2014-05-31 UTC
chrissaad joined the channel
benwerd and caseorganic joined the channel
# 01:58 aaronpk wow facebook broke the stupid external link redirects
# 01:58 aaronpk clicked on a link in a facebook message thread from a few months ago and it tells me "Something Went Wrong"
# 01:59 aaronpk even though the link it tells me there's a problem with works fine if I visit it directly
krendil, emmak, mlinksva_ and j12t joined the channel
tantek joined the channel
# 03:22 KartikPrabhu tantek: I know how to do www to non-www and http to https separately but was hoping there would be a blurb to do both in one fell swoop
# 03:31 KartikPrabhu the interwebs don't seem to know either. I found the individual redirect codes for both cases but no one-shot solution to both
# 03:32 rascul best to put that in the config though and not use htaccess
# 03:32 rascul disable htaccess altogether if you can, htaccess is mostly just extra file reads for no good reason
dybskiy joined the channel
# 03:35 rascul not gonna hurt anything, i was just mentioning a better way :)
tantek, catsup, rektide, jacus, binbasti, ddysart_, danfowle1, dybskiy, nloadholtes and KevinMarks joined the channel
# 05:32 bear you want to redirect www to non-www first otherwise you will be then redirecting https www which will most likely get a cert error unless you have a wildcard cert
# 05:36 KartikPrabhu bear: true but I wanted to avoid multiple redirections and do the whole thing in one swoop
# 05:36 bear I've never seen it down with .htaccess
# 05:36 bear because of how rewritecond rules are handled
# 05:38 bear you visit a url for a blog, that will get sent to a tracker, then sent to cdn and then to the content
# 05:39 bear this is the fanciest i've seen it - but my htaccess fu is very old
# 05:39 bear RewriteCond %{HTTP_HOST}
^(www\.)(.+) [OR]
# 05:39 bear RewriteCond %{HTTP_HOST}
^(www\.)?(.+)
# 05:41 bear but all of the above is why I hug my nginx server every night
# 05:41 bear because your building an OR for https and not https checks
# 05:42 bear no - with nginx you can specify the server name to be www.example.com or example.com
# 05:45 Loqi Ok, I'll tell them that when I see them next
dybskiy joined the channel
glennjones, smcgregor, dybskiy, brianloveswords and squeakytoy joined the channel
carlo_au joined the channel
garietyxxx joined the channel
Garbee_ joined the channel
# 09:39 garietyxxx Shouldn't the http/https be resolved automatically y indieauth?
# 09:42 KartikPrabhu garietyxxx: do a !tell to aaronpk... I'm sure he'd like to fix that :)
netweb joined the channel
# 09:52 Loqi Ok, I'll tell them that when I see them next
bnvk, edge226, tantek, barnabywalters, nloadholtes, dns53, friedcell, garietyxxx, carlo_au, ttepasse, brianloveswords and chrissaad joined the channel
carlo_au and KevinMarks_ joined the channel
# 16:22 aaronpk !tell kylewm thanks, fixed! I switched the site over to ssl-only and forgot to update the setting in the instagram app
# 16:22 Loqi Ok, I'll tell them that when I see them next
# 16:26 bnvk goin ok, got the formatting pretty much down
# 16:27 bnvk figuring out downloading of attachments now
# 16:30 Loqi kylewm: aaronpk left you a message 7 minutes ago: thanks, fixed! I switched the site over to ssl-only and forgot to update the setting in the instagram app
snarfed joined the channel
# 16:44 aaronpk I think I can avoid all the rabbit holes of comparing equivalencies of URLs by checking to see if the actual username from the providers matches
# 16:50 aaronpk no because they all return a username after authenticating
# 16:51 aaronpk well yes, the known URL pattern happens first, so if I see a rel-me link like "http://twitter.com/aaronpk" then I know what username to expect
# 16:52 aaronpk so if I just don't try to reconstruct it and instead just compare the username afterwards then it should avoid that problem
# 17:08 aaronpk ah crap but google returns a "uid" which is not the username
# 17:25 aaronpk alright i'm just gonna have to hackily compare URLs
bnvk, ScruffyDan and petermolnar joined the channel
# 17:54 aaronpk ok... pushing out a change. would love some help testing
# 17:57 petermolnar although my https-only page only works if I explicitly enter the whole url
# 17:58 petermolnar it's semi-fixed: it now finds the urls but tries to backlink to http instead of https and therefore the urls are invalid
snarfed, j12t and garietyxxx joined the channel
# 18:31 snarfed aaronpk: worked for me w/twitter, g+, and github (all https)
dariusdunlap joined the channel
# 18:32 Loqi dariusdunlap: tantek left you a message on 5/29 at 7:18pm: well done with IndieAuth login to your blog! Is that a plug-in to add to /WordPress ?
