| 09:19 | <DennisRasCollege> | I can't seem to find any documentation or article about how to detect image transparency. Any help on where I should read more about that? Basically I want to draw an image and have the user only be able to click the image where it's non-transparent, am I making any sense? |
| 09:19 | <DennisRasCollege> | This is with canvas (I supposed) or SVG if that's my only option |
| 09:23 | <Lachy> | DennisRasCollege, use getImageData(), then use the CanvasPixelArray returned to get the colour of the pixels, with which you can check the value of the alpha channel |
| 09:23 | <Lachy> | that depends on the image having the same origin though, so that getImageData works |
| 09:24 | <DennisRasCollege> | What do mean with having the same origin? |
| 09:25 | <DennisRasCollege> | The position? |
| 09:25 | <Lachy> | I mean, there are security restrictions in place if you've painted an <img> onto a canvas, and that image comes from a different domain than the web page. |
| 09:25 | <DennisRasCollege> | Oh I don't have that problem though, but thanks |
| 09:44 | <DennisRasCollege> | Lachy: wait is getImageData for the canvas or an image object? |
| 09:45 | <Lachy> | it's a method on the canvas context |
| 09:45 | <Lachy> | look up the spec |
| 09:45 | <DennisRasCollege> | But what do I do if I have layered images? |
| 09:45 | <DennisRasCollege> | Yeah I just saw that, which is why I'm asking |
| 09:46 | <Lachy> | You could create a new hidden canvas, paint just that one image you want to check, and then figure out what the co-ordinates would be on the other canvas from that. |
| 09:47 | <DennisRasCollege> | That looks like a lot of processing? |
| 09:48 | <DennisRasCollege> | Say with 100 images |
| 09:50 | <Lachy> | if you want to know where the transparent pixels are for every single image, there's no way to get around the amount of processing you have to do. |
| 09:51 | <Lachy> | it would be less if you could just get the remaining transparent pixels after adding several of the images together. |
| 09:55 | <DennisRasCollege> | yeah but it's not the canvas transparency I was looking for, sorry about that |
| 09:55 | <DennisRasCollege> | anyway my goal is to be able to make a click detection which takes image transparency into consideration |
| 09:56 | <DennisRasCollege> | and i dont know what to search for |
| 09:58 | <Lachy> | if you're donig it on click detection, then you only need to check the specific pixels on a limited set of the images each time, which should reduce the amount of processing you need to do |
| 10:00 | <Lachy> | so when a user clicks co-ordinates x,y, you check which images are covering that location, and then you can check the co-ordinates on those specific images, by painting them to another blank, transparent canvas. |
| 10:01 | <DennisRasCollege> | Yeah that makes sense |
| 10:01 | <DennisRasCollege> | And I guess the background canvas should have images added in the same order as the z-index right? |
| 10:01 | <DennisRasCollege> | highest index first |
| 10:02 | <DennisRasCollege> | This could be fun to do :) |
| 10:05 | <hdhoang> | I don't think it has to: if any image contributes an opaque pixel, the result would not be transparent |
| 10:05 | <DennisRasCollege> | What do you mean? |
| 10:11 | <phrearch> | hello |
| 10:12 | <phrearch> | im trying to find out the default language per user, but the websocket handshake doesnt include accept-language |
| 10:12 | <phrearch> | in the header |
| 10:13 | <phrearch> | is there a way to add this parameter to the header? |
| 10:14 | <jgraham> | phrearch: No |
| 10:14 | <jgraham> | It sounds like a dubious approach to the problem anyway (is accept-language generally right?) |
