#indiewebcamp 2015-12-22

2015-12-22 UTC
tantek, KartikPrabhu and snarfed joined the channel
#
GWG
https://core.trac.wordpress.org/ticket/30783 - Maybe it will work out this time.
#
GWG
tantek: This is launching the next improvements to my Post Kinds enhancements.
KevinMarks and todrobbins joined the channel
#
tantek.com
edited /2016-01-01-commitments (+225) "/* Why 2016-01-01 */ Answer 2016 and 2015 separately"
(view diff)
snarfed, mlncn and lukebrooker joined the channel
#
kylewm
is mp-syndicate-to[]= the recommended way to pass syndication targets to a micropub endpoint now? cc: aaronpk
#
kylewm
i see woodwind is still using the unprefixed syndicate-to[]
Jihaisse joined the channel
#
kylewm
I think I had a hand in suggesting mp-* and now I'm regretting it :P
tantek joined the channel
#
KevinMarks
what does known want?
snarfed and johnstorey joined the channel
#
GWG
What is the state of current parsers, If there are both mf1 and mf2 on a page, how will most of the current ones respond if they aren't in the same positions in the page?
#
GWG
I need to solve this.
#
aaronpk
lol kylewm
#
aaronpk
it's *technically* not too late to change that since it's not super widely implemented yet
lukebrooker joined the channel
#
KevinMarks
the mf2 rules say to ignore mf1 special names if mf2 h- container is found
#
KevinMarks
not all parsers do this yet, though
#
aaronpk
that was a relatively recent addition IIRC
#
GWG
KevinMarks: The issue is that if I file the patch suggested by the core contributor, it would leave the classic in and add the mf2.
#
GWG
So, if the parsers don't support that...
#
aaronpk
IMO that is acceptable and we would fix the parsers to support that ASAP
#
GWG
I'd better write the patch
#
kylewm
KevinMarks: known technically uses mp-syndicate-to[] but syndication via micropub doesn't work at all yet
#
aaronpk
kylewm: I would rather iterate on this now before there are more implementations and before I write up the w3c version of micropub
#
kylewm
I'm not sure how to proceed with it aaronpk, just seems like a couple aesthetic judgment calls need to be made
#
kylewm
I know I really don't like the mp-* prefix in the JSON versions
#
kylewm
because in the JSON format, there is already a way to differentiate control parameters from properties
loic_m_ and [tantek] joined the channel
#
[tantek]
aaronpk "IMO that is acceptable and we would fix the parsers to support that ASAP" is a horrible way of approaching this. We already have bugs against the parsers to fix issues discovered against current Wordpress usage and the parsers are not all instantly fixed.
lukebrooker joined the channel
#
[tantek]
So no it is not acceptable. The unit tests suggestion is the way to go so that the support is not broken in the future.
#
[tantek]
GWG if you are going to keep mf1, just make sure that the overriding mf2 class names go on the same exact elements as the mf1 classes.
[aaronpk] joined the channel
#
[aaronpk]
kylewm: there isn't an mp prefix in the json version
#
[tantek]
Kevinmarks it is not "new". Much of that compat/interaction was defined at least a year ago.
#
[tantek]
Kevinmarks sorry that was not you that said new.
#
[tantek]
Aaronpk you said "recent addition" - it's not really. It's almost a year old. (If not 2?)
#
kylewm
aaronpk: the JSON form has mp-action here still, is that not current? https://indiewebcamp.com/Micropub#Update
#
[aaronpk]
I said relatively recent, as in there haven't been many more additions since then
#
[tantek]
goes back to sleep (lagging a bit EST)
#
[tantek]
There have been many new minor tweaks since then. See mf2-parsing-issues.
#
[aaronpk]
tantek: i would much prefer the parser I am using prioritizes mf2 markup since it's more descriptive, so I would think of it as a bug that mf1 markup is causing the mf2 parsed result to include other data
#
[tantek]
That's already specd that way but the parsers still have bugs.
#
[tantek]
This is my point. You said we would fix the parsers. My point is the parsers already have bugs and no they are not instantly fixed as you claim they would be.
