#indiewebcamp 2016-04-15

2016-04-15 UTC
rrix, shiflett, KevinMarks and FergusL joined the channel
#
GWG Evening
Vendan, shiflett_ and Garbee_ joined the channel
#
KevinMarks call for /mood support in twitter: https://twitter.com/puppytube/status/720815422610669568
#
@puppytube why does twitter keep adding useless features when this is the only feature that would improve the website [photo] (twitter.com/_/status/720815422610669568)
#
KartikPrabhu "improve" :P
cleverdevil- joined the channel
#
bear they want to make twitter look like that personal blog software that was popular a while ago
#
bear (dang, cannot thing of the name)
#
KartikPrabhu FB gave us moods by expanding "likes"
#
GWG What is Livejournal?
#
Loqi LiveJournal is a declining blogging silo https://indiewebcamp.com/LiveJournal
#
bear yea, LiveJournal!
#
bear gwg++
#
Loqi gwg has 152 karma
#
GWG I'll take declining silos for $500, Alex.
#
bear :)
#
GWG acegiak did something with mood lately.
#
KevinMarks hm, we should credit LJ wiht all the things it invented
M-RyanRix joined the channel
#
kylewm livejournal++
#
Loqi livejournal has 1 karma
[dym_cx] and [kylewm] joined the channel
#
[kylewm] huh, is the slack bridge down?
[dym_cx] joined the channel
#
[dym_cx] something's wrong
#
Loqi [dym_cx]: [shaners] left you a message on 4/8 at 6:55pm: I fixed the logo knockout. re: "the c carving is somehow a bit short, in comperrison with the other two" http://indiewebcamp.com/irc/2016-04-08/line/1460166946667
#
[dym_cx] oh, now it's up
#
[dym_cx] LJ had the best tool for discovery (within its limits) – interests.
#
[dym_cx] kylewm: 100% of innovation related to LJ should be attributed to bradfitz++
#
[dym_cx] bradfitz++
#
Loqi bradfitz has 2 karma
#
KevinMarks yes bradfitz pioneered lots
#
[dym_cx] whois bradfitz
#
[kylewm] yeah blows my mind how much stuff we depend on daily that brad came u pwith
#
[dym_cx] speaking of discovery – still no XFN (or analogue) with an easy access for less technical people
DePhraser joined the channel
#
KevinMarks how do you mean?
#
bear there has been talk about discovery in the past but without a use case from the users point of view to point at, it always ends up being a plumbing discussion
#
KartikPrabhu I do discovery just fine. People I follow share/like/repost things others have written. I read the other site and if I like it I follow that.
#
KartikPrabhu and so my reading network grows
#
KartikPrabhu I have never understood the use-case for "automatic discovery" or something
#
miklb that's like Twitter telling me who to follow based on someone I choose to follow.
#
KartikPrabhu I know what it is. But i have never had any use for that
#
miklb sorry, meant I'm not a fan of it either
#
KartikPrabhu oh i see, yeah I agree
#
KevinMarks bradfitz called it "the ex-girlfriend finder"
#
KartikPrabhu Twitter does help me find others to follow, but not by its "suggestions" but by reposts/favs by people I do follow already. So same as a /reader
#
bear I have found that twitters suggested follow does pop up people I would not have seen - at least on my twitter account that is for industry stuff
#
bear as my Ops follow graph is probably very similar to others in the ops realm
#
miklb I could see that in a more defined account.
#
KevinMarks http://classicprogrammerpaintings.tumblr.com/
shiflett, wolftune, j12t_, awolf, KartikPrabhu, strugee, finchd and gRegorLove joined the channel
#
aaronpk good morning!
#
aaronpk interesting, xpath does not appear to have the problem I thought it did
#
aaronpk added tests to php-mf2 to confirm anyway
bjoern and strugee joined the channel
#
aaronpk never parse html with regex
j12t joined the channel
#
aaronpk i mean i always knew that, but thought i could cheat anyway
KartikPrabhu joined the channel
#
gRegorLove PHP parser PR: https://github.com/indieweb/php-mf2/pull/93 Could use a double-check on my xpath query, but appears to be working right.
#
aaronpk that's going to match only an <a> tag that has a rel and href attribute
#
aaronpk is that correct?
#
aaronpk should probably add a test case that also has a class attribute
#
gRegorLove Correct
#
gRegorLove Oh yeah, I had one but forgot to put it in.
