#caseorganictantek: it was perfect to have this discussion early on, as it set the roadmap for where people currently where and where they wanted to go, as well as educated newer attendees
#gjonesslowly getting there - building a blog engine from zero is always a bigger job than you think - noted image size issue, trying to fix the file uploading to add a different image size at the moment : )
#neuro`gjones: what are you developping your blog engine with?
#tantekcan you pick a nice cafe near downtown that is good for working with laptops (has wifi) ?
julien51 joined the channel
#@aralWorking on my new talk on avoiding Digital Feudalism by making Experience-driven Open (#xo) products to realise the goals of #indiedata (twitter.com/_/status/378095962650316800)
smus, matthewlmcclure_ and julien51 joined the channel
#ShaneHudsonHas anybody yet done POSSE and/or web mentions with a flat file cms? I'm making my own and not sure whether to make it hybrid, with a database, or not
#cweiskeI always wonder what people think they are gaining when using flat files and not a database
#ShaneHudsoncweiske: I don't have a problem with databases but quite like the way it works, plus I know database corrupting won't lose content etc
#ShaneHudsonMy cms wasn't originally going to be flat, but ended up that way once I got into the details of planning
barnabywalters, cyberjar09, andreypopp and bnvk joined the channel
#pfenwickHow to get rick-rolled (multiple times) in one easy step: Tell the world that tweeting to you with a #todo tag will add that tweet to your todo list automatically.
#barnabywaltersI’ve just learnt that lack of a good relative URL resolving technique plagues javascript as well — apparently this is a widespread issue :/
#tantekadactio, assuming we can getup and get to Taylor St by 10:45 on Saturday morning, we should be good to go.
#barnabywaltersit mods twitter.com, detects which notes are POSSEd copies of content you own on your own domain, pulls the original in and corrects permalinks and such things
#tantekhow many indieweb folks' POSSE copies does it work with?
#barnabywaltersit works for aaronpk and my own notes, not for tantek’s yet because of the extra complexity and large quantities of extra HTTP requests required
#aaronpkso far I've got two separate attempts, one where I just made stuff up, and one where I am attempting to implement the pseudocode in the RFC
#@aralComing to the sad realisation that UK is basically as bad as US (if not worse in some ways). Which sucks ’cos I love Brighton. #indiedata (twitter.com/_/status/378170507113541632)
#tantekaaronpk - I hesitate to ask - which RFC? ;)
#barnabywalterstantek: I have to request each trailing URL in tweets, which I’m doing at the moment. In your case I also have to resolve all the rel-syndication URLs on the potential POSSE to see what they redirect to
#tantekaaronpk - I'd say use the URL Specification instead
#barnabywaltersif the webignition url library is okay, why not just copy+paste their URL resolving code? the dependency problem they have is nothing to do with URL resolving
#tantekbarnabywalters ok I can see how that might be a problem
#barnabywaltersthey require the intl extension to easily handle IRIs
#aaronpkbarnabywalters: a few of the tests in the webignition code seemed wrong
#barnabywalterstantek: it’s fine for my original post discovery endpoint because I’m caching the results, but for browser-side code it’s going to be making a *lot* of requests
#tantekbarnabywalters - yes, it's a real issue. it's not reasonable to have to follow all the rel-syndication links to look for redirects
#barnabywaltersaaronpk: okay then, let’s all ditch PHP and use Go. they already wrote the function ;)
#tantek*however* what if we filtered by domain when looking for URLs to resolve to try for redirect matches?
#tantek.comedited /original-post-discovery (+453) "/* Algorithm */ check syndication URLs one at a time and stop when match is found, add more detail to the algorithm and nesting of steps too" (view diff)
#tantekbarnabywalters ok I used a specific name for each
#barnabywalterstantek: looks good, might benefit from added separation between candidate URL discovery and validation process. I might write a pseudocode version
#tantekbarnabywalters - bolding the terms is a good suggestion
#ShaneHudsonaaronpk: Hopefully that means you're very close!
#tantek.comedited /original-post-discovery (+1228) "add Algorithm Notes and Use-cases used to design/add the "syndication URL has the same domain as the POSSE permalink" steps" (view diff)
#ShaneHudsontantek: Heh makes sense if there is no one to play with! Great idea though :)
#ShaneHudsonI've always been careful to avoid posting about chess though... once upon a time someone called Shane Hudson was a programmer/chess player/scammer. It took a lot to knock him off of google lol
#tantek.comedited /original-post-discovery (+140) "bold terms in algorithm (and make them consistent) to make them easier to follow (almost like variables / parameters!)" (view diff)
#bnvkMet some guy from Harvard's Berkman Internet & Society who has some open social / decentralized project yesterday http://idcubed.org I told him about IndieWeb and was showing him demos
#aaronpkas in, i'm not sure if the tests are actually correct
#Loqibarnabywalters meant to say: aaronpk: all then tests passing? awesome. I will spend this evening trying to break it, then fix/merge as necessary
#ShaneHudsonDon't suppose any of you know webrtc inside out? I've got a strange problem with FF<->Chrome
#ShaneHudsonWell, strange in that it doesn't work and I see no reason why not!
#barnabywaltersaaronpk: looks like it’s working well enough to merge at the moment — user+pass URLs are not a big deal
scor joined the channel
#aaronpkbarnabywalters: *almost* all tests passing
#aaronpkthere's something wrong with resolving base "http://example.com" path "http://example.com/" to "http://example.com/"
#tommorrisbarnabywalters: you going on Saturday to Brighton for the chit-chat about microformats2 parsers? ;)
#barnabywaltersaaronpk: from what I can see of the test results, the failing tests aren’t a huge deal. I can’t see any which are in common use which are deal-breaking
#aaronpkbarnabywalters: how do you run a specific test with phpunit?
#barnabywaltersthere are two things which need doing for php-mf2 v1.0.0 — making the URL resolving work, which aaronpk seems to have a handle on, and making consistently escaped output the default (only) behaviour
#aaronpkI should probably package that up as its own composer library
#barnabywaltersaaronpk: absolutely. want to do a functional URL-focused lib? there are a bunch of smaller things I’ve implemented which need a better home
#bnvktantek: yah, I find the anti-pattern concept really useful. This fellow definitely does not "dogfood" he wants to create a shiny, polished, beautiful, and widely used product- immediately, heh!
jschweinsberg joined the channel
#barnabywalterstantek: if I wrote a test suite for cassis, would you use it? if I’m going to start adding stuff I want to be testing it
#tantekI do have private test code I should upload
#barnabywalters!tell tantek your homepage feed permalink URL mf are parsing strangely, did you change it recently? Looks like you have class=value on children of the .u-url element, and I’m applying value-class pattern to everything, not just datetimes. Pretty sure what I’m doing is spec compliant, do we need to change anything?
#barnabywaltersaaronpk: okay, php-mf2 v0.1.21 has NO DEPENDENCIES YAY thanks to you. It seems to work fine, I’ve deployed it. when can we expect to see pin13 pick it up?
#Loqitantek: barnabywalters left you a message 5 minutes ago: your homepage feed permalink URL mf are parsing strangely, did you change it recently? Looks like you have class=value on children of the .u-url element, and I’m applying value-class pattern to everything, not just datetimes. Pretty sure what I’m doing is spec compliant, do we need to change anything?