#
[aaronpk]
kylewm: oh yeah you're right. It's that way because it makes the two formats analogous other than serialization. Still up for taking suggestions tho
#
[tantek]
anyway lots of specific issues on mf1 vs mf2 interactions on the mf2-parsing-issues page. If you have a specific use case / example you are worried about please check those specific issues and proposed / actual resolutions.
#
[tantek]
It's possible you'll find a new use case but we won't know until you can write it down as example markup.
#
kylewm
aaronpk: can you remember where we were writing those alternative microformats JSON serializations? cannot find them on the wiki at all...
lukebrooker, pauloppenheim, KevinMarks_, todrobbins, KevinMarks__, cweiske, j12t, KartikPrabhu, edsu, Leeky, glennjones, elima, M-kegan, M-Kodo, M-RyanRix, johnstorey and Pierre-O joined the channel
#
GWG
!tell tantek The problem with keeping the mf1 on the same elements as the mf2 is that the mf1 implementation in WordPress is incorrectly implemented, but hentry is used by a lot of themes as a styling element, so it would break half the themes out there.
#
Loqi
Ok, I'll tell him that when I see him next
j12t joined the channel
ttepasse and [jonnybarnes] joined the channel
#
[jonnybarnes]
aaronpk: <https://telegraph.p3k.io has an invalid cert (cert being sent is for other domains) fyi
j12t, modem, loic_m_, mlncn and frzn joined the channel
#
cweiske
loic_m, attack 1.2.3.4
#
cweiske
!loic_m attack 1.2.3.4
#
cweiske
I thought you were an IRC interface of http://sourceforge.net/projects/loic/
#
cweiske
low orbit ion cannon
nitot, singpolyma, sammachin and yakker joined the channel
j12t and [aaronpk] joined the channel
#
[aaronpk]
jonnybarnes: yeah I haven't actually launched anything there yet :-)
xregetic joined the channel
#
[aaronpk]
kylewm: is it on the brainstorming page?
#
[aaronpk]
Or was it in a gist?
elima, friedcell and tantek joined the channel
#
aaronpk
I think I need to write up the webmention status page thing as an extension
Pierre-O, KevinMarks_ and [shaners] joined the channel
#
[shaners]
aaronpk kylewm: is this what you're looking for?
wolftune joined the channel
#
aaronpk
ah yeah cool
#
kylewm.com
edited /micropub-edit-examples (+656) "add "B" that uses single-character control parameters"
(view diff)
j12t, snarfed, ttepasse, KartikPrabhu, Lancey and gRegorLove joined the channel
#
gRegorLove
Howdy, #indiewebcamp
ttepasse joined the channel
#
aaronpk
yep I really need a spec for webmention status
#
aaronpk
i'm implementing both ends of that right now and need some instructions to follow
snarfed and [kevinmarks] joined the channel
#
[kevinmarks]
you're implementing both ends and the editor of the spec? handy
ttepasse and loic_m_ joined the channel
#
kylewm
aaronpk++ status spec is really important
#
Loqi
aaronpk has 39 karma
ttepasse joined the channel
#
KartikPrabhu
what is salmention?
#
Loqi
Salmentions are a protocol extension to Webmention to propagate comments and other interactions upstream by sending a webmention from a response to the original post when the response itself receives a response (comment, like, etc.) https://indiewebcamp.com/salmention
ttepasse, squeakytoy, snarfed, Pierre-O1 and loic_m_ joined the channel
#
GWG
afternoon
snarfed, loic_m__, tantek and marcthiele joined the channel
#
marcthiele
Good evening from Krefeld, Germany
#
marcthiele
Got a question. I recently added webmention support to my blog o
#
marcthiele
at the beyond tellerrand website. Using Kirby and brid.gy.
#
marcthiele
Is there any way to get a post which is about a month old to be checked by brid.gy?
loic_m_ joined the channel
#
snarfed
marcthiele: when you sign up for bridgy, it looks at recent posts, but not arbitrarily far back
#
snarfed
it didn't find that post when you signed up?
#
snarfed
if not, there's no easy way to make bridgy process it manually, but you can send the webmentions yourself if you're a bit technical: https://brid.gy/about#source-urls
#
marcthiele
Thanks. Seen this. I thought there might be a way to trigger a specific post.