#
gRegorLove Need to handle those rel="bookmark somethingelse" cases too.
miklb and loic_m_ joined the channel
#
gRegorLove PR updated. And now I sleep.
#
aaronpk gRegorLove++
#
Loqi gRegorLove has 60 karma
#
aaronpk boom. all 19 webmention tests passing now.
#
aaronpk i had to walk down that dependency chain quite a bit for that
friedcell, miklb, quails, M-Kodo, glennjones, petermolnar, M-RyanRix, John_Duh and [kevinmarks] joined the channel
#
[kevinmarks] hm, if I ping the wrong endpoint do you delete the one to the right endpoint?
#
[kevinmarks] (my code sends to all endpoints found at the moment)
M-RyanRix1 joined the channel
#
aaronpk Yes I do
#
aaronpk you can technically sneak in a pass if you send to the right endpoint last
#
aaronpk I don't know how to catch that really
#
KevinMarks hah, so if I reverse sort the endpoints…
#
KevinMarks you could explicitly say it pinged the wrong one
#
aaronpk It does in the http response
#
KevinMarks hm, I'm using a set, so order may not be preserved
#
KevinMarks hm, test 20 is tricky as using mf2py to find rels doesn't distinguish link from a
#
KevinMarks OK, I think that fixed it
#
aaronpk #20 doesn't matter the difference between <link> and <a>
#
aaronpk it's testing to make sure you don't consider the lack of an href attribute to be the same as an empty string
#
KevinMarks hm. It seems to find both, but apparenly sends to the right one
#
aaronpk hm i should move it so that the wrong one is last
#
aaronpk that way it deletes it if you send to the wrong one second
#
aaronpk updates
#
aaronpk d
#
KevinMarks so using a set is a bad idea
#
aaronpk yes
tantek and [kevinmarks] joined the channel
#
[kevinmarks] ok, list used now
j12t joined the channel
#
tantek first to post successful comments on all 21! https://webmention.rocks/test/21 :D
#
Loqi [Webmention Rocks!] Test #21
halorgium, M-Kodo, tvn and tantek joined the channel
#
petermolnar long time no see, indieweb
#
Loqi petermolnar: GWG left you a message on 4/11 at 8:18am: I want to talk about Webmention Again http://indiewebcamp.com/irc/2016-04-11/line/1460387927651
#
petermolnar !tell GWG sorry, I was on holiday for a while, hit me with the questions any time
#
Loqi Ok, I'll tell them that when I see them next
#
j4y_funabashi hey petermolnar, nice holiday?
#
petermolnar yes, thank you, although that's #indiechat topic :)
#
j4y_funabashi ah ok was not aware that existed :D
frzn, Pierre-O, M-Kodo, strugee, hs0ucy and mlncn joined the channel
#
GWG petermolnar, I have been working on the Webmentions plugin.
#
Loqi GWG: petermolnar left you a message 1 hour ago: sorry, I was on holiday for a while, hit me with the questions any time http://indiewebcamp.com/irc/2016-04-15/line/1460721749839
#
GWG And since you had written a webmention plugin, I wanted your input
j12t, loic_m_, John_Duh, Jammy_Stuff and mlncn joined the channel
#
petermolnar I a
#
petermolnar I haven't read your commits but I saw the changes
#
GWG One of them involved adjusting the webmention handler so there could more easily be multiple handlers.
#
GWG I renamed the built-in one the synchronous handler.
bkardell, tvn and singpolyma joined the channel
#
sknebel aaronpk: idea for another webmention.rocks test case: webmention endpoint on a subdomain with non-ascii characters
shiflett joined the channel
#
aaronpk interesting
#
aaronpk i thought domain names could only have ascii characters?
#
aaronpk utf8 in URLs could work tho
#
aaronpk tho i think those need to be url encoded in order to be valid too right?
#
sknebel http://www.bücher.de is a valid domain (alkthough they put a redirect behind it)
#
sknebel there is an encoding ("punycode") to ascii that happens before DNS lookup
#
aaronpk when I type that in my browser it goes to www.xn--bcher-kva.de
#
sknebel exactly
#
sknebel but you can put the form I posted into your HTML
#
aaronpk curl http://www.bücher.de/ says "Could not resolve host: www.bücher.de"
#
sknebel and your browser knows that it has to convert that
#
aaronpk ah, that makes sense
#
aaronpk hm yeah then an HTML parser should be handling that
#
aaronpk actually what is the right place for that transformation to happen? should a microformats parser return the ascii version? or should the HTTP client be doing that?
