| 15:57 | <Domenic> | JakeA: wanderview_PTO: Bluebird put up their 3.0 cancelation docs. https://petkaantonov.github.io/bluebird/web/docs/api-reference.html#cancellation |
| 15:59 | <JakeA> | Domenic: any info on why they went for "don't care" rather than abort? |
| 16:00 | <Domenic> | JakeA: the usual reason, errors are annoying. You can still see it with .finally though. It's basically exactly what you/we were planning. |
| 16:00 | <JakeA> | Domenic: holy shit, even down to onCancel taking a function, this is unbelievably close |
| 16:01 | <Domenic> | Heh yeah |
| 16:01 | <Domenic> | Still not a big fan of that, but shrug |
| 16:02 | <JakeA> | Your idea of returning a function was fine too, I don't have strong feelings either way |
| 16:02 | <Domenic> | I gotta get back to working on that soon i guess |
| 16:02 | <JakeA> | Curious to find out how they deal with resolved but not settled promises too |
| 16:02 | <Domenic> | You should open an issue asking the exact question |
| 16:03 | <Domenic> | I don't recall enough of the details to relay it |
| 16:03 | <JakeA> | I'll do that now |
| 16:03 | <JakeA> | Domenic: I promised anne something by mid-June, I'm hoping to get untangled with I/O video work by the end of this week, so I can actually do some proper work after that |
| 16:13 | <JakeA> | Domenic: https://groups.google.com/forum/#!topic/bluebird-js/Pc7TCuEo3B4 |
| 20:26 | <JonathanNeal> | Is there a specific document for ADA compliance of web sites that differs from, say, Section 508? |
| 20:27 | <caitp-> | seems like it would be hard to regulate without documentation :x |
| 22:01 | <Domenic> | rniwa: if we said custom elements could define their own constructor, are you willing to run those author-defined constructors during parsing, or...? |
| 22:01 | <Domenic> | (also during cloning) |
| 22:21 | <rniwa> | Domenic: we have to, right? |
| 22:22 | <rniwa> | Domenic: even in Dimitry's design of having a constructor having a callback |
| 22:22 | <rniwa> | Domenic: we'd still have to run it during parsing |
| 22:22 | <rniwa> | Domenic: i don't think there's a way around it |
| 22:22 | <Domenic> | rniwa: no, not in Dmitry's design. In Dmitry's design you just run the HTMLElement constructor like you currently are. |
| 22:22 | <rniwa> | Domenic: yeah but then you'd have to call user defined callback right after that |
| 22:22 | <Domenic> | Right, after parsing, but before returning control to the user |
| 22:23 | <rniwa> | Domenic: I don't think that makes sense. |
| 22:23 | <Domenic> | It's the same as createdCallback is today |
| 22:23 | <rniwa> | Domenic: that exposes "uninitialized" elements |
| 22:23 | <Domenic> | It does indeed expose elements that have not had their createdCallback run |
| 22:23 | <rniwa> | Domenic: because all those custom element callbacks would see other elements that have not been initialized |
| 22:24 | <rniwa> | Domenic: that's precisely why createdCallback is broken. |
| 22:24 | <Domenic> | If we don't allow that then we have to run constructors sync during parsing, before the parser continues creating more elements |
| 22:24 | <Domenic> | It seems fine. Don't mess with your siblings while you're being created. |
| 22:24 | <Domenic> | But I understand if you disagree. |
| 22:24 | <Domenic> | It just seems like the better solution than introducing author code into the parser. |
| 22:24 | <rniwa> | Domenic: I just don't understand what the problem of introducing author code into the parser is |
| 22:25 | <rniwa> | Domenic: we already do this for sync script elements |
| 22:25 | <rniwa> | Domenic: what's so difficult about running script there? |
| 22:25 | <rniwa> | Domenic: I do see that cloning case is hard |
| 22:25 | <Domenic> | Hmm why is cloning different than parsing |
| 22:25 | <rniwa> | but I don't think invoking script during normal parsing is hard at all |
| 22:25 | <rniwa> | Domenic: because there's a bunch of UA internal code that clones nodes internally |
| 22:25 | <rniwa> | Domenic: e.g. editing |
| 22:25 | <rniwa> | Domenic: and a bunch of Range methods |
| 22:25 | <rniwa> | and a few other places |
| 22:26 | <rniwa> | Domenic: Gecko for examples clones the document for priting |
| 22:26 | <Domenic> | sure, but there's a bunch of UA internal code that parses too, right? |
| 22:26 | <Domenic> | hmm |
| 22:26 | <rniwa> | Domenic: I don't think so. |
| 22:26 | <rniwa> | Domenic: if you're talking about innerHTML, sure |
| 22:26 | <rniwa> | Domenic: or copy/paste |
| 22:26 | <rniwa> | Domenic: but those things can already run scripts anyway so there's nothing new |
| 22:27 | <Domenic> | Well if WebKit is OK running author code during parse and cloning then that's good to know. |
| 22:27 | <Domenic> | I got the impression Gecko wasn't, but I'm not sure. |
| 22:29 | <rniwa> | Domenic: they're not okay with running the code during cloning |
| 22:29 | <Domenic> | Welp. |
| 22:30 | <Domenic> | During, or in reaction to? |
| 22:30 | <Domenic> | I guess they must be OK with in reaction to since they propose clonedCallback |
| 22:33 | <rniwa> | Domenic: during cloning |
| 22:33 | <rniwa> | Domenic: I'm not sure if they want to add clonedCallback anymore |
| 22:33 | <rniwa> | Domenic: at least in v1 |
| 22:33 | <Domenic> | Yeah it's listed as v2 |
| 22:33 | <Domenic> | It's pretty obvious though, gotta have it to emulate native elements |
| 22:34 | <smaug____> | Domenic: webkit is ok to run random scripts during cloning? |
| 22:35 | <Domenic> | That was my interpretation of rniwa above, but I might be reading too much into it. |
| 22:35 | <rniwa> | smaug____: I need look into it |
| 22:35 | <rniwa> | smaug____: we might just need to punt cloning in v1 and say it'll be broken once cloned |
| 22:36 | <smaug____> | right |
| 22:36 | <smaug____> | cloning is really hard |
| 22:36 | <Domenic> | i don't think that's feasible |
| 22:36 | <smaug____> | anything can happen when a script runs |
| 22:36 | <rniwa> | Domenic: what is not feasible? |
| 22:36 | <Domenic> | cloning is fundamental to the core issues of custom elements; punting on cloning is like punting on constructors. |
| 22:37 | <Domenic> | (which, y'know, Chrome did, but I'm not happy about.) |
| 22:37 | <smaug____> | node (the original or clone) gets adopted to another document, window is closed, node is moved to another place in the document, etc.. |
| 22:42 | <aklein> | punting cloning would also punt users of <template> |
| 22:42 | <aklein> | I think that'd be undesireable |
| 23:03 | <hober> | if "punting cloning" means "behaves like <canvas> when cloned," i don't think it's all that bad. |
| 23:15 | <aklein> | hober: you'd have to ask rniwa what he meant, but I assumed it meant "don't run the element's constructor when it's cloned", which is not what happens to <canvas> |