markup *or* design"? |
| 19:33 | <tusho> | So now I'm off in typography and markup la-la land... |
| 19:34 | <Windstoss> | I'm just trying out getElementsByClassName(). In Safari, how to use forEach for the result? In Firefox, I do Array.forEach(document.getElementsByClassName('foo'), function() {}). |
| 19:35 | <Windstoss> | Safari hast forEach implemented, so I don't want to use the traditional for loop if possible. |
| 19:41 | <tusho> | Hm, |
| 19:42 | <tusho> | is <h2> even right for a subtitle? |
| 19:42 | <tusho> | Doesn't that make this 'tree': |
| 19:42 | <tusho> | - My title |
| 19:42 | <tusho> | \- My subtitle |
| 19:42 | <tusho> | and then, under that subtitle, my published line... |
| 19:42 | <tusho> | that seems wrong |
| 21:08 | <tusho> | Hmm. |
| 21:08 | <tusho> | Can <h2> be in a <p>? |
| 21:08 | <tusho> | If so, <p><h2>My fancy subtitle.</h2> Published blah.</p> may be more correct. |
| 21:11 | <gsnedders> | tusho: No |
| 21:11 | <gsnedders> | tusho: in text/html, the <h2> creates an implied </p> |
| 21:11 | <tusho> | Hrmph. |
| 21:11 | <tusho> | Because my current way isn't really Working Out. |
| 21:11 | <gsnedders> | tusho: in XHTML, it's non-conforming |
| 21:12 | <tusho> | It's making me have to make (header h2, header p) inline. |
| 21:12 | <tusho> | :| |
| 21:20 | <tusho> | Anyone have markupy suggestions? |
| 21:20 | <tusho> | I already have a <header> including a <h1> and these two elements.. |
| 21:22 | <gsnedders> | tusho: what sort of subtitle is it? I'd probably just use a <p> for it |
| 21:22 | <tusho> | gsnedders: like |
| 21:23 | <tusho> | <h1>Killing Gnomes</h1><h2>A great idea with a terrible implementation.</h2><p>Published 1970-01-01.</p> |
| 21:23 | <gsnedders> | tusho: I'd just use a p emenet |
| 21:23 | <tusho> | which, in the end, I'm trying to display like <p>A great idea with a terrible implementation. Published 1970-01-01.</p> |
| 21:24 | <tusho> | gsnedders: OK then |
| 21:24 | <tusho> | it feels like I should mark a subtitle SOMEHOW, though. |
| 21:25 | <gsnedders> | tusho: Heck, if you want that, just a single p element |
| 21:25 | <gsnedders> | tusho: a p element is markup! :P |
| 21:25 | <tusho> | *g* |
| 21:25 | <jcranmer> | display: run-in ? |
| 21:25 | <gsnedders> | tusho: Also, put the date in a time element |
| 21:25 | <tusho> | yes, I am gsnedders |
| 21:25 | gsnedders | wishes that was better supported |
| 21:25 | <jcranmer> | that seems to be the desired formatting, FWICT |
| 21:26 | <gsnedders> | I use display: run-in on the print styles for my blog, though only Prince XML and Opera actually manage to render the print style correctly |
| 21:26 | <tusho> | hmm |
| 21:26 | <tusho> | I'm all for browser snobbery |
| 21:26 | <tusho> | but I'd like to support ff,opera,safari |
| 21:26 | <gsnedders> | Browsers just tend to have rubbish support of print stuff which doesn't help |
| 21:27 | <gsnedders> | tusho: display: run-in isn't supported in Gecko, so that won't work |
| 21:27 | <tusho> | heh |
| 21:27 | <jcranmer> | gsnedders: patches welcome, I suppose |
| 21:28 | jcranmer | wonders idly if IE 8 will support run-in |
| 21:28 | <gsnedders> | jcranmer: I don't have the interest to write stuff like that for Gecko |
| 21:28 | <gsnedders> | jcranmer: It should, IIRC b1 does |
| 21:28 | <tusho> | jcranmer: that won't be of much interest to me: |
| 21:28 | <tusho> | <meta http-equiv='X-UA-Compatible' content='IE=5'> |
| 21:29 | <tusho> | (i hope it can emulate that far back ;)) |
| 21:29 | <gsnedders> | tusho: that just enters quirks mode |
| 21:29 | <tusho> | aw |
| 21:29 | <tusho> | i want full ie 5 emulation |
| 21:29 | <tusho> | can I have my money back, microsoft? |
| 21:29 | <jcranmer> | content='IE=1'> |
| 21:29 | <tusho> | jcranmer: see, that's not as fun |
| 21:29 | <tusho> | IE 1 would probably just ignore my css |
| 21:29 | <gsnedders> | IE 5.5, IE7 and IE8 are supported |
| 21:29 | <jcranmer> | or IE=40 |
| 21:29 | <tusho> | but ie 5 is just clever enough to really break my page |
| 21:29 | <tusho> | ah, 5.5 is close enough |
| 21:29 | <gsnedders> | (IE5.5 is what quirks mode has been ever since) |
| 21:30 | <tusho> | oh, really? |
| 21:30 | <tusho> | crazy |
| 21:30 | <gsnedders> | But AFAIK it's just supported as switches within the IE7 engine |