#
sknebel that's a good question
#
sknebel probably the client
#
sknebel a microformat parser at least for some fields already outputs UTF-8
#
aaronpk yeah that's a good test to add :)
#
sknebel and it makes for nicer display if the URL (if it is displayed somewhere) is shown in the native script
#
voxpelli If the parser then the URL will never be printed correctly
#
voxpelli So yeah: the client
#
aaronpk yeah i think it's a http client responsibility
#
aaronpk since that's how the browser handles it
#
aaronpk i should throw in an emoji one while i'm at it
#
aaronpk http://👍.webmention.rocks
#
sknebel yeah
#
Loqi definitely
#
sknebel (for source and target URLs the same thing might become relevant, e.g. if people compare a punycode-string to a unicode representation and don't realize two URLs are identical)
#
sknebel display is interesting because as you observed, browsers sometimes show the punycode version, to protect against phishing attacks where I make a URL that looks like paypal.com but has a strange identical-looking characters somewhere
#
aaronpk but isn't this the responsibility of an http client?
#
aaronpk like if i'm using the curl library, it should be able to take an input URL of http://👍.webmention.rocks and know how to make that http request
#
sknebel yes
#
sknebel was only meant as extra-info why your browser showed you something strange
#
aaronpk i'm hesitant to add too many tests that are testing http-level stuff
#
aaronpk right now most of the tests are actually testing HTTP rel parsing, but there isn't a good test suite out there for that right now, and it's pretty important to get it right for Webmention, so it's relevant
awolf and j12t joined the channel
#
sknebel I think the emoji-subdomain endpoint would be enough for now -> tests if the HTML parser understands it and if the HTTP client can POST something there in one go
#
sknebel + many people probably don't even know that you can do that kind of URL now
#
aaronpk yeah, i'm going to hold off for now
#
aaronpk i'm in the middle of working on tests for updating posts, which is more relevant to the spec
#
sknebel fair enough
#
sknebel want me to add a github issue so the idea doesn't get lost?
Zegnat joined the channel
#
aaronpk i've got it in my notes
#
sknebel ok
#
sknebel bye, gotta run
mattronix joined the channel
#
aaronpk tschüss
myfreeweb, KevinMarks_, jonnybarnes, anarchivist, M-RyanRix, uranther, voxpelli, gmacon, ramsey, davbo and JasonO- joined the channel
#
GWG I wonder if anything I did to the Webmentions plugin would cause it to fail more tests
#
aaronpk now you can test it!
MylesBraithwaite, tbrb, bear, hs0ucy, daf, nelson_, Pierre-O, tommorris, DePhraser, bergie, kronda, anarchivist and Phae joined the channel
#
GWG aaronpk: I think I may wait till pfefferle gives his Thumbs Up or Down to the other changes I made.
jden joined the channel
#
aaronpk you should definitely feel free to throw stuff at webmention.rocks whenever you want, it's made to be a useful debugging tool
#
aaronpk you can always delete your posts from there if you want ;-)
#
Loqi +1
ramsey joined the channel
#
GWG aa
#
GWG aaronpk: Does it work in the other direction?
Garbee joined the channel
#
aaronpk other direction?
#
GWG Can it send me a test webmention? You are testing a site's ability as a sender of webmentions, what about receiving?
#
aaronpk ah yeah. I will need to add that, but it's more complicated on my end so I am doing that after I finish this half
#
GWG How will you protect against abuse?
#
aaronpk require people to indieauth login as their domain, and only allow them to send to targets on that domain
#
aaronpk in the mean time, this is a great tool for testing receiving webmentions https://github.com/voxpelli/node-webmention-testpinger
gRegorLove joined the channel
#
GWG aaronpk: That works.
#
aaronpk I actually used that to test receiving mentions for webmention.rocks
KartikPrabhu joined the channel
#
aaronpk helped me catch a few xss things i missed :)
tantek_ joined the channel
#
voxpelli aaronpk: nice!
#
GWG I really should try voxpelli's thing, but I haven't.
#
voxpelli GWG: I think pfefferle used it quite a bit, he did a reversed one as well
#
GWG voxpelli: I like the idea of it as a hosted service though, ala webmention.rocks
#
voxpelli GWG: as it pings an endpoint I made it deliberately local to avoid it being a cause of spam
Phae, KartikPrabhu, jden and tommorris joined the channel
#
gRegorLove tantek: Can you perform merges on microformats/tests repo? https://github.com/microformats/tests/pull/47
#
kylewm i believe KevinMarks can if not
bergie and ramsey joined the channel
#
GWG voxpelli: I just am procrastinating as I'd have to reconfigure my firewall.