| 10:15 | <jgraham> | But in general the idea is that you send application-specific data over the socket, not in the handshake |
| 10:16 | <phrearch> | yea indeed. maybe ill have to set preferences then |
| 10:16 | <phrearch> | im using the django framework, but it relies on http headers for proper language detection. cant use that with websocket requests :( |
| 10:17 | <hsivonen> | phrearch: you can write navigator.language into the socket after opening it |
| 10:17 | <Lachy> | DennisRasCollege, he means that once you determine the set of images that need to be checked for transparency, you can paint them all on a new canvas, and then just check the one pixel, rather than having to check each image individually. |
| 10:18 | <phrearch> | hsivonen: aha thanks! |
| 10:18 | <phrearch> | seems like i have to send the language in each websocket request though |
| 10:19 | <hsivonen> | phrearch: is your server stateless between Web Socket packets? |
| 10:19 | <phrearch> | no, it maintains state |
| 10:20 | <phrearch> | you mean if it keeps track of the connections? |
| 10:20 | <hsivonen> | phrearch: I mean: did you mean it would forget the language associated with a connection after establishing the language once? |
| 10:21 | <hsivonen> | if you meant you need te send the language once per Web Socket connection, then, yes |
| 10:21 | <phrearch> | hm no, it could stay in memory, but the whole webframework is built around the notion of a request object available, which is not there in this case(only a reference to the client connection) |
| 10:22 | <phrearch> | hm, think you gave me the answer :) |
| 10:23 | <hsivonen> | I wonder if anyone has done research on how ofter the browser's UI language actually matches the user's content language preference on sites that try to select the content language from the UI language |
| 10:23 | <hsivonen> | *often |
| 10:24 | <phrearch> | most sites use language selection i guess |
| 10:24 | <phrearch> | *manual* |
| 10:24 | <hsivonen> | I use en-US browsers but I want my banking UI to be in Finnish, because banking requires vocabulary that's not only domain-specific but country-specific, too |
| 10:25 | <hsivonen> | phrearch: yeah, manual language selection is likely to actually reflect what the user wants |
| 10:28 | <phrearch> | its actually hard to set chromium in something else than en-us hehe |
| 10:30 | <phrearch> | ill try to detect the browser language first, and add an option for the user to override the setting |
| 10:31 | <phrearch> | maybe the initial bootstrapping http request is suitable for that |
| 10:53 | <DennisRasCollege> | Lachy_: that's not the result im looking for though, gonna get back to you in a bit after dinner |
| 10:54 | <Lachy> | DennisRasCollege, ok. You haven't been particularly clear about what exactly you want to achieve |
| 12:10 | <MikeSmith> | hsivonen: plh points out to me that the message "The document validates according to the specified schema(s)." may lead people to assume that validator.nu is only doing schema validation |
| 12:14 | <MikeSmith> | not sure what better wording could be put in there |
| 12:15 | <MikeSmith> | maybe make it a hyperlink to the description on about.validator.nu of the types of checks it is doing? |
| 12:16 | <MikeSmith> | "The document validates according to the specified schema(s) and to additional constraints checked by the validator." |
| 12:16 | <MikeSmith> | maybe |
| 12:22 | <hsivonen> | MikeSmith: good point |
| 12:23 | <MikeSmith> | hsivonen: any thoughts about better wording? |
| 12:23 | <hsivonen> | MikeSmith: your wording seems OK |
| 12:24 | <MikeSmith> | ok |
| 12:24 | <hsivonen> | jgraham: it seems that the WebM download prompt is what confuses @font-face in Opera Mobile 11 on Symbian |