| 21:33 | <gsnedders> | jcranmer: they're claiming support for IE8 final of display:run-in |
| 21:34 | <gsnedders> | (but that comes just as part of the full CSS 2.1 support inevitably) |
| 21:34 | <gsnedders> | it seems IE8b1 doesn't, though |
| 21:34 | <jcranmer> | I guess FF needs to get its run-in going, then |
| 21:36 | <gsnedders> | IE8 should be pretty close to getting the print styles of my blog right, too |
| 21:36 | <gsnedders> | (you need some CSS3 support to get it fully right) |
| 21:37 | <tusho> | i think everyone should base their designs on Acid3 |
| 21:37 | <tusho> | ;) |
| 21:41 | <gsnedders> | You need CSS 2.1 + CSS 3's Generated Content for Paged Media and Paged Media modules for the print version of my blog :P |
| 21:54 | <tusho> | Wow. Link? :P |
| 21:55 | tusho | is aiming for trivial markup, trivial design with his blog |
| 21:55 | <tusho> | so it doesn't really need much |
| 21:57 | <gsnedders> | tusho: http://gsnedders.com/ |
| 21:57 | <gsnedders> | tusho: It's not conforming HTML 5, I know |
| 21:57 | <tusho> | <html lang="en-gb-x-sneddy"> |
| 21:57 | <tusho> | someone took a leaf out of hixie's book :-P |
| 21:58 | <gsnedders> | tusho: Oh, never :P |
| 21:58 | gsnedders | is yet to write a spec for it though |
| 21:58 | tusho | considers en-gb-x-sdjfklsfjdoijafoidjfkoiakjfoiajfoiawjoaiewudrcioarucawoxpkaiosafjaidfjdslfk |
| 21:58 | <gsnedders> | tusho: that's non-conforming |
| 21:58 | <gsnedders> | tusho: each subtag has a max. length of eight characters |
| 21:58 | <tusho> | Stop shattering my hopes and dreams. |
| 21:59 | <gsnedders> | (I'd be using en-gb-x-gsnedders if I could) |
| 21:59 | <tusho> | gsnedders: That's what you get for not having an IRC-friendly nickname! Sucks to be you. ;) |
| 22:00 | <gsnedders> | tusho: How's it not IRC-friendly!? |
| 22:00 | <tusho> | IRC nicknames are 7 or less characters. :-P |
| 22:00 | hsivonen | is happy to have a 7-letter surname |
| 22:00 | <tusho> | I have a three letter surname, but I don't like it. :p |
| 22:00 | <sneddy> | tusho: Fine. |
| 22:01 | <tusho> | sneddy: Now you get extra unicorns in your daily servings. |
| 22:02 | <sneddy> | I so have to write a blog post in response to the fact that someone found my blog searching for "claiming a love" |
| 22:02 | <Lachy> | tusho, in which clients are IRC nicknames limited to 7 characters? |
| 22:02 | <tusho> | Lachy: It's just common convention. |
| 22:02 | <tusho> | I mean, not as a hard limit, but what I'd class as an "IRC nickname" |
| 22:04 | tusho | wonders if his blog needs a little more debris than "Here are the latest entries in my blog. If you like, you can see a list of older ones too." and "Licensed under CC by-sa." |
| 22:05 | <sneddy> | (fwiw, I started to write a blog post about last May) |
| 22:06 | <tusho> | sneddy: your url design is mine! |
| 22:06 | <tusho> | (/slug and nothing more) |
| 22:06 | <tusho> | well |
| 22:06 | <tusho> | technically I have some extra stuff |
| 22:06 | <tusho> | though most of that is redirects so I can be pedantic |
| 22:06 | <sneddy> | tusho: Habari has no easy way to change it :P |
| 22:06 | <tusho> | e.g. /slug/comments redirects to /slug#comments |
| 22:07 | <tusho> | so that the comment form can PUT to /slug/comments |
| 22:07 | <tusho> | and /slug/comments/N redirects to /slug#comment-N |
| 22:07 | <tusho> | so that my baleet button can DELETE to /slug/comments/N |
| 22:07 | <tusho> | oh, and I won't be able to use "login" or "post" as a slug |
| 22:07 | <tusho> | that's about it I think |
| 22:08 | <tusho> | oh, and /slug/edit shows my edit form |
| 22:08 | <gsnedders> | /atom, /tag, /user, and /admin are forbidden off the top of my head |
| 22:09 | <tusho> | ah, my atom will be /index.html or /index.atom |
| 22:09 | <tusho> | or, you know, / |
| 22:09 | <tusho> | content negotiation ;) |
| 22:09 | <tusho> | oh, /archives |
| 22:09 | <tusho> | that's it |
| 22:09 | <tusho> | (i don't have tags or anything because they don't really bother me) |
| 22:09 | <tusho> | */archive actually |
| 22:09 | <tusho> | and /archive is just a semi-pretty list of every post ever |
| 22:10 | <tusho> | simplistic :p |
| 22:11 | <tusho> | oh, and on a markupy note, universal edit button is so broken |