#
marcthiele
It found the post, but not all reactions to it.
#
snarfed
oh, in that case i might take a look
#
snarfed
which post?
#
marcthiele
Strange enough it found only Facebook likes, but nothing on Twitter. as you can see at the bottom.
loic_m joined the channel
#
snarfed
looking
#
snarfed
right, looks like you signed up twitter @btconf user on 12/14, about a month after that post, so that makes sense
#
snarfed
sorry for the inconvenience!
#
marcthiele
So it was too "old" already when I signed up?
#
marcthiele
This means I can only add it manually?
#
snarfed
right now, yes
#
snarfed
i'm not currently thinking about how to go farther back than it already does when you sign up, but feel free to file a feature request! https://github.com/snarfed/bridgy/issues
#
marcthiele
Either this or something to manually shoot an older url to brid.gy to be checked
loic_m and marcthiele_ joined the channel
#
KevinMarks
if you tweet a link to it, does bridgy decide to look again?
#
snarfed
KevinMarks: the new tweet, yes, but not the old tweet with the responses he wants
#
KevinMarks
ah, right
tantek joined the channel
#
marcthiele
Yes, KevinMarks, I tried this already.
marcthiele joined the channel
#
aaronpk
snarfed: do you think it'd be simple enough to make a field you could put in a link to an old post and bridgy would go check it?
#
snarfed
aaronpk: honestly, no. :/ it's not really designed for that mode
#
snarfed
i mean, it's obviously doable, but nontrivial
#
aaronpk
it wouldn't require extra backend infrastructure would it?
#
snarfed
i'm not sure
#
aaronpk
I don't really know how the queueing system works but I would imagine it might be as simple as putting a new job onto the queue the same way it does when it finds a new post
#
snarfed
aaronpk: yeah, i understand the idea. the problem is that most of the work happens *before* then, in creating the job to enqueue
marcthiele joined the channel
#
snarfed
mostly since that job is per response, not per post, since posts regularly get new responses
#
snarfed
if you want to look, it'd basically be generalizing the poll method here to handle a single silo post as well as the current "grab everything" mode: https://github.com/snarfed/bridgy/blob/master/tasks.py#L108
#
snarfed
aha. thanks for filing https://github.com/snarfed/bridgy/issues/579 marcthiele!
elima joined the channel
#
marcthiele
Yep. I think twittes API could be a problem as well though
#
snarfed
marcthiele: why?
#
marcthiele
If the post is too old, I mean. Not?
#
snarfed
no. they actually put a fair amount of work in a while back to make all tweets over history accessible, including to the api
#
snarfed
(twitter is also just one of five silos we support. definitely the most popular though!)
#
marcthiele
Ah ok. As this are not many posts I'd love to get the mentions for, I am currently thinking of doing it manually. But that is quite some work as well ;)
#
snarfed
marcthiele: you're definitely right, it is. sorry we don't have a better answer!
modem, marcthiele and wolftune joined the channel
#
aaronpk
huh, someone is trying to log in on indieauth.com as me
#
aaronpk
i just got an SMS verification code
#
snarfed
ominous
#
aaronpk
they tried twitter too
#
snarfed
i just realized i have no idea what to do if that happens
#
snarfed
change passwords?
#
snarfed
i know how to protect ahead of time, but not during
#
tantek
if what happens?
#
Loqi
tantek: GWG left you a message 10 hours, 6 minutes ago: The problem with keeping the mf1 on the same elements as the mf2 is that the mf1 implementation in WordPress is incorrectly implemented, but hentry is used by a lot of themes as a styling element, so it would break half the themes out there. http://indiewebcamp.com/irc/2015-12-22/line/1450785099723
#
tantek
GWG, sounds like this needs further discussion.
#
snarfed
tantek: what aaronpk just said (someone trying to log into sites as him)
#
aaronpk
there isn't really anything they can do
#
snarfed
well, change passwords
#
snarfed
unsophisticated but does help some
#
KevinMarks
talking of logins, anyone want to check my made up email a login link model for me?
