@bltavares↩️ thanks. I've registered on webmention a while ago, but integrating static with POSSE and micropub to publish on the go tho is not so trivial yet (twitter.com/_/status/1028071386064056320)
ski77777 and tantek__ joined the channel; schmarty left the channel
@voxpelli↩️ I already have a social feed setup using Jekyll: https://voxpelli.com/social/ I can post from apps using my MicroPub endpoint. I can receive mentions and replies from my WebMention endpoint. Building my presence through composable parts that can each shine in their own way. (twitter.com/_/status/1028198622054305797)
petermolnarthe essay-length tootstorms on json-ld are all missing one single thing: the end user perspective. Nobody want's to deal with this, not devs, not users. Instead of all those documents, what they should have done, it a [things] to json-ld converter, where things could be: rss, atom, mf2, mf1, rdfa, microdata, schema.org, etc.
petermolnarbesides, the I still believe json-ld is the wrong way, I'd rather stick to schema.org if I have to choose; at least that's an addon for HTML and not a completely separate entity
jjuranI made a simple optimization to my language's garbage collector, and now the run time for the webmention script I wrote at IWC Baltimore dropped from 15.3s to 6.3s.
ZegnatI am sure I am doing it wrong. But the way I am doing it wrong only means that tools that explicitly *prefer* the older standard over the new one fail. All other tools failing are because they do not implement the new standard at all, and that’s exactly the point, those failures are expected.
ZegnatOnly support old standard: fail. Only support the new standard: succeed. Support both standards, prefer new: succeed. Support both standards, prefer old: fail.
ZegnatI can’t tell nginx not to advertise the old standard as available, which is why the “prefer old” could happen. Still, I don’t mind punishing tools that do not prefer new :P
ZegnatBut my textarea had a width of 100%, so you can click anywhere on the page to get it in focus. And then I used calc() to calculate a left and right padding for sizing the effective text area.
ZegnatI needed a mediaquery to switch from the calc() method to a default padding at small sizes though. Which is where the problems began. I’ll find a different way to do it.
ZegnatDepends if you may want to give Loqi a self-contained interface via DM. E.g. people who have auth’d via DM can also modify the spam filter via DM. That could be neat.
aaronpkneither of these authenticates the current IRC connection. i would need to add some hooks to de-register people when they disconnect or change nicks in order to do that.
aaronpknothing right now sends action without a URL, so if you only are worrying about supporting what works right now then you can make that assumption.
aaronpkbut once i hooked up my site to send replies and deliver my content to my followers, it was easier to switch that to handling the ActivityPub inbox deliveries
Zegnat[jgmac1106]: you have been working with this, do you know if it is possible with css grids to do a 3 column grid where the middle column will never be wider than 20ch?
ZegnatYou need to set something for the middle, sknebel. So that would be 1fr for all, and then a max-width on the middle element? I don’t think that works, because the grid doesn’t change depending on sizing of its items. I think.
aaronpkyeah especially cause this is what XRay sends Accept: application/mf2+json, application/activity+json, text/html, application/json, application/xml, text/xml
Zegnatmaybe it fails because application/activity+json and text/html are of the same q-value (same preference) and you aren’t breaking the tie, rhiaro?
aaronpkspecifically it looks for an "id" property as well as an "attributedTo" property, then it fetches the attributedTo (profile URL) and looks for an inbox
Zegnatsknebel, I think there might have been something weird about the button element again... I should just put a wrapper element around everything that relates to forms *facepalm*
ZegnatThe trick is going to be that H3 should be a button element (submit button for save), but the button text should really always align to M1. So basically it will have a big right-padding extending the clickable area...
ZegnatSo the button you will see in the final design seems to be 1/3 the width of the editable textarea below, but in reality (and important for touch) you can touch anywhere in the top right to submit
ZegnatSure. But just holding your mouse key down doesn’t make anything happen. So I never really understand why I would need to style that. And as soon as you release, the form submits and you will be off of the page not seeing any styling anymore.
ZegnatHere is a thing I haven’t fixed yet: there is a :focus style on the Store button, but the bottom outline falls outside of the element and I don’t know why
ZegnatGot rid of <header> entirely, no more nested grid, just using the main grid from <body> for the three columns now. Liking this design more and more