| 00:06 | <Hixie> | boogyman: what i'm talking about is not loading a resource that has a media query that doesn't match |
| 00:07 | <boogyman> | and I was stating that I've had decisions on both sides of that fence. |
| 00:08 | <Hixie> | ah |
| 00:09 | <boogyman> | in my opinion, it can/should be a UA's decision on how to load a resource, not something that's defined in a specification (if that is what you were implying). |
| 00:10 | <Hixie> | if the resource is just not going to be used, seems dumb to load it |
| 00:10 | <Hixie> | anyone got an idea for the name of a flag that indicates that a resource should be loaded even if the browser thinks it's already been loaded? |
| 00:10 | <Hixie> | the opposite of "deduping" |
| 00:10 | <Hixie> | "force-unique-instance" or something |
| 00:13 | <boogyman> | Hixie: it may seem illogical to you, but that doesn't mean others cannot make a case for it. I just think that's outside of the bounds of any industry specification. |
| 00:13 | <Hixie> | do you think that if the page's entire markup is: <!DOCTYPE HTML><p>Hello</p> ...the browser should fetch "hello.png" from the server? |
| 00:16 | <boogyman> | If it's not defined in the document, of course not, but if <!DOCTYPE html><head><link type=text/css media="max-width:960px" src="foo.css"><link type=text/css media="min-width:1000x"></head>... is, then let the UA decide when to load those resources |
| 00:17 | <Hixie> | if we offer authors the feature to control whether a resource is fetched or not by including or not including a reference to the resource in the file, it seems reasonable that they would ask (as they have) for the ability to control whether a resource gets downloaded or not even when it _is_ mentioned in the file |
| 00:18 | <Hixie> | since otherwise they could just change their code to do the uglier thing of conditionally not including it at all |
| 00:46 | <Hixie> | oops, my e-mail was so long it got caught by the filter |
| 00:52 | <boogyman> | Hixie: I am saying that Authors should have no say in the matter, but User Agents should be able to decide their strategy, eg desktop agents may choose to load all because of an assumption of higher bandwidth capacity, where palm clients may choose to only load what currently matches... and maybe some UA's decide to provide a way for an end-user to control their preference. Point being, I... |
| 00:52 | <boogyman> | ...think the spec would overstep it's bounds my making that a what/when decision. I do believe however it's within the bounds of the spec to define how something is loaded based upon the what/when. |
| 00:55 | <Hixie> | boogyman: authors have a say in the matter, you just agreed to that (by not including the element) |
| 00:56 | <Hixie> | btw if anyone is interested in the stuff i've been talking about here for the past few weeks, i finally posted my e-mail on the subject |
| 00:59 | <boogyman> | Hixie: I think it goes over the bounds of the UA to load any resource that is not defined in the source, either initially or given some author action (like post-render definition) |
| 01:02 | <JonathanNeal> | erlehmann, annevk, caitp, okay, I’m free to write something if we’re still looking for a polyfill. |
| 01:30 | <caitp> | I don't really get why a <base href="host.com/base">; + <a href="/base/foo"></a> will resolve the anchor's url to "host.com/base/foo"; rather than "host.com/base/base/foo"; |
| 01:30 | <caitp> | who do I blame for that |
| 01:30 | <caitp> | well |
| 01:30 | <caitp> | sorry, add a slash on the end of the base href |
| 01:31 | <caitp> | the slash is implicit <_< |
| 01:32 | <caitp> | just making things behave weirdly and inconsistently for no reason other than the expectation that it's probably what most people want --- except for the people that don't @_@ |
| 03:59 | <JonathanNeal> | What does the “currently-sorting flag” look like? |
| 04:01 | <JonathanNeal> | e.g. http://www.whatwg.org/specs/web-apps/current-work/multipage/tables.html#currently-sorting-flag - what does an element’s flag look like? Is it an attribute or something that can be seen in the dom or a kind of private state? |
| 04:44 | <zcorpan> | JonathanNeal: private state |
| 04:44 | <JonathanNeal> | zcorpan: oh okay, thanks |
| 04:46 | <JonathanNeal> | Have any browsers experimentally added table sortable support? |
| 04:48 | <caitp> | it's not listed on chromestatus, didn't see a bug for gecko yet, but it didn't work in nightly. I'd be pretty shocked if netsurf or servo are there yet |
| 04:50 | <caitp> | (so I'm guessing that's a no unless webkit has started on it) |
| 04:52 | <caitp> | might not be a bad candidate feature for blink-in-js though, at least as a preliminary thing |
| 04:53 | <JonathanNeal> | Okay. |
| 05:01 | <JonathanNeal> | Well, I’ve made some progress, and I’ll make some more on the plane tomorrow. |
| 05:04 | <JonathanNeal> | I don’t get how the UI is supposed to work, but I can experiment with the attributes working stand-alone. |
| 09:16 | <MikeSmith> | Hixie: damn you may have set another record with the length of the omnibus reply |
| 09:17 | MikeSmith | sets aside a week in his schedule to try to read through it |
| 10:13 | <MikeSmith> | Hixie: fwiw, I think in cases like this, you could better take the "here's the tentative proposal" part (which nobody is otherwise going to read until they've read the other hundreds of lines that precede it) and move it to the very beginning of the message -- along with a statement like "For the rationale behind this proposal, read the other 90% [literally] of this message that follows the actual proposal" |
| 10:14 | <MikeSmith> | Hixie: I think otherwise there aren't 7 people in the universe that are going to ever read far enough into that message to realize you've got a concrete proposal in there |
| 10:14 | <annevk> | MikeSmith: presumably those that do will tweet and then it'll sort itself |
| 10:15 | <MikeSmith> | annevk: yeah I suppose so |
| 10:16 | <MikeSmith> | even then I wish there was a way to link into subsections of archived e-mail messages |
| 10:17 | <MikeSmith> | the "Here's how this would handle the use cases listed above" part is great too but it also kind of gets buried at the end |
| 16:11 | <JonathanNeal> | Worked on some of the sortable js on the plane. |
| 17:03 | <smaug____> | annevk: FYI, the patch which makes Notification's click cancelable in Gecko is waiting for review |
| 23:35 | <gsnedders> | jgraham: also ergh merging PRs is so much harder with failing tests :( |