#
KevinMarks
when someone clicks it, I increment a counter, append their email, the counter and random to a string, and md5 that
#
KevinMarks
then I convert from hex to newbase60 and store it in a db table
#
KevinMarks
then email them it
#
aaronpk
probably should avoid md5. at least use sha256 instead
#
bear
+1 to avoiding md5
#
KevinMarks
if they click the link, I check the string is in the db, log them in by setting a cookie, and delete the db entry
#
aaronpk
kylewm: do you think there's value in defining a list of error codes that describe webmention errors like "source doesn't have a link to target" vs "target URL doesn't exist"? or is that fine as just an optional human readable description
#
KevinMarks
I know md5 is frowned on for actual hashing, I am just trying to create something unique enough that it's not guessably sequential
#
aaronpk
KevinMarks: why not just a random sequence of 20 newbase60 chars?
#
bear
KevinMarks - I would use sha256 and also remove the incrementing counter and just create a uuid
#
KevinMarks
I don't want to just use random, as I'm not sure it's reliably random
#
ben_thatmustbeme
sorry KevinMarks just makde me think of this
#
bear
personally I would hash the email and a random UUID then convert that to newbase60
#
KevinMarks
sha256 is goign to make very long strings to email
#
bear
for your use case md5 is ok as the token is short lived and only used to identify the email click is valid
#
KevinMarks
md5 is 22 newbase60 chars, sha256 is 44
#
aaronpk
use emoji instead and you can get it down to like 10
#
KevinMarks
right, emailing clickabel emoji links is so going to work
#
aaronpk
what could possibly go wrong
[shaners] joined the channel
#
[shaners]
aaronpk: thanks for adding me to the indieweb github org
#
[shaners]
The ruby parser is now under the indieweb org
gRegorLove_ joined the channel
#
aaronpk
[shaners]: \o/
#
aaronpk
https://github.com/indieweb/microformats2 can I rename that to something that includes "ruby"?
#
[shaners]
I’ll do some repo cleanup things
#
[shaners]
Yeah. Like that. And Travis, etc.
#
aaronpk
k I'll let you take care of that :)
[aaronpk] joined the channel
#
[aaronpk]
wtf the links aren't coming through from the slack bridge
#
tantek
shaners++ thanks shaners for finding the rubymf2 parser a new home!
#
Loqi
shaners has 14 karma
#
bear
*cough* circlecic *cough*
#
bear
s/circlecic/circleci/
#
Loqi
bear meant to say: *cough* circleci *cough*
#
[shaners]
bear: why?
#
[shaners]
tantek: it’s long over due. I kept trying to get others to own this, but no one would. So, I’m gonna.
#
bear
shaners - i'm being silly - I work for CircleCI now ;)
#
tantek
what is CircleCI?
#
Loqi
It looks like we don't have a page for "CircleCI" yet. Would you like to create it? https://indiewebcamp.com/s/10AE
#
KevinMarks
hm, a uuid4 would be simpler
#
bear
KevinMarks - yea, a good uuid generator solves so many of the "now how/what do I need for *this* token..." items
#
[shaners]
bear: oooooh
#
[shaners]
are yall hiring?
#
bear
we will be in february
#
KevinMarks
newbase60.numtosxg(int(uuid.uuid4().hex,16))
#
bear
(wow - did I just mangle feb)
#
[shaners]
nm: I found the jobs page. (I’m looking)
#
bear
CircleCI is a SaaS company delivery Continuous Integration and Delivery to developers at https://circleci.com
#
loqi.me
created /CircleCI (+136) "prompted by tantek https://indiewebcamp.com/irc/2015-12-22/line/1450823465580 and dfn added by bear"
(view diff)
#
__number5__
bear: congrats. I like CircleCI
#
bear
thanks
#
bear
but truthfully, Travis or Circle -- just as long as devs are doing tests then we all win
#
tantek
what is Travis?
#
Loqi
It looks like we don't have a page for "Travis" yet. Would you like to create it? https://indiewebcamp.com/s/10AF
#
[shaners]
What is TravisCI?
