| 05:20 | <royiv> | Can someone in WHATWG, W3, the IETF define the *semantics*, and not just the grammar, of application/x-www-form-urlencoded in a real standard like an RFC or something? Like, are "k" and "k=" semantically equivalent? Can we all agree that either dupes are silly, or at least, if we are having them, the last arg takes precendence? The ambiguity such a critical format has is astonishing. |
| 05:20 | <royiv> | I think at least WHATWG says "k" and "k=" are equivalent. |
| 06:02 | <annevk> | royiv: if the URL Standard is unclear somehow, please file an issue, but sounds like it’s clear enough |
| 08:42 | <domfarolino> | Per spec, does the concept of a "Document" exist when something like an image resource is navigated to and displayed? |
| 08:44 | <domfarolino> | Specifically what I'm getting at is: Could a link rel=stylesheet header be used to style an image or JSON resource in any way? (Since we of course couldn't include a rel=stylesheet in our HTML, since there is no HTML) |
| 08:50 | <Ms2ger> | Yep |
| 08:51 | <Ms2ger> | https://html.spec.whatwg.org/multipage/browsing-the-web.html#read-media |
| 08:52 | <Ms2ger> | You could put an image in an iframe and then dynamically a link element, if I'm not mistaken |
| 09:34 | <annevk> | Indeed, use an <iframe> and contentDocument it |
| 10:59 | <annevk> | foolip: yt? Mind if I push some changes to your XPath branch? |
| 11:10 | <MikeSmith> | tehres an XPath branch? |
| 11:11 | <MikeSmith> | ah, DOM |
| 11:12 | <foolip> | annevk: not at all, go ahead |
| 11:52 | <hsivonen> | Does the Web Platform have a built-in way to convert between JS Strings and Uint16Array? |
| 11:52 | <annevk> | hsivonen: I don't think so |
| 11:52 | <hsivonen> | annevk: thanks |
| 11:54 | <hsivonen> | I wonder how Qt on Wasm deals with this |
| 11:56 | <annevk> | hsivonen: Uint16Array.from(str) allocates something of the right size, but you need more to get the values out it seems |
| 11:58 | <annevk> | hsivonen: I'd ask fitzgen |
| 11:59 | <hsivonen> | annevk: thanks. I guess the answer to how Qt deals with this is in the Qt source code |
| 12:50 | <nox> | Shouldn't there be a way to view a JS string as a Uint16Array? |
| 12:50 | <nox> | Without allocating anything I mean. |
| 12:51 | <annevk> | nox: maybe, there's been some talk about a StringView thingy |
| 12:51 | <hsivonen> | nox: not possible in cases where the backing storage isn't actually UTF-16 |
| 12:51 | <nox> | Oh right, silly me. |
| 13:08 | <oskarlh> | Not the most efficient solution, but..... Array.prototype.reduce.call(SOME UINT16ARRAY, (str, cc) => str + String.fromCharCode(cc), ""); |
| 13:09 | <oskarlh> | and new Uint16Array((function *() { for(let i = 0; i !== str.length; ++i) { yield str.charCodeAt(i); } })()) |
| 13:09 | <gsnedders> | hsivonen: we should have a way to be able to get a JS string as a Uint16Array where impls don't have to copy the string if it's already a contiguous UTF-16 array |
| 13:10 | <gsnedders> | s/should/should at least/ |
| 13:11 | <hsivonen> | gsnedders: can Uint16Array be read-only? |
| 13:13 | <gsnedders> | hmm, maybe not |
| 13:13 | <gsnedders> | And I don't know how useful being able to do zero-copy would be even if there was, given idk how often you'd end up wanting to copy strings represented as a rope… |
| 13:28 | <hsivonen> | I expect Wasm bindings to need a way to get the string written into a caller-supplied Uint16Array instead of creating a new one |
| 13:28 | <hsivonen> | But I guess I should find the relevant Qt bits before speculating about use cases |
| 13:29 | <gsnedders> | (the ex-Opera employee in me wants to all be like, "but what about Big Endian implementations?!" |
| 13:29 | <gsnedders> | ) |
| 13:30 | <hsivonen> | Does TenFourFox have little-endian Web-visible ArrayBuffer semantics by now? |
| 13:31 | <hsivonen> | The next Debian release will drop big-endian MIPS as an official platform, so s390x will be the only remaining big-endian architecture with an official Debian release |
| 14:18 | <gsnedders> | There are ways to replace the document object while keeping the same window, right? |
| 14:20 | <annevk> | gsnedders: initial about:blank |
| 14:24 | <gsnedders> | annevk: i.e., when navigating away from it? |
