#indiewebcamp 2013-10-26

2013-10-26 UTC
bnvk joined the channel
#
JonathanNeal tantek: http://www.crunchbase.com/company/manymoon
#
JonathanNeal manymoon === do
#
tantek my goodness their actual email announcement that shaners linked to includes the literal phrase "incredible journey" (!!!)
#
tantek per: http://ourincrediblejourney.tumblr.com/
#
tantek (!!!) https://do.com/done
#
tantek I want to know how much they spent on buying the domain name do.com
#
JonathanNeal Looks like do.com used to belong to ddo.com
#
JonathanNeal http://web.archive.org/web/20060104043143/http://www.do.com/
josephboyle joined the channel
#
tantek aaronpk - 48 line http_rels() php function ready for testing. I've been trying it on aaronparecki.com in case you've been seeing a strange useragent.
#
aaronpk sweet!
#
tantek apparently Link: header syntax can be kind a nuts - e.g.
#
tantek Link: <http://example.org/query?a=b,c>
#
tantek rel="d e", <http://example.org/>
#
tantek rel=f
#
tantek (found here: https://groups.google.com/forum/#!topic/hal-discuss/iXrmGnkmBcw )
#
aaronpk yeah :(
#
tantek so my code handles that :)
#
aaronpk wow. did you find some good test data?
#
tantek yeah that Link: above ^^^ :)
#
tantek and your site
#
tantek I figure that's pretty good ;)
#
aaronpk hah
#
aaronpk is this the real spec? http://www.w3.org/Protocols/9707-link-header.html
#
tantek now should I post this function as a gist or separate file?
#
aaronpk I wouldn't mind bundling it up as its own project, adding a test suite, and publishing to composer
#
aaronpk here is my current test suite for parsing link headers https://github.com/indieweb/mention-client-php/blob/master/tests/LinkHeaderParserTest.php
#
aaronpk and here's a relatively simple component bundled up how I would bundle this one up https://github.com/indieweb/date-formatter-php
#
tantek reading that spec now and seeing if I missed anything obv
#
tantek e.g. "Relationship values are case-insensitive"
#
JonathanNeal THAT'S RIGHT
#
tantek aaronpk, hmm, I'm ignoring all non-rel attribute on the Link: header
mikeal joined the channel
#
aaronpk where do they go in the mf2-parsed version?
#
tantek we ignore them there too except for special-casing alternate
#
tantek http://microformats.org/wiki/microformats2-parsing#rel_parse_examples
#
aaronpk maybe that's the answer then
#
tantek I designed that to satisfy actual use-cases, rather than full expression of what a <link> can do
#
tantek yeah - though I'm not going to bother with adding alternate-specific support
#
aaronpk well if you want, I can wrap this in a test suite real quick
#
aaronpk is it all a single function?
#
tantek yes
#
tantek quite tight
#
aaronpk I can just namespace it under IndieWeb without making it into a class
#
tantek yes that would be great. in reading that spec I need to add just a bit more defensiveness.
scor joined the channel
#
tantek aaronpk - right now I'm doing the curl in the function, I wonder is it better to split it into two functions, one that just extracts the rels from a headers string.
#
aaronpk ah yes, that would be ideal
#
aaronpk likely some people may want to fetch the headers some other way and still use the code to parse the rel values
tpinto joined the channel
#
tantek I'm going to postpone making the "extract rels from a headers string" function into CASSIS for now - but that shouldn't be too hard
#
tantek since it's just string manipulation
#
tantek then Node folks could use it too
#
aaronpk the function returns an array, yes?
#
tantek yes
#
tantek right now, just a php array, with the rel values as indices, which then are arrays of URLs themselves
#
aaronpk great
#
aaronpk did benwerd drop the Link from his headers?
#
tantek it was there earlier today
#
tantek also I think this is the more recent spec for HTTP Link: http://tools.ietf.org/html/rfc5988
#
tantek except they made the mistake of putting a registry in there (which HTML has promptly ignored and is officially using the microformats existing-rel-values registry)
#
aaronpk new rule: spec authors should be required to give a complete list of examples that can be used as tests
#
tantek heh - I don't think we pass that rule ;)
dybskiy joined the channel
#
aaronpk this one was particularly helpful: http://tools.ietf.org/html/rfc3986#section-5.4
#
aaronpk I know, we should fix that!
#
aaronpk I was able to turn that bit of the RFC into a test suite very easily: https://github.com/indieweb/php-mf2/blob/master/tests/Mf2/URLTest.php#L245
#
aaronpk simple test suite added, all tests failing ;) https://github.com/indieweb/link-rel-parser-php/blob/master/tests/BasicTest.php
#
tantek well alright then. ok, running some test code of my own with more examples from that spec
#
aaronpk feel free to drop your code in here: https://github.com/indieweb/link-rel-parser-php/blob/master/src/IndieWeb/link_rel_parser.php#L6
#
tantek should we worry about multiline header fields?
#
aaronpk have you encountered any in the wild yet?
#
tantek only in the spec
#
tantek specs
#
aaronpk oh wait you mean a header field with a newline as one of the characters? that seems weird
#
tantek newline then space is allowed as a way to break long header fields apparently
#
aaronpk wow
#
tantek so I'm doing this now: preg_replace('/\r\n[ \t]+/',' ',$h)
#
aaronpk I suppose that's what they had in mind
#
tantek correction: preg_replace("/(\r\n|\r|\n)[ \t]+/", " ", $h)
#
tantek sheesh
#
aaronpk heh
#
aaronpk http doesn't specity \r\n vs \n?
#
tantek well http does, but I'm coding http_rels just a bit