| 12:24 | <hsivonen> | jgraham: If I don't call .play(), the fonts apply |
| 12:26 | <hsivonen> | jgraham: I updated webm.html5.org with a special case for Opera Mobile 11.x for Symbian, but now the page never finishes loading and I don't know why |
| 12:27 | <hsivonen> | anyone: Please let me know if webm.html5.org also needs special casing for Opera Mobile 11 for Android |
| 12:28 | <hsivonen> | Opera Mobile 11 on Windows 7 seems OK |
| 12:28 | <jgraham> | hsivonen: Right. I think the bug is some bad interaction between the video download stuff and resource loading |
| 12:29 | <hsivonen> | jgraham: I suspect having ever had a <video> element in the DOM confuses the browser to think the page hasn't finished loading even if I remove the <video> from the DOM |
| 12:30 | <hsivonen> | curiously, the browser say "probably" to the canPlayType query |
| 12:30 | <hsivonen> | which seems wrong, since it, in fact, cannot play the type |
| 12:31 | <jgraham> | Interesting |
| 12:31 | gsnedders | blames foolip |
| 12:31 | <jgraham> | gsnedders: I think this is platform specific |
| 12:32 | <gsnedders> | jgraham: If the canPlayType thing is what I think — I doubt it is. |
| 12:32 | <hsivonen> | It's sad that canPlayType sucks the most in various mobile browsers where it is needed the most, since site authors can't test in every flavor of mobile browser |
| 12:32 | <jgraham> | My understanding (without trying it) is that it uses the platform media player to do the actual playing |
| 12:32 | <gsnedders> | That is correct. |
| 12:32 | <foolip> | I didn't know we had any support for <video> on Symbian at all |
| 12:32 | <foolip> | in fact I'm pretty sure we don't |
| 12:33 | <jgraham> | foolip: We invoke the platform media player on seeing a <video> element |
| 12:33 | <jgraham> | Or something |
| 12:33 | <hsivonen> | foolip: you most certainly show a black rectangle with a play button and prompt to download the file when calling .play() or pressing the button |
| 12:33 | <jgraham> | I'm not quite sure how it works |
| 12:33 | <foolip> | jgraham, I've never heard from a platform team implementing <video> on Symbian |
| 12:33 | <hsivonen> | foolip: and it never seems to fire any of the events one might expect after calling .play() |
| 12:33 | <hsivonen> | like Konqueror... |
| 12:34 | <DennisRasCollege> | Lachy: No and that's my fault, I'm sorry about that. It's just that I don't want to seem like someone who hasn't done research (when answers might be SO simple even though they aren't for me) since I have done a lot of research |
| 12:35 | <hsivonen> | anyway, Opera Mobile 11 now has the questionable honor shared with Konqueror to be a browser that can't be subjected to a WebM .play() attempt |
| 12:35 | <foolip> | hsivonen, that all sounds very strange, probably the feature has been enabled without actually implementing the platform bits |
| 12:36 | <DennisRasCollege> | Lachy: I'm creating a 'game' sort of where a lot of sprites overlap each other and I need a click detection which detects which sprite is clicked on even though they overlap eachother and got transparent pixels within the sprite image |
| 12:36 | <hsivonen> | foolip: I think I even saw a marketing video that showed the download prompt as a feature |
| 12:36 | <hsivonen> | foolip: but the video may have demoed Android |
| 12:37 | <hsivonen> | so I'd like to know if I should blacklist Opera Mobile 11 on Android, too |
| 12:38 | <jgraham> | hsivonen: Maybe you could try the android emulator |
| 12:38 | <foolip> | hsivonen, it looks like we have implemented some support for it |
| 12:38 | <jgraham> | Maybe I should get a phone that actually lets me use a web browser :) |
