| 05:43 | <zewt> | Hixie: i wonder if this amounts to a valid WebVTT use case: http://pastebin.com/pCDhnTWk |
| 06:19 | <nessy> | zewt: you'll need to convert to valid tags and timing ;-) |
| 07:59 | <Ms2ger> | Why is the view argument to TextDecoder.decode optional? |
| 10:03 | <Stevef> | For those zero or more people that frequent this channel intersted in the fate of hgroup in HTML5 I have updated one of my proposals http://www.w3.org/html/wg/wiki/ChangeProposals/hgroup replacement element defined here http://www.html5accessibility.com/tests/subline.html |
| 14:09 | <zewt> | Ms2ger: to allow flushing the decoder for stateful encodings, i think |
| 14:09 | <zewt> | (without having to construct a dummy ArrayBuffer, which would just be busywork) |
| 14:15 | <zewt> | Hixie: btw got another of those mails, same formatting different address, so whatever it is is spoofing senders |
| 14:18 | <zewt> | Hixie: haven't squinted at it to see if there's anything useful (headed to work now), but here are both of them, fwiw: http://pastebin.com/CWnKASuT http://pastebin.com/EVDNBS9x |
| 14:41 | <smaug____> | hmm, could all the elements which load something have attribute for after-load-event loading |
| 14:42 | <smaug____> | could be nice for iframes for example |
| 14:42 | <smaug____> | in case of less essential content, like various "like" buttons |
| 15:03 | <Velmont> | smaug____: Hehe, like @defer? |
| 15:04 | <smaug____> | close to that |
| 15:05 | smaug____ | can't remember when 'defer' scripts start loading |
| 15:05 | <smaug____> | is it right after DOMContentLoaded |
| 15:05 | <gsnedders> | I think so. |
| 15:05 | <smaug____> | "the defer attribute is present, then the script is executed when the page has finished parsing" |
| 15:05 | <smaug____> | so, I don't want that |
| 15:06 | <smaug____> | I want something to happen even later |
| 15:07 | <odinho> | Yea, for |
| 15:07 | <jgraham> | Well there isn't really anything after "load" |
| 15:07 | <jgraham> | Even async scripts are run before that |
| 15:08 | <jgraham> | YOu can ofc just run stuff from the load event |
| 15:08 | <odinho> | for <img> I'd wanted it to be "whenever, after page has finished loading". etc. So that javascript can choose what to load etc. |
| 15:09 | <smaug____> | odinho: well, then you can just have <img> |
| 15:10 | <odinho> | smaug____: No, images are prefetched even before script is run. |
| 15:11 | <smaug____> | but if you have <img> |
| 15:11 | <smaug____> | without src |
| 15:11 | <smaug____> | when just set .src using js |
| 15:11 | <smaug____> | s/when/then/ |
| 15:11 | <odinho> | Ah, that won't work in browsers without js/without it turned on. |
| 15:11 | <smaug____> | well, you said "So that javascript can choose what to load etc." :) |
| 15:12 | <odinho> | smaug____: Well, compat was a prerequisite :P |
| 15:13 | <Wilto> | Hixie: Since I know you’re leafing through the respimg threads over the next few days, http://www.w3.org/community/respimg/2012/08/04/picture-in-drupal-8/ and http://www.w3.org/community/respimg/2012/08/04/picture-in-the-html5-spec/ might be worth a read for the current state of Florian Rivoal’s `picture` proposal. |
| 15:14 | <smaug____> | hmm, if there was load-after-pageload flag, I wouldn't be surprised if certain kinds of pages would misuse it, and actually load ads etc first, and essential content after pageload |
| 15:14 | <smaug____> | that seems to happen with some Finnish news sites |
| 15:17 | <Ms2ger> | "A responsive, proportional, nestable, and mobile-first 6-column grid framework." |
| 15:17 | <Ms2ger> | Nice buzzword density |
| 15:18 | <smaug____> | well, 'A' and 'and' aren't buzzwords :) |
| 15:22 | <odinho> | Wilto: You should probably say why you think srcset isn't good enough as specced now. As that was what Hixie wondered last time. |
| 15:23 | <tobie> | Hixie: were use cases and requirements for App Cache ever documented somewhere? I've found bits and pieces here and there but nothing concrete. |
| 15:26 | <jgraham> | tobie: Generally the "documentation" is mail to the whatwg list |
| 15:26 | <tobie> | jgraham: yeah - what I meant by bits and pieces here and there. |
| 15:29 | <jgraham> | tobie: I doubt that anyone went through that input and wrote up a summary. Hixie typically doesn't do that in an explicit fashion. |
| 15:35 | <tobie> | jgraham: thanks. |
| 15:45 | <Hixie> | tobie: the use case was basically "write an app that works whether the user is on a reliable network or not" |
| 15:46 | <Hixie> | Wilto: thanks |
| 15:47 | <Wilto> | odinho: Yeah, that’s more what http://www.w3.org/community/respimg/2012/06/18/florians-compromise/ is about — I figured everyone was sick of me throwing that link around, by this point. |