| 22:11 | <tusho> | rel="alternate" type="application/x-wiki"? |
| 22:12 | <tusho> | an edit page is not in any way an alternate version of a document and it won't give you back an application/x-wiki page either... |
| 22:12 | <tusho> | they could have, you know, just done rel="edit" |
| 22:14 | <gsnedders> | Anyone want to bully me into going to a specific uni? |
| 22:14 | <tusho> | No. :p |
| 23:02 | <gsnedders> | Philip`, jgraham: You want to meet up next Monday (20080707), or not? |
| 23:07 | <tusho> | Hmm. |
| 23:07 | <tusho> | Why can't I style 'article' in Firefox3, I wonder... |
| 23:07 | <tusho> | It seems to just ignore my styles. |
| 23:07 | <tusho> | Ah. |
| 23:07 | <tusho> | I need to specify display: block |
| 23:16 | <mikedoty> | Hey! Someone here might be able to help me! I'm playing with the Canvas tag. I do a toDataURL and save that to a variable. Then I try to use that variable as the first parameter for drawImage. It somehow errors, and all subsquent calls to toDataURL fail. Is this a known issue by any chance? I'm using Opera 9.27 |
| 23:18 | <othermaciej> | mikedoty: you can't just use a data URL as a parameter to drawImage |
| 23:18 | <othermaciej> | mikedoty: you would have to make it the src of an img element and draw that |
| 23:18 | <mikedoty> | Hm, I do x = new Image(); x.src = toDataURL() then use x |
| 23:18 | <mikedoty> | That's what I meant, sorry |
| 23:19 | <othermaciej> | mikedoty: but it might be a better choice to use getImageData / putImageData |
| 23:19 | <mikedoty> | It is like here http://www.psyguygames.com/paint.html |
| 23:19 | <othermaciej> | if you just want to reuse the image content across the same session |
| 23:19 | <mikedoty> | Ok, I'll reexamine those |
| 23:19 | <othermaciej> | toDataURL is mainly useful for when you want to save it for a later page load |
| 23:20 | <mikedoty> | I think Opera doesn't support get/put though :( |
| 23:20 | <mikedoty> | This would not be a big deal but I had hoped to have it working in the Wii browser |
| 23:26 | <othermaciej> | I'm pretty sure they do |
| 23:26 | <othermaciej> | at least in the newest Opera |
| 23:30 | <mikedoty> | Yeah, I think it says 9.50 does, but the Wii's version is a bit older and I fear it probably doesn't. I could take a look though. |
| 23:31 | <mikedoty> | I think I might be able to skate by with reloading the page when I need to prepare new images (an uncommon occurence). It is hackish but shouldn't be too unsightly |
| 23:31 | <othermaciej> | I think it was added at the same time as toDataURL |
| 23:47 | <Philip`> | mikedoty: That's a security-related bug - if you draw an image from a remote location onto a canvas, then that prevents you later calling toDataURL/getImageData on the canvas (because then your script could access the image's data, which is bad since the image contain be something secret) |
| 23:48 | <Philip`> | mikedoty: but most implementations did/do consider images created from data: URLs to count as 'from a remote location' (or at least not from somewhere they know is safe) |
| 23:49 | <Philip`> | mikedoty: which is a bug since in most cases they should consider data: images to be safe |
| 23:50 | <mikedoty> | Ok, although I'm super disappointed in that, I understand the reasoning and thank you for bringing some light on what was a very dark situation :) |
| 23:51 | <Philip`> | There is a way to work around that security feature in some versions of Opera, but that workaround is a security vulnerability so maybe they've fixed it on the Wii |
| 23:52 | <mikedoty> | Well, what I'm doing is prerendering levels into images, which I just need to do at the start of a level... I think I can get away with reloading the page to start a new level probably |
| 23:52 | <Philip`> | http://www.opera.com/support/search/view/883/ only says it's fixed in 9.5, though |
| 23:55 | <Philip`> | (Draw unsafe stuff on a canvas; call ctx.createPattern(canvas, ...) on that canvas; create a new canvas and set ctx2.fillStyle = pattern and draw stuff, and it won't realise the new canvas is unsafe) |
| 23:57 | <Philip`> | (But if there's a 9.28 then it'll probably be fixed in there) |
| 23:57 | <Philip`> | (where "fixed" means it'll stop working and break any attempts to utilise that hack for noble purposes) |