| 12:38 | <foolip> | hsivonen, not very well though, this is the first time I've heard about it |
| 12:39 | <hsivonen> | jgraham: I didn't know there was any useful kind of emulator |
| 12:39 | <jgraham> | I have no idea if it is useful |
| 12:39 | <jgraham> | http://developer.android.com/guide/developing/devices/emulator.html |
| 12:39 | <Lachy> | DennisRasCollege, ok, then in that case, you will need to check each image sequentially, acording to their z-index from front to back, until you find a non-transparent pixel at the clicked coordinates |
| 12:39 | <hsivonen> | jgraham: do you mean what's in the SDK? I have only read bad things about it. |
| 12:39 | <hsivonen> | jgraham: yeah, that one |
| 12:39 | <foolip> | hsivonen, it looks like it will reply "probably" to absolutely anything and try to download it |
| 12:39 | <MikeSmith> | foolip: hey, I got your reply. I had been pinging you at the time because I was talking with Silvia and Eric Carlson and Frank Olivier about media stuff and wanted to ask you about… something |
| 12:39 | <jgraham> | hsivonen: Oh, OK |
| 12:39 | <hsivonen> | jgraham: I didn't get it to boot when I last tried |
| 12:40 | <MikeSmith> | foolip: can't remember now what it was |
| 12:40 | <foolip> | MikeSmith, ok |
| 12:40 | <MikeSmith> | anyway, thanks for having followed up about it |
| 12:40 | <jgraham> | hsivonen: I can probably find someone with an android phone and ask them to try |
| 12:40 | <hsivonen> | jgraham: yeah |
| 12:40 | <hsivonen> | anyone on this channel with an Android phone and Opera Mobile 11 on it? |
| 12:41 | <hsivonen> | I guess I should get an Android phone at some point |
| 12:42 | <hsivonen> | but I have been dragging my feet waiting for the dust to settle around Firefox for Android requirements and for a WebM hardware decoder to appear in Android devices |
| 12:42 | <foolip> | hsivonen, I borrowed one from a colleague, what do you want me to test? |
| 12:42 | <hsivonen> | foolip: if loading http://webm.html5.org/ prompts to download or completes with useful information or something else |
| 12:43 | <jgraham> | hsivonen: Oh, WebM for android. I think you just put me off buying an android device for another round :) |
| 12:44 | <foolip> | hsivonen, it plays a 1 second video in fullscreen of "WebM" with a green tick next to it, then I see "Congratulations!" |
| 12:44 | <hsivonen> | jgraham: I don't actually know anything, but since the hardware design was announced on the WebM blog, I have been assuming that it's expected to find its way to Google's OEMs some day |
| 12:44 | <hsivonen> | foolip: great |
| 12:45 | <hsivonen> | foolip: and congrats for shipping WebM support for Android! |
| 12:45 | <jgraham> | hsivonen: Yes, I know about the same as you, I just hadn't considered it :) |
| 12:45 | <foolip> | hsivonen, I actually had no idea that WebM worked yet :) |
| 12:45 | <foolip> | we're just using the platform player on android I think |
| 12:46 | <hsivonen> | I guess now I have to edit the android advice to say that Opera Mobile 11 works in addition to Firefox |
| 12:46 | <hsivonen> | foolip: what version of Android did you test on? |
| 12:47 | <foolip> | hsivonen, I don't know, how do I check? |
| 12:47 | <foolip> | I can ask |
| 12:47 | <hsivonen> | foolip: I don't know |
| 12:47 | <foolip> | hsivonen, 2.3.3 says the owner |
| 12:47 | <foolip> | let's assume that's true |
| 12:48 | <hsivonen> | foolip: ok. that one is supposed to ship with a WebM decoder in the system |
| 12:48 | <hsivonen> | so now we don't know what happens in Opera on downlevel Android |
| 12:48 | <hsivonen> | assuming that you indeed are using the system bits |
| 12:52 | <hsivonen> | I wonder if there's a market: URL for Opera |
| 12:53 | hsivonen | guesses market://details?id=com.opera.browser |