| 15:47 | <Wilto> | That one was posted to the list not long ago, as well. |
| 15:50 | <odinho> | Wilto: But that post does not answer it. :-) |
| 15:51 | tantek | really dislikes framing a solution as a compromise as that implies there is something technical/design about it that has been compromised (deliberately suboptimal). |
| 15:53 | <odinho> | Wilto: I'm designing a web site now (magazine), and there is nothing missing in the srcset, as it is currently specced for what I want to do. |
| 15:53 | <tobie> | Hixie: that's… broad! |
| 15:54 | <Hixie> | tobie: use cases ususally are |
| 15:54 | <Hixie> | tobie: at least, use cases for entire sections are :-) |
| 15:55 | <Hixie> | if you have a more narrow feature for which you want use cases, let me know :-) |
| 16:42 | <Ianrr> | Im trying to use html5lib to parse HTML (python). I cant find any good examples of how to pull specific tags etc. Can any point me in the right direction |
| 16:43 | <gsnedders> | Ianrr: Depends on what tree-builder you're using. |
| 16:44 | <gsnedders> | Ianrr: Basically, it just uses some other library to represent the tree, so you'd have to look at the documentation for that. |
| 16:44 | <Ianrr> | etree from lxml |
| 16:46 | <gsnedders> | I'd start off looking at the ElementTree documentation, then the lxml documentation. |
| 16:46 | <Ianrr> | ahh ok, so how does html5lib fit into this? |
| 16:46 | <gsnedders> | In short, you probably want something like doc.find("//{http://www.w3.org/1999/xhtml}p";) |
| 16:46 | <Ianrr> | it builds the tree? |
| 16:46 | <gsnedders> | Ianrr: html5lib just parses the input data into a tree. |
| 16:47 | <Ianrr> | so what is the benefit of using say html5lib or just strait up lxml? html5lib is a more robust parser? |
| 16:48 | <gsnedders> | Yeah. libxml's HTML parser won't cope very well with random content on the web, and has a history of infinite loops and other such things. |
| 16:48 | <Ianrr> | Ok great. Thats what I found, it couldn't handle a lot of HTML i threw at it |
| 17:00 | <Ianrr> | After I parse out something and then do something like lxml.etree.tostring(some_tag), it prints all the html tags like "</html:ul>" is there a way to get things to print in normal HTML |
| 17:00 | <Ianrr> | Normal HTML meaning no scape before it |
| 17:00 | <Ianrr> | scope |
| 17:00 | <gsnedders> | Ianrr: lxml.etree.tostring returns XML. |
| 17:01 | <Ianrr> | ahh |
| 17:01 | <gsnedders> | Ianrr: See html5lib.HTMLSerializer (I think the class is called) |
| 17:01 | <Ianrr> | Thank you |
| 17:01 | <gsnedders> | Some name along those lines. :P |
| 17:18 | <Ianrr> | It seems like etree.tostring() has a method where I can specify xml or html but it still seems to print <html:TAG> for everything |
| 17:25 | <Ianrr> | Here is a short sample of what i'm doing and the output http://pastebin.com/RUSQ1FRb |
| 17:33 | <Ianrr> | Figured it out, I had to add namespaceHTMLElements=False |
| 17:33 | <Ianrr> | to html5lib.HTMLParser |
| 18:26 | <Wilto> | odinho: I mean, it’s a functional markup pattern, for certain. We just have to recognize that unless it continues to be developed and adapted in parallel with media queries, we’d forever be limited to resolution, width, and height as detemining factors, and pixel-based values. |
| 18:28 | <Wilto> | Otherwise we have a very specific set of markup that solves a very specific set of use cases. For my part, I use ems in all my media queries; it’s a growing practice amongst developers. In that way, `img srcset` has already already fallen out-of-touch with current practices. |
| 18:38 | <tantek> | I recall someone here saying they'd either never heard of people having problems with cloud services (specifically Google's) or themselves had not had any. This is a good article documenting some specific problems, and why "just use the cloud" is an insufficient answer (to anything) http://gizmodo.com/5932161/why-the-cloud-sucks |
| 18:38 | <tantek> | (authored by none other than Apple co-founder Steve Wozniak) |
| 18:39 | <tantek> | also referenced, Google Calendar, iCloud etc. all the fanboy services. |
| 19:03 | <Ianrr> | I want to replace the contents of an HTML <body></body> with a string. I'm using lxml and etree, any ideas where I would start? |
| 20:17 | <jgraham> | Ianrr: body = root[1]; body.clear(); body.text = "your string"; |
| 20:17 | <jgraham> | Or something like that |
| 20:24 | <Ianrr> | thanks i'll give it a try |
| 20:30 | <Ianrr> | jgraham: If I wanted to insert HTML, would I do something like i = etree.tostring(THE_HTML) then body.insert(i) |
| 23:08 | <zewt> | (heh, I've been bit by the confusing namespaceHTMLElements thing myself) |