# 18:32 aaronpk snarfed: but you probably entered https in the sign-in field?
# 18:33 snarfed but re-scanning complains about all providers, since all rel-mes link to https
KartikPrabhu joined the channel
# 18:36 snarfed complains that it can't find rel-me links for tw, g+, and gh
# 18:37 aaronpk right, doesn't work right now because of what I just described there ^
# 18:38 snarfed is google authenticator fully supported yet? scanning the barcode worked, but i couldn't find a description of what the rel-me link would look like
# 18:38 aaronpk yeah it works, but it's more like registering an account on indieauth.com
# 18:39 aaronpk but it does work, I use it all the time, especially on slow networks because it avoids all the http requests involved in oauth requests
# 18:40 aaronpk go try to sign in to the wiki, and you'll see it listed as an option
# 18:41 snarfed minor feature request: maybe make enter in the token text box submit
jonnybarnes joined the channel
# 18:57 Loqi jonnybarnes: luxagraf left you a message on 5/30 at 1:32pm: - thanks for all your help debugging my ssl issues. everything is working now.
caseorganic joined the channel
# 19:11 aaronpk petermolnar: snarfed: ok I think I fixed it, so if you enter http for your URL
# 19:14 snarfed (and btw, obviously don't feel obligated to fix this for me…but i'm happy to help test!)
# 19:14 aaronpk ohh does your site respond on both http and https?
edge226 joined the channel
# 19:15 aaronpk should work for petermolnar since he returns 301 on http
tilgovi joined the channel
# 19:17 aaronpk without the 301 link, your http and https pages are technically completely separate URLs
# 19:17 aaronpk whereas returning 301 from http to https makes them equivalent
# 19:17 aaronpk oh goddammit... now there's a trailing slash issue
# 19:25 aaronpk i'm making a bunch of changes right now so i'm not surprised, heh
# 19:26 jonnybarnes clicked log in, typed in url, clicked google link, was back at indiewebcam.com but not logged in, so I tried again, and was then logged in
Kopfstein joined the channel
glennjones and chrissaad joined the channel
# 20:49 KartikPrabhu apparently I can't query multiple post types in Django without creating a whole another database with all the posts in it! :|
friedcell, tantek and nloadholtes joined the channel
# 21:54 KartikPrabhu aaronpk: I have noticed that sometimes Quill does not post anything when I use the Homescreen bookmark. but it works if I re-log in. Any reason?
# 21:55 aaronpk you mean you click the home screen bookmark and see the interface, but a post after that fails?
# 21:55 aaronpk does the debug output show you anything useful in the http headers?
# 21:56 KartikPrabhu no. it still shows the last response. I think it doesn't even hit my server as I return some response for everything
# 21:57 aaronpk clicking "post" should make it show something in the debug section under the button, do you mean it doesn't even do that?
# 21:59 KartikPrabhu no. the fields I filled become blank, but the debug section still shows the "Last response received" blurb
# 21:59 aaronpk hmm strange! could you check the browser console to see if it's getting some weird error there?
# 21:59 KartikPrabhu which is the very last successful one, not the one you'd expect out of the currect micropub request
brianloveswords joined the channel
# 22:00 tantek.com edited /site-deaths (+249) "move justin.tv to upcoming since June 8 hasn't happened yet. Editorially user data backups available til 2014-12-31. update dates for loom, myopenid checks" (
view diff )
# 22:02 tantek going to move them to company specific pages accordingly
grantmacken and snarfed joined the channel
KevinMarks joined the channel
# 22:36 KartikPrabhu aaronpk: unable to repeat that situation now. WIll keep an eye on it though
scor joined the channel
ttepasse and j12t joined the channel
# 22:50 aaronpk thoughts on whether it's worth trying to resurrect OpenID support for indieauth.com?
# 22:51 aaronpk I'm thinking no, given that the OpenID group has moved on to Open ID Connect
# 22:51 aaronpk also becomes less important if indieauth.com can auth against your site treating it as an oauth provider (basically that's what openid connect is)
# 22:57 tantek aaronpk - until someone asks for it specifically (i.e. their own site natively acts as an openid provider for themselves), we can probably punt it
# 22:59 aaronpk yeah, I know of at least a couple people who are their own openid provider who have mentioned it, but I think it's probably not important anymore since it's officially deprecated and everything
edge226 joined the channel
# 23:10 aaronpk I do still want to fix the openid provider side so it can be used to sign in to pypi.org
scor joined the channel
# 23:28 aaronpk so.. pypi is sending as assoc request with session_type set to an empty string.
# 23:31 aaronpk I had to force session_type to 'no-encryption' rather than an empty string. that would appear to be a bug in the ruby-openid gem
KevinMarks joined the channel
# 23:40 tantek.com edited /Yahoo (+656) "Citizen Sports looks like it was just an app/tool, minimal user data if any, no sign of any site data" (
view diff )
# 23:44 tantek huh, looks like Yahoo restored the Yahoo! WebPlayer
chrissaad joined the channel
# 23:55 tantek.com edited /site-deaths () "(-991) citizen sports no sign of site content, Yahoo! WebPlayer was scheduled for shutdown, but restored/rescued somehow. also a service, not a content site" (
view diff )