| 12:58 | <hsivonen> | OK. I think I'm now done with webm.html5.org updates for today. |
| 12:58 | <hsivonen> | foolip: thanks |
| 12:58 | <hsivonen> | anyone: If you find that Opera Mobile 11 does not play WebM on downlevel Android, please let me know |
| 13:11 | <miketaylr> | hsivonen: "For Android versions earlier than 2.3 only H.264 is supported, while from version 2.3 onward WebM is also supported." |
| 13:11 | <miketaylr> | from http://www.opera.com/docs/specs/productspecs/ |
| 13:17 | <hsivonen> | miketaylr: is that *really* 2.3 or 2.3.3? |
| 13:17 | <miketaylr> | i'm not certain... i'm stuck on 2.2 :/ |
| 13:17 | <hsivonen> | miketaylr: IIRC, 2.3.0 didn't have WebM even though it was originally announced |
| 13:18 | hsivonen | goes edit webm.html5.org again |
| 13:18 | <miketaylr> | i'll ask around and have that updated if it's incorrect |
| 13:18 | <hsivonen> | miketaylr: thanks |
| 13:19 | <hsivonen> | Hmm. Opera Mobile itself doesn't report the Android version, so it's not possible to give targeted upgrade advice |
| 13:28 | <hsivonen> | In general, giving browser advice to Android users sucks. Firefox wants an ARM v7 CPU, and you don't know if the visitor has one. |
| 13:29 | <hsivonen> | Opera Mobile 11 needs Android 2.3, but you don't know what version of Android the user has if the user is already running Opera Mobile or Opera Mini |
| 13:29 | <gsnedders> | hsivonen: You have no way on ARM to tell what sort of CPU you have apart from blindingly trusting the OS |
| 13:29 | <gsnedders> | (or trying and seeing if you crash) |
| 13:30 | <hsivonen> | gsnedders: as I understand it, Firefox reads stuff under /proc to figure out if the CPU supports neon |
| 13:30 | <hsivonen> | also to blacklist the JIT on certain Samsung devices |
| 13:34 | <gsnedders> | hsivonen: AFAIK short of reading /proc there's nothing you can do, as I understand it |
| 13:35 | <gsnedders> | Like, there's nothing analogous to CPUID. |
| 13:35 | <hsivonen> | gsnedders: how does Linux itself find out what sort of ARM CPU it has? |
| 13:39 | <gsnedders> | hsivonen: This I don't entirely understand, and looking at this again I'm confused… I *believe* it is set at compile-time. |
| 13:39 | hsivonen | wonders how Verizon Opera Mini differs from Opera Mini except by being a version behind and having an ugly icon |
| 13:39 | <gsnedders> | hsivonen: looking at it on a BeagleBoard, "Features" is incomplete, for example. |
| 13:40 | <hsivonen> | gsnedders: so if the kernel isn't device-specific (e.g. Debian) one can't tell? |
| 13:41 | <gsnedders> | hsivonen: I believe this to be the case. I'm just looking on Ubuntu on ARM, FWIW. |
| 13:43 | hsivonen | didn't know Ubuntu shipped on ARM |
| 13:45 | <gsnedders> | hsivonen: Only an unofficial port |
| 13:46 | <hsivonen> | http://en.wikipedia.org/wiki/Linux_distribution_comparison#Architecture_support seems to list a distro as "Yes" if it has ever supported a given architecture |
| 13:46 | <hsivonen> | which misrepresents e.g. PPC support today |
| 14:15 | <Lachy> | Opera Mobile for Android build previewing getUserMedia() support has released http://my.opera.com/core/blog/2011/03/23/webcam-orientation-preview |
| 14:16 | richt | it remains to be seen if we're going to bleed out on this cutting edge with getUserMedia() |
| 14:21 | <hsivonen> | so where does HTML5 get the idea that cloning a script clones the "already started" flag? |
| 14:23 | <hsivonen> | whoa! Hixie is not on IRC. |
| 14:23 | <hsivonen> | anyway, the context is https://bugzilla.mozilla.org/show_bug.cgi?id=644034 |
| 14:24 | <Lachy> | Hixie hasn't been on IRC since at least yesterday. I don't know why. |
| 14:24 | <jgraham> | I think someone said his server is being reimaged |