#
Loqi
It looks like we don't have a page for "TravisCI" yet. Would you like to create it? https://indiewebcamp.com/s/10AG
#
@jgmac1106
Thanks to @fortyfoxes I now have #fedwiki up & running. Now its time to watch @holden tutorials and get ready to try this semester #indieweb
(twitter.com/_/status/679430494291943424)
#
[shaners]
TravisCI is a SaaS company delivery continuous integration and delivery to developers at <https://travis-ci.org
#
loqi.me
created /TravisCI (+138) "prompted by [shaners] https://indiewebcamp.com/irc/2015-12-22/line/1450823778304 and dfn added by [shaners]"
(view diff)
#
[shaners]
Travis is [[TravisCI]]
#
loqi.me
created /Travis (+21) "prompted by tantek https://indiewebcamp.com/irc/2015-12-22/line/1450823740956 and dfn added by [shaners]"
(view diff)
#
bear.im
edited /CircleCI (+31) "add see-also"
(view diff)
#
bear.im
edited /TravisCI (+60) "add see-also"
(view diff)
#
[shaners]
Is there a go to license for libraries in the indieweb org?
#
aaronpk
I usually use Apache or MIT, but no I dont think there's much consistency
ttepasse joined the channel
#
tantek
CC0 would be preferred over MIT. CC0 is basically an internationally vetted version of MIT.
#
aaronpk
what no that doesn't sound right at all
#
[shaners]
I prefer CC0 myself
#
tantek
MIT still has "branding" recognition as a code license but literally there is no reason to use it instead of CC0 at this point.
#
aaronpk
CC0 actually attempts to have the owner give up the copyright
#
aaronpk
in the MIT license, the authors retain copyright of the work
#
tantek
Right, it's the international way to do public domain, which is effectively the desired impact of MIT
#
bear
the biggest reason I don't like CC0 over MIT is that CC0 allows for copying without noting origin of the work
#
tantek
bear - yes that's deliberate so as to not burden every single minor use of code
#
bear
and that is the reason I don't like it - even minor uses of code should credit the source material
#
tantek
if you can cite where there has been harm from "without noting origin" and/or where MIT was successfully used to sue someone to force them to note the "origin", I'll consider it. Otherwise it's an empty promise/threat.
#
bear
that's a straw man arguement - of course people will always steal/borrow code without credit and the vast majority of it is done without knowledge
#
bear
but having it listed as MIT gives a polite nudge to the person using it to at least show credit someplace
#
bear
it establishes the rules and boundaries for good social sharing of code
#
[shaners]
bear: i don’t care about due props given.
#
bear
[shaners] - then CC0 will work for you
#
tantek
bear - no it's not a strawman argument - it's pointing out that the "without noting origin of the work" is a chickenlittling/strawman argument
#
[shaners]
HTML, HTTP, etc were public domain and look at how they spread and succeeded
#
tantek
bear, a culture of attribution and citation goes beyond any particular license
#
tantek
and that's the thing to reinforce (that culture), not attempting to do so via legalese/licensese
#
bear
[shaners] http and html are not public domain - they are copyright the The Internet Society
#
tantek
original versions were public domain IIRC per Weaving The Web
#
tantek
only later versions got tangled up in people's paranoia about licensing and stuff
#
tantek
fear driven nonsense
#
bear
i'm not in a good place to be arguing about this so I'll bow out and go for a walk
#
tantek
has spent far too much time arguing this stuff with so many people at W3C etc.
#
kylewm
aaronpk: I can't think of a good reason to define webmention error codes. vouch-required is the only one i can imagine caring about and that is already handled
#
bear
my only point to make is that calling what I feel to be a sound opinion on why MIT is preferred over public domain "fear driven nonsense" is not productive or helpful and just makes me not want to help
#
tantek
bear - MIT is obsolete and people are still clinging to it purely out of habit / reputation, that's my point. Lots of licensing work has happened since the MIT license got put out there, and the reasons for MIT have all been rationally disposed of, especially for *anything* international (which is nearly everything today)
KartikPrabhu and snarfed joined the channel
#
[shaners]
i’m using CC0
elima joined the channel
#
kylewm.com
edited /FreeMyOAuth (-1) "update flickr icon"
(view diff)
Pierre-O1, Gold and e-lima joined the channel