| 14:25 | <gsnedders> | (Sorry, I'm far too far away from remembering how document loading works!) |
| 14:30 | <annevk> | gsnedders: yeah, sorry |
| 14:30 | <annevk> | gsnedders: but only if it's same-origin with the initial about:blank and there might be differences across UAs |
| 15:16 | <Ms2ger> | Domenic, bikeshed update fetches from bikeshed-data, which is broken: https://github.com/tabatkins/bikeshed-data/issues/3 |
| 15:17 | <Domenic> | Oh I see |
| 15:17 | <Domenic> | Yeah I guess the web interface uses a different set of data :-/ |
| 15:17 | <Domenic> | TabAtkins is on vacation so we should probably try to work around |
| 15:17 | <Domenic> | Or maybe foolip can help |
| 15:21 | <Ms2ger> | People could add the link-defaults thing for now |
| 15:22 | <Domenic> | Why don't we just fix the promises guide? |
| 15:26 | <TabAtkins> | Ah yeah I turned off my computer before I went on vacation. ...I should put bikeshed-data on some actual infra |
| 15:27 | <TabAtkins> | (am on the beach right now in Haiti) |
| 15:29 | <TabAtkins> | I should be able to run it on my server as well. I think I'm rate-limited on cron there, but I can at least run it a few times a day as a backup. |
| 15:34 | <foolip> | TabAtkins: if you want to run it daily Travis would do the trick, and if more often is needed GitHub Actions can now do that. |
| 15:35 | <TabAtkins> | Kk, I'll look into that when I'm back next week |
| 15:36 | <foolip> | TabAtkins: is bikeshed-data up and running now, or is there a simple command someone else could run to get it updated as a one-off? |
| 15:37 | <TabAtkins> | If I grant access to the repo, anyone can download it and run the command. |
| 15:38 | <TabAtkins> | You want to handle it? |
| 15:39 | <TabAtkins> | Or Domenic ? |
| 15:39 | <TabAtkins> | (it just requires a local Bikeshed install to already exist) |
| 15:46 | <foolip> | TabAtkins: I can do an update, but what is the command to run? |
| 15:47 | foolip | finds https://github.com/tabatkins/bikeshed-data/blob/master/__init__.py#L14 |
| 15:54 | <foolip> | TabAtkins: I figured it out, if you give me write access to the repo I can push an update |
| 15:57 | <foolip> | TabAtkins: if you sign up for the beta for your user on https://github.com/features/actions I could do the automation part too |
| 16:18 | <TabAtkins> | foolip: sent you an invite |
| 16:21 | <foolip> | TabAtkins: received! On my way home now, but can push an update within 6 hours |
| 16:23 | <foolip> | TabAtkins: do you want to do the GitHub Actions thing too? Going by commit times it looks like you've been running continuously, but I'd hourly enough? |
| 16:23 | <foolip> | Is |
| 16:24 | <TabAtkins> | Should be enough yeah |
| 16:24 | <TabAtkins> | I run it every ten minutes right now |
| 16:30 | <annevk> | Oh, it changed from the 24h cycle? I clearly have not been doing much editing as of late |
| 16:31 | <annevk> | Or maybe this is not Shepherd... |
| 16:51 | <foolip> | TabAtkins: https://github.com/tabatkins/bikeshed-data/commit/165a92ef0fcdfafafcd96fbb2aadeb28cf1cd02c pushed, just done manually |
| 16:52 | <foolip> | annevk: this isn't Shepherd, no :) |
| 19:10 | <TabAtkins> | annevk: Shepherd's cycle is still 24h (or less). But bikeshed-data pulls from more sources than that, most particularly wpt which updates constantly. |
| 19:13 | <TabAtkins> | foolip: danke |
| 19:14 | <foolip> | TabAtkins: did you click through https://github.com/features/actions or shall we deal with that next week? |
| 19:14 | <foolip> | Domenic: did my manual update fix anything? |
| 19:14 | <TabAtkins> | I clicked, but it just put me in the beta queue |
| 19:16 | <TabAtkins> | Domenic, annevk: note also the --skip-manifest flag documented at https://tabatkins.github.io/bikeshed/#cli-update if there are every problems with the data being out of date |
| 19:17 | <TabAtkins> | It's just a lot slower than the normal update process |
| 19:17 | <Domenic> | foolip: I'm not sure, Ms2ger is the person to ask |
| 19:17 | <foolip> | Domenic: oh, right |
| 19:23 | <annevk> | TabAtkins: please enjoy the beach 🏖 |
| 19:23 | <TabAtkins> | Oh I am |
| 19:24 | <annevk> | 🥳 |
| 19:27 | <TabAtkins> | https://twitter.com/tabatkins/status/1166794352753963009 |