| 14:24 | <jgraham> | So I blame wilhelm |
| 14:26 | <jgraham> | hsivonen: Your TCs need better pass conditions |
| 14:26 | <jgraham> | Or need better descriptions of what's going on |
| 14:27 | <hsivonen> | jgraham: view source :-) |
| 14:27 | <jgraham> | hsivonen: I did :) |
| 14:27 | <hsivonen> | jgraham: also, it's not clear if I should call the specced behavior "PASS"! |
| 14:28 | <jgraham> | hsivonen: Making the descriptions "cloned script ran" and "cloned script did not run" would be clear |
| 14:29 | <hsivonen> | jgraham: changed |
| 14:29 | <karlcow> | Hixie decided to give up finally, and grow veggies in Southern America in a remote farmland. |
| 14:29 | <jgraham> | karlcow: You are confusing Hixie and annevk :p |
| 14:29 | <karlcow> | ahaha |
| 14:30 | <hsivonen> | Opera passes, FWIW |
| 14:32 | <wilhelm> | I am indeed to blame! But I like to think of it as an accidental visit to rehab. (c: |
| 16:44 | <MikeSmith> | is anybody here using django-cms? |
| 16:44 | <MikeSmith> | or tried it? |
| 16:58 | <jgraham> | MikeSmith: Is django-cms something different to django? |
| 16:58 | <MikeSmith> | yeah |
| 16:58 | <MikeSmith> | it's a specific application of django |
| 16:59 | <MikeSmith> | as far as I understand it |
| 16:59 | <MikeSmith> | which isn't very far |
| 16:59 | jgraham | finds the website |
| 17:00 | <smaug____> | Google did it again. http://chrome.blogspot.com/2011/03/talking-to-your-computer-with-html5.html Calling their own API as part of HTML5 "HTML5 speech input API" |
| 17:05 | <MikeSmith> | smaug____: I think they have written a draft for that… maybe "HTML5" now means "existence of a draft for an open standard" |
| 17:06 | <MikeSmith> | anyway, I'm happy until PHP starts getting called HTML5 |
| 17:06 | <smaug____> | MikeSmith: they have written a (not very good, IMO) proposal for HTML Speech XG |
| 17:06 | <smaug____> | there are also other proposals |
| 17:06 | <MikeSmith> | yeah |
| 17:08 | <MikeSmith> | we need a isPointInPath(x, y) to determine if something really is HTML5 |
| 17:08 | <MikeSmith> | with a non-zero winding rule number |
| 17:08 | <smaug____> | Microsoft has a lot more sane proposal for HTML+Speech |
| 17:08 | <MikeSmith> | I haven't actually seen that one, I guess |
| 17:09 | <smaug____> | well, MS has 3 proposals, but 2 of those are nice ones |
| 17:09 | <smaug____> | :) |
| 17:10 | <MikeSmith> | that's a good way to do it |
| 17:47 | <MikeSmith> | jgraham: sometime soon, I would like to get your help with figuring out where we need to be going with trying to get a common test runner set up |
| 17:47 | <MikeSmith> | I can go to where you are if needed |
| 17:48 | <MikeSmith> | you still in Linköping, right? |
| 18:15 | AryehGregor | realizes that foo.insertBefore(bar, foo.firstChild) works to insert bar as the first child of foo, by magic |
| 18:15 | <AryehGregor> | Would prependChild() have killed anyone, honestly? |
| 18:17 | <Lachy> | AryehGregor, does that work even if foo has no children, and foo.firstChild returns null? |
| 18:18 | <AryehGregor> | Lachy, yes, because foo.insertBefore(bar, null) is the same as foo.appendChild(bar). |
| 18:18 | <AryehGregor> | As I said, magic. |
| 18:18 | <Lachy> | ok |
| 18:27 | <zewt> | that's not magic, that's avoiding redundant entry points |
| 18:28 | <zewt> | when you have several exactly equivalent ways of doing the same thing, all you do is fragment code, with everyone choosing one at random |
| 18:29 | <AryehGregor> | You make the code shorter, clearer, and more readable. |
| 18:29 | <AryehGregor> | Python's startsWith() and endsWith() methods for strings are a great example. |
