| 14:37 | <Domenic> | devsnek: there is no way to modify the behavior of a writable so that when you write to it it goes into some other readable instead of its original underlying sink. Similarly there is no way to modify the behavior of a readable so that when you read from it it reads from a writable instead of from its original underlying source. |
| 14:37 | <Domenic> | So yeah focusing on the use case and some other solution makes more sense to me. |
| 14:38 | <devsnek> | cool ok |
| 14:38 | <devsnek> | also |
| 14:38 | <devsnek> | there was an issue about duplex streams open on whatwg a long time ago |
| 14:38 | <devsnek> | it got closed but was seemingly never solved? |
| 14:44 | <Domenic> | Any object containing { readable, writable } is a duplex right? |
| 14:46 | <devsnek> | Domenic: sure but thats kinda messy |
| 14:46 | <Domenic> | Why? |
| 14:49 | <devsnek> | Domenic: i think the concept of a stream with data flowing in two directions is primitive enough to warrant a unique type |
| 14:49 | <devsnek> | an object where you could getReader and getWriter as an example |
| 14:49 | <Domenic> | Nah I think Node's conflating of two separate streams into a single object causes a ton of pain |
| 14:50 | <Domenic> | Readable and writable streams as separate objects that never have to occupy the same space is great |
| 14:50 | <Domenic> | And then you have this general concept of "two related streams" represented by { readable, writable } |
| 14:50 | <Domenic> | That could either be duplex, or transform. The relationship between the two sides is up to the creator of that pair. |
| 14:50 | <Domenic> | But pipeThrough() works with anything that conforms to that object signature |
| 14:51 | <Domenic> | (As does any userland code that assumes the same pattern) |
| 14:53 | <devsnek> | so { readable, writable } is a good pattern to follow? |
| 15:08 | <JakeA> | Do we have a webidl way to create async iterators? If not I'll open an issue and propose prose |
| 15:08 | <TimothyGu> | I don't think so |
| 15:09 | <Domenic> | devsnek: yeah it is |
| 15:10 | <Domenic> | I guess we should doc this more |
| 15:12 | <devsnek> | i btw i opened https://github.com/whatwg/streams/pull/950 |
| 15:13 | <Domenic> | Oh right yes I am excited to look at that today |
| 15:13 | <Domenic> | But, I specifically put a large block on my calendar starting in 15 minutes to work on the hard thing I've been procrastinating on at work. So, later. |
| 15:14 | <devsnek> | 👋 |
| 15:51 | <JakeA> | TabAtkins: is it possible to export a dictionary? I'm trying to reference https://w3c.github.io/manifest/#dom-imageresource from another spec |
| 16:13 | <TabAtkins> | JakeA: yeah def. You're wanting them to mark it up so Bikeshed can target it? |
| 16:16 | <TabAtkins> | I think I can just point Shepherd at it, actually - as long as there's an idl block and a dfn, the heuristics should hook them up. |
| 16:17 | <TabAtkins> | I've just arrived from Buenos Aires tho, so I'm dead right now |
| 16:19 | <gsnedders> | That corpse is typing? disconcerting. |
| 16:50 | <JakeA> | TabAtkins: hah, no probs. Get some sleeps. |
| 18:51 | <howdoi> | is there a way to add animation to the splash screen that appears via the manifest.json? (does it even makes sense to do so?) |
| 21:00 | <gsnedders> | also: ergh, we could really do with someone maintaining CSSOM/CSSOM View |
| 21:14 | <TabAtkins> | gsnedders: emilio is the editor now |
| 21:15 | <ecobos> | TabAtkins: technically not from cssom-view though. I guess I should, I've been looking at compat issues in the area lately |
| 21:15 | <ecobos> | gsnedders: if there's any issue that is particularly urgent let me know and I'll try to get to it ASAP |
| 21:17 | <gsnedders> | ecobos: mostly just actually defining terminology btter for other things to hook into, but I think that actually runs into the lack of definition of layout v. visual viewport :( |
| 21:18 | <gsnedders> | (about to change trains, will respond more when I get home) |
| 21:19 | <ecobos> | k, ttyl |
| 21:24 | <gsnedders> | or alternatively that train arrived just when there's a 20 minute gap between local trains :( |
| 21:30 | <gsnedders> | basically I want easy to refer to terms of outerWidth (width of the "client window"), innerWidth (width ofthe viewport *including* scrollbars), clientWidth (width of the viewport *excluding* scrollbars) |
| 21:31 | <gsnedders> | which is a bit weird given it isn't totally clear whether the viewport actually includes the scrollbar or not in CSS2 |
| 21:31 | <gsnedders> | (which also affects things like vw) |
| 21:33 | <gsnedders> | really all of this is a bit awkward because most of the underlying things are deliberately undefined because we don't want to assume too much about the UA |
| 21:33 | <gsnedders> | but these things are all web exposed in one way or the other |
| 21:33 | <gsnedders> | c.f. that long discussion about what behaviour should be with disappearing scrollbars in some CSS F2F last year |