#bearhave you tried to add the git url ones to setup.py's depednecy_links arguement?
#bearthis allows you to reference a package in install_requires but not have it to be on pypi
#KartikPrabhubear: yes that is what I am doing now. But gives this error for ronkyuu: "Could not find a version that satisfies the requirement ronkyuu>=0.2.8 (from connection) (from versions: 0.2.1alpha, 0.2.2, 0.2.3, 0.2.4, 0.2.6, 0.2.7)"
#KartikPrabhuseems like somehow 0.2.8 does not exist even though it is at 0.2.8 at the moment
#KartikPrabhuall these things suggest to me that people really have no intrinsic problem with silos... unless they do something bad which they inevitably do
#bearsilos are never perceived as bad in the beginning
#bearbecause early adopters all have confirmation bias about using them
#KartikPrabhuthe strange bit is every silo gets a fresh clean start wrt to people's opinion and everyone moves to it... people don't see a pattern
#KartikPrabhumy current thought is that even if indieweb tools become easy to use, people won't switch to them because "why"?
#KartikPrabhuclearing "owning your data" is not a motivation for most people
#LoqiKartikPrabhu meant to say: clearly "owning your data" is not a motivation for most people
#bearthe key is, IMO, even IndieWeb is a silo - but it's how big of a shadow does the silo cast
#bearwith IndieWeb the "shadow" is a person or their family
#bearand as long as the interop is all open formats / standards - then that is a Good Silo (tm)
#KartikPrabhubut then you are also responsible for your own data. Now if my Twitter account is compromised I can yell at twitter. But if my indiewebsite gets compromised it is purely my fault
#tantekKartikPrabhu: keep cranking on your own site
#tantekeventually they'll see *something* (unpredictable what that will be) that you've done there, and it will inspire them to ask you how you did it, and how they can do it too
pbeaulieu joined the channel
#KartikPrabhutantek: well personally I don't really care if people are themselves excited by this or not. Just a pattern of not caring I have observed
#KartikPrabhuparticularly since I am on Twitter and FB and G+ they just use that anyway
#tantekKartikPrabhu: then perhaps consider figuring out how to reduce your own use of Twitter, FB, G+
#tantekby replacing it with using your own site instead
#KartikPrabhui exclusively tweet through POSSEing but I still POSSE it
#tantekthat's fine. are you also POSSEing to FB and G+ or are you still silo posting there?
#KartikPrabhuPOSSEing to G+ too.... don't really do any FB anymore
#KartikPrabhuI have also stopped posting comments on G+ as G+ comments don't have permalinks
#tanteksuggestion for http://indiewebcamp.com/User:Kartikprabhu.com - how about moving the "Done" things to their own section at the bottom, and re-assess your itches to see which of those you're most excited about solving
#tantekHave you written a post (even brief) about how you "have stopped posting comments on G+ as G+ comments don't have permalinks" ?
#tantekthat would be good to see written up as an explicit policy decision
#KartikPrabhuno. have been thinking of work-arounds for now
#tantekalso, KartikPrabhu re: Ghost - we have pretty decent documentation of it at /Ghost - feel free to add to that if you know anything knew. I for one am not too optimistic about it.
#KartikPrabhuerr it is the package manager for python that does not work for some simple things I want to do
#tommorrisI'm just debating whether or not to apply to be an invited expert for the W3C Social WG.
#tommorrison the one hand, tantek is involved which means it's probably not going to be too much of a pointless gabfest and might actually get shit done.
#tommorris"This questionnaire is open for answers until 23:59, Boston time on 2060-12-31."
#tommorrispushes the buttons and joins the IRC channels and hopes things don't turn out to be too irritating.
fourtonfish joined the channel
#kylewm.comedited /User:Kylewm.com (+389) "/* Itches */ added photo albums to itches, note declining interest in building stuff for hypothetical gen 2ers instead of for myself" (view diff)
#tommorristantek: so I've just applied as an invited expert to the Social WG
#tommorrison the basis that I'm a social software developer, I know some things about both microformats and upper-case Semantic Web, and I know more about the Facebook API than is healthy for one person to know.
#tommorris06:09 <tommorris> on the one hand, tantek is involved which means it's probably not going to be too much of a pointless gabfest and might actually get shit done.
#tommorris06:09 <tommorris> on the other, W3C WG...
#tommorrisI'll log into w3 wiki and add myself to the list of people who've applied for membership.