| 18:30 | <AryehGregor> | When they're useful, it's clear that you do in fact want to use them rather than some other methods, and they read much better than non-special-purpose alternatives. |
| 18:30 | <zewt> | "insert before the first child" is perfectly readable to me |
| 18:30 | MikeSmith | finds himself strongly agreeing with AryehGregor |
| 18:31 | <AryehGregor> | E.g., if url.startsWith("http://"): vs. if url.substr(0, 7) == "http://": or if re.match("^http://", url): (I don't even remember if the latter two are correct, but you get the point) |
| 18:31 | <AryehGregor> | zewt, no, because it's not obvious what happens if there's no first child. |
| 18:31 | <zewt> | it's not obvious if you don't know the API |
| 18:32 | <AryehGregor> | Actually, Python is a great example of usually providing only one good way to do things, while still being very easy to learn and read. |
| 18:32 | <zewt> | (and even then it seems pretty straightforward) |
| 18:32 | <AryehGregor> | zewt, you can't expect the typical developer to know API details. Even if you know the API, it's not so logical that you wouldn't have to think for a moment, unless maybe you're *really* familiar with it. |
| 18:33 | <AryehGregor> | prependChild() would be shorter and clearer. |
| 18:33 | <zewt> | that's just it--a Python API wouldn't have both insertBefore(obj, null) and insertAtBeginning(obj), it'd only have the former |
| 18:33 | <zewt> | (there's no "native" tree or linked list type in Python, though) |
| 18:33 | <zewt> | (oddly) |
| 18:34 | <AryehGregor> | Well, why does the API have appendChild(), then? I guess according to you that shouldn't exist either? |
| 18:35 | <MikeSmith> | http://my.opera.com/operamobile/blog/the-components-of-opera-mobile-11-on-android is a nice article |
| 18:35 | <zewt> | because you can't implement appendChild with insertBefore |
| 18:35 | MikeSmith | wonders who Dag Olav Norem is |
| 18:35 | <AryehGregor> | Sure you can. a.appendChild(b) == a.insertBefore(b, null). |
| 18:36 | <zewt> | er, yeah (mixing up in my head due to multitasking)--and that's exactly why there shouldn't be multiple ways to do that |
| 18:37 | <zewt> | instead we have multiple methods to remember |
| 18:38 | <AryehGregor> | http://docs.python.org/py3k/library/stdtypes.html#mutable-sequence-types |
| 18:38 | <AryehGregor> | That has append(), extend(), and insert() even though you can get the same effect by array slicing. |
| 18:38 | <AryehGregor> | Also remove(). |
| 18:45 | <zewt> | it boils down to what you consider an "obvious way to do something"; to me, with a node-based data structure like a tree, it makes sense to have a single "insert child" method |
| 18:45 | <zewt> | which I think is not, at least, comparable to s[len(s):len(s)] = [x] :) |
| 18:50 | <zewt> | jeeezus |
| 18:50 | <zewt> | mail from ferraiolo on webapps is a poster child for why html mail is a bad idea. heh |
| 18:50 | <AryehGregor> | Man, text-decoration is a horrible property in every possible respect. |
| 18:50 | <zewt> | i can't even tell what his mailer is trying to do with that gigantic quote header |
| 18:50 | <AryehGregor> | It sets several orthogonal things with no way to set them separately and no way to override it in descendants. |
| 18:55 | <zewt> | one can also only wonder why "blink" is in there at all |
| 18:55 | <zewt> | "let's resurrect the single most hated thing webpages have ever done!" |
| 18:56 | <zewt> | (okay, pop-unders are probably higher on that list) |
| 18:57 | <kennyluck> | I wonder why no browser implementor is interested in the "text-decoration-skip" property. At least it's a canceler. |
| 18:57 | <zewt> | what the heck does this mean? "SVG1, CSS1-only, and CSS2-only user agents may implement the older model and still claim conformance to this part of CSS 2.1. (This does not apply to UAs developed after this specification was released.)" |
