Vendanjust odd question, does anyone else use a language that requires a build step, or is mostly everyone on "source code is the executable" languages?
kylewmnothing too exciting about it really, html5lib just punted on implementing the "javascript disabled" case, and didn't bother to parse noscript tags
tommorrisalso be sure to throw salt over your shoulder, not walk under any ladders, avoid black cats crossing your path, knock on some wood and pray to the lesser gods that the OpenSSL devs are doing their thing competently.
barnabywaltersmorning rhiaro — php-mf2-shim doesn’t touch the twitter API at all, it scrapes information out of twitter.com HTML — what problems are you having with it?
rhiarobarnabywalters: I haven't tried it yet, I just noticed the example had "client($twitterClientID, $twitterClientSecret, $twitterAccessToken, $twitterAccessTokenSecret);"
barnabywaltersaaronpk do you actively use your twitter shim? Would be interested in knowing what prompted use of the API over HTML scraping — stability?
barnabywaltersfwiw I’ve had minimal problems with the HTML scraping approach, but that could change at any moment. After all, Twitter are obliged to let people know about API updates but not front end code changes
GWGWell, basically I want to have a generic post display object and specific post type objects that use the same functions so I can detect and pass an object of the correct type
Vendan"I'm using it more for the common interface" vs. "you mentioned interfaces, but that doesn't work with having a generic functionality replaced by more custom ones"
barnabywaltersgreat! I will certainly have a look over it, I considered building one as Go practise a while ago but always had other things to do. will be interesting to see how you’re implementing it
barnabywaltersVendan: probably not, although I’ve been considering rewriting my feedreader backend in Go for the websocket support. I do have some relevant mf parsing experience and might be able to provide helpful hints though
barnabywaltershm, my second issue is actually not quite right. The contents of <base> are taken into account, but are not resolved against the document URL
barnabywaltersaaronpk: apache, but I’m considering changing to nginx for a really weird obscure reason to do with custom HTTP methods and audio streaming
aaronpkbasically it means i can make php apps that speak websockets without dealing with any of the weirdness that comes with actually doing websockets in php
barnabywaltersaaronpk: that’s really cool! One of the reasons I wanted to use Go was because it was self contained with fewer moving parts, which still stands, but this looks fun as an interim solution
barnabywaltersI’m like modular approaches to things right up until they mean I have a dozen extra bits of software to compile in a particular way and hook up together in a particular way :)
Loqibarnabywalters meant to say: I like modular approaches to things right up until they mean I have a dozen extra bits of software to compile in a particular way and hook up together in a particular way :)
barnabywaltersVendan: i.e. when the go binary which is doing websockets stuff is the same one which is acting as a machines external-facing server, with no reverse proxies inbetween?
VendanI've got a collection of other indieweb libraries too: webmention, distributed indieauth, and PuSH. And there's a micropub endpoint in my main server app, but there's not too much point in pulling that out.
barnabywalterstantek: maybe the best solution is to add a video_width parameter to auto_link, so that it doesn’t have to make any assumptions about the context
tantekand if it doesn't, then consider filing bugs on whatever browser you're using, and feel free to use my auto-embedded webm posts as real world examples
cweisketantek, not exactly another level of complexity. it's just that you have a permanently open connection for each browser, which does not work with the traditional model of web servers
VendanI'm not sure about data-uri for images, I'm planning on making a post-process step for my micropub endpoint that pulls the images out, decodes them, and rewrites the tags to point at them
aaronpkwhen you upload an image into the editor, the editor actually sends it up to the quill server and the quill server is supposed to return a URL for where the image lives. I'm short-circuiting that by just returning the data URI
aaronpkso the question then is... what's the best way to do raw image/file uploading in micropub, where uploading the image doesn't actually create a photo post
kylewmI have this problem in general with post editors -- you want to upload the image but the post hasn't been created yet, where do you upload hte image?
Vendanit'd be a little more complex, but we could go the route of email, and do basically a multifile post to the micropub endpoint with the photos all as files
Vendanyeahm I know that, hence the "I'm planning on making a post-process step for my micropub endpoint that pulls the images out, decodes them, and rewrites the tags to point at them"
aaronpkyeah, seems like the post-processing of updating the <img> tags is gonna be about the same whether the img tag contains an http URL or a data URI
snarfedPSA for bridgy users: if you've been missing any facebook responses over the last few weeks, try re-authing. (FB recently granted us a few new permissions.)
tantekaaronpk: re: "where uploading the image doesn't actually create a photo post" why not? that is, why not have uploading the image create a private (you only) photo post, and then the article authoring UI allow you to use any photo from your photo posts, or enter a URL of a .jpg/.png./gif etc. to embed?
aaronpktantek: i support private posts right now where i whitelist people who can see it. it was the same workflow/code to make it support visible to only me or to only certain other people
benborgesso i came back from a www.neutrinet.be meeting (associative internet provider at early stage) where we got a presentation of this : http://labriqueinter.net/ and this https://yunohost.org/
benborgesbut seriously, this box is crazy ! just imagine your own indieweb self hosted website at HOME running on a Debian 7 on decent hardware at home
benborges"website in a box" this means, we could customize a Known or wordpress instance, tie it to a domain, and have this ready for a gift, it means we could distribute the indieweb phisically, in people's home !!
benborgesanyway, i'm going to sleeep, because tomorrow i want to assemble this little box and SSH into it and actually move one of my domain to the box, spice it up with some indieweb and then see how it goes. this box also solve a big issue to me : email, i'm finally going to selfhost part of my emails, it's going to be fun !
aaronpk“No one moves where they tweet because some other tool has better formatting or profile customization. That’s because a tiny percentage of the value Twitter brings comes from the software itself. It’s all about the network”Š—”Šthe connection with other users and the content they create.”
aaronpkand ev highlighted this: "the ability for Medium responses to live on their own gives both more motivation to invest in them (as a creator) and more likelihood the significant ones will be found"
benborgesi never got into livejournal, always felt akward in front of it, and at the i was compulsively registering to anything new that was arising, blogger was my first blog, then moved to b2evolution, then wordpress
benborgesi was first a music producer, under CC, in 98 uploading mp3 online and retaining control on this data was not easy task, but that challenge ultimately got me into more web, less music, until it became my job, haha :)