#willnorristhe problem with auto-highlighting the entire textarea is that it prevents highlighting a subsection of it. This drove me crazy on https://webmention.herokuapp.com/ where I wanted to highlight just the “/api/webmention” URL, not the full <link> element (/cc voxpelli). Tantek’s URL box is a little different since it’s less likely that you want to select a substring. I’ve not found a site that strikes a good balance here.
#voxpelliwillnorris: ah, yeah, I know what you mean – drives me crazy as well, not sure why I did it that way :/
wolftune joined the channel
#voxpelliwillnorris: one could probably do a better one which only selects the entire text if no selection has been made, but haven't kept up to date on the javascript API:s for selecting text so not sure how complex that would be
#willnorrisyeah, something like that would be ideal
fmarier joined the channel
#willnorrisso I think this is kinda neat. I now have GET requests to https://willnorris.com/api/webmention served as a static site with a webmention form, but POST requests to the same URL are proxied to voxpelli’s webmention.herokuapp.com. Best of both worlds: a customized “visible endpoint that teaches” while still using an external service, which is necessary for static sites. (blog post and nginx config forthcoming when I have a moment)
#aaronpkwillnorris: nice! Mine only has a link to the wiki for a get request
#tantekI like this whole idea/pattern that API endpoints should provide user friendly HTML responses to a GET.
#willnorriscurrently, submitting the form displays the JSON response from the heroku app directly in the browser. I still need to add some JS to send the request and display a nice response.
Acidnerd joined the channel
#aaronpkMy HTML form under each post adds a hidden field called "from_browser" or something, and if that's set, my webmention endpoint returns an HTML response
#aaronpkThe Webmention spec doesn't actually say you need to return JSON
#tantekwillnorris: perhaps simply redirect to the TARGET URL which would display the result of processing the webmention
#aaronpkOh yeah wait maybe that's what mine does. Can't remember what I decided
#tantekaaronpk same thing - rather than an HTML response saying "got it" etc, just reload the page and perhaps highlight the new comment, fave, repost etc.
#tantekmakes sense, for testing, documentation, discovery, and user friendliness
#voxpelliwillnorris: updated the herokuapp now – if you come up with any more ideas for tweaks and extensions then I'm all ears – here, on GitHub etc :)
#tantekbret - would it be reasonable to stub a /access_token article with brief dfn and link to oauth/indieauth for more details?
catsup joined the channel
#voxpellisnarfed: yeah, saw that one – thinks this is rather something editing multiple pages at once and sending pings for +10 pages in a very short amount of time
barnabywalters joined the channel
#voxpellisnarfed: my endpoint is currently set to never fetch a page from a domain more often than one per minute and to never queue more than 10 pages – might be a bit too hard from my side
#snarfedvoxpelli: probably yes. maybe use something like a leaky bucket, so you throttle over time but still allow bursts
#voxpellisnarfed: I let 10 pages queue up and fetches them over time, but perhaps I should allow more – since I don't save them to the database but just keep them in memory right now I don't want to accept too many – rather have the pinger retry later on themselves
johncash and indie-visitor joined the channel
#LoqiWelcome, indie-visitor! Set your nickname by typing /nick yourname
#KartikPrabhuso the latest Medium backlash suggests that people do really want to put things in a pre-made template
#bearwillnorris - I wonder if you could do that redirect via nginx by having the heroku app send back a POST for success to a url with the url as a query parm
#bearyou should be able to sniff that query parm (or even header?) in the /success location handler
#willnorrisbecause I’m not sure if this realistically can be done without JS (though I’d certainly be interested to see what bear comes up with). I’m using nginx as a reverse proxy to forward this request to heroku, so I’m not sure what all I can do with that response… I’ll have to read up on that
#willnorrisin which case I feel even less inclined to come up with a non-JS solution :)
#willnorriswell, I do have a pretty page for the initial GET request (http://willnorris.dev/api/webmention). I’m not sure if I can still serve a custom response page for POST requests since I’m doing the reverse proxy to heroku. the nginx module bear mentioned may be one option
#bearyea, i'm researching this now and it seems convulated
chrissaad joined the channel
#willnorris(I’m also trying to avoid building nginx from source, which it looks like that module may require)
#willnorriserr, no nevermind. apt’s nginx-extras includes it, which is what I’m running
#tantekI think I remember JonathanNeal sharing that a while ago
tantek joined the channel
#tantek!tell aaronpk, good news / bad news. remember those session start errors we were seeing on tantek.dev/falcon ? now the same errors are happening on tantek.com/falcon, but I still haven't figured out what php.ini etc. magic I need to fix it - in either (or both) place(s).