| 18:58 | <zewt> | "you can claim conformance to this if you implemented it before the spec was released, but not after"? spec grandfathering? that seems ... sort of absurd, heh |
| 18:58 | <zewt> | am I misreading that? |
| 19:00 | <kennyluck> | Backward compatible on UA that claims to conform to http://www.w3.org/TR/CSS2/ I guess. They should have linked to the date-specific version. |
| 19:57 | <Hixie> | does opera have a pushstate implementation? |
| 20:04 | <karlcow> | Hixie: for history stuff? |
| 20:04 | <karlcow> | not in a public build yet |
| 20:05 | <aho> | 11.01 and 11.50 (webgl build) don't support it yet |
| 20:06 | <aho> | if anyone wants to try it... here is a nice demo page: http://html5demos.com/history |
| 20:25 | <Hixie> | karlcow: k |
| 21:05 | <hober> | Hixie: nice to finally meet you, however briefly |
| 21:18 | <Hixie> | hober: indeed! |
| 21:20 | <Lachy> | Hixie, just wondering when you next plan to work on the device streaming api? |
| 21:22 | <Hixie> | soonish. the w3c html wg deadlines keep getting in my way. |
| 21:23 | <Hixie> | history api is my next job, since that's urgent (shipping already, fixes must bedone asap) |
| 21:23 | <Hixie> | then device, then updating websockets, then 147 and 144 |
| 21:33 | <Hixie> | tomorrow's the deadline on http://www.w3.org/2002/09/wbs/40318/issue-122-objection-poll/results if anyone cares |
| 21:33 | <Hixie> | not really sure what to say on that one, both options seem kinda poor and neither seem related to the issue i originally raised |
| 21:34 | <Hixie> | othermaciej: http://www.w3.org/2002/09/wbs/40318/issue-31-80-objection-poll/ seems corrupted |
| 22:04 | <AryehGregor> | othermaciej, here's a page that seeks to convert Subversion users to the DVCS way of thinking, if you're interested: http://hginit.com/00.html |
| 22:29 | <Lachy> | didn't anyone submit a no-change proposal for issue-122? |
| 22:30 | <Hixie> | apparently not |
| 22:30 | <Lachy> | that sucks. why don't the survey's include a no change option by default anyway? |
| 22:30 | <Hixie> | i didn't expect it to have any effect on the spec since it was an issue on another document |
| 22:34 | <Lachy> | what do you mean? It's about the lady of shallott example in the spec. How is that about another document? |
| 22:35 | <Hixie> | it's about the lady of shallott example in faulkner's document, not the spec |
| 22:35 | <Hixie> | i raised the issue. then retracted it, but the chairs let it live anyway. |
| 22:35 | <Hixie> | then it somehow mutated into changes to the spec |
| 22:41 | <Lachy> | oh, well if it's about the alt-techniques document, then I can ignore it as an irrelevant spec |
| 22:41 | <Hixie> | it was about alt-techniques |
| 22:41 | <Hixie> | but the changes proposed are on the html spec |
| 22:42 | <Lachy> | hmm, it seems the second change proposal relates to the spec. The first proposal only refers to the alt-techniques document |
| 22:43 | <Hixie> | really? |
| 22:43 | <Hixie> | i thought both applied to both |
| 22:44 | <Hixie> | step 2 of the details of CP 1 affects HTML5 |
| 22:44 | <Hixie> | and CP2 only affects html5 |
| 22:44 | <Lachy> | oh, wtf. I'm confused now |
| 22:44 | <Lachy> | the description of the issue at the beginning of the questionnaire only refers to there being a problem in alt-techniques. |
| 22:45 | <Hixie> | yeah this does seem highly misleading |
| 22:45 | <Hixie> | and confusing |
| 22:45 | <Hixie> | and both proposals seem net negative to me |
| 22:45 | <Hixie> | hence my not being sure what to do |
| 22:53 | <jgraham> | Hixie: Point this out and assume that reasonable action will be taken to remedy the evident problems? |