#
voxpelli :)
#
aaronpk There was a version of some of those a test running online but I couldn't log in when I tried last week
daf, KartikPrabhu, j12t, anarchivist, loic_m__, tvn_ and tantek joined the channel
#
KevinMarks_ "the absurd architectural diagram of a GUI on top of an API on top of a GUI on top of a miasma." https://18f.gsa.gov/2014/09/08/the-encasement-strategy-on-legacy-systems-and-the/
#
KevinMarks_ sounds familiar
davbo_, awolf, bjoern__, nelson_, KartikPrabhu1, uranther, tantek, gRegorLove, cleverdevil, gRegorLove_, loic_m__, M-RyanRix, kronda, ramsey, Phae, bkardell, KevinMarks, j12t, bergie, tommorris, mlncn, jden and anarchivist joined the channel
#
bear cool - ronkyuu handles the new webmention.rocks tests (with a bugfix for #15)
davbo_, tommorris, Pierre-O, frzn and Vendan joined the channel
#
kylewm so php strip_tags is weird... strip_tags('a << b') just returns 'a '
#
kylewm bear++
#
Loqi bear has 121 karma
#
gRegorLove kylewm: Yeah, it doesn't handle "incorrect" HTML well
j12t_, gRegorLove_, nelson_, John_Duh, mlncn, bergie, ramsey and uranther joined the channel
#
Loqi Just generated the first draft of this week's newsletter! http://indiewebcamp.com/this-week/2016-04-15.html I'll generate it again in an hour and it will be sent out at 2pm.
M-Kodo joined the channel
#
M-RyanRix http://notes.whatthefuck.computer/ Arcology works!
#
M-RyanRix I need to re-add a rel-me but
#
M-RyanRix well, I guess there is a rel-me on the bottom
jden, catsup and [dym_cx] joined the channel
#
[dym_cx] "arcology" sounds like a baby trying to say "archaeology"
#
[dym_cx] "arcology" sounds like a baby trying to say "archaeology" Jump Mark as read
#
[dym_cx] "arcology" sounds like a baby trying to say "archaeology" Jump Mark as read
#
[dym_cx] test
gRegorLove and gRegorLove_ joined the channel
#
[dym_cx] wtf with slack»irc bridge again
gRegorLove_ joined the channel
#
M-RyanRix heh
#
M-RyanRix [dym_cx]: An Arcology is a self-sustaining community theorized by one of my favorite architects/artists https://en.wikipedia.org/wiki/Arcology
frzn joined the channel
#
Loqi Just generated this week's newsletter! You still have a few minutes to make changes, and I'll re-generate it 10 minutes before it gets sent out at 2pm. http://indiewebcamp.com/this-week/2016-04-15.html
KartikPrabhu, wolftune, lukasros, rMdes, frzn, hs0ucy, snarfed, link0, jonnybarnes and Phyks joined the channel
#
aaronpk kylewm: i wouldn't rely on strip_tags for anything. there's a good sanitization library here: http://htmlpurifier.org/
#
GWG Is this just the lull before the IWC?
#
aaronpk that's what I use in XRay and webmention.rocks
#
aaronpk GWG: this is the aaron quickly finishes the validator before IWC
#
GWG aaronpk: I'm looking forward to hearing people talk Indieweb.
#
aaronpk me too!
#
GWG You are more likely not to sleep through it
#
aaronpk depends on how late i stay up :P
tantek joined the channel
#
GWG aaronpk: Good point
wolftune joined the channel
#
kevinmarks.com edited /2016/Nuremberg/Guest_List (+349) "/* Remote Participants */" (view diff)
quails joined the channel
#
aaronpk testing the tester
#
KevinMarks hm, did loqi stop calling twitter search?
#
Loqi grins profusely
#
aaronpk streaming isn't working but i don't know why
#
aaronpk Loqi will find them with the polling
#
Loqi grins profusely
#
aaronpk thought it was working, but i might be hitting twitter rate limits or something
#
aaronpk their rate limiting errors on the streaming api tend to be a bit obtuse
#
KevinMarks I wonder if their otages yesterday knocked things over
#
KevinMarks s/otag/outag/
#
Loqi KevinMarks meant to say: I wonder if their outages yesterday knocked things over
#
KevinMarks might need to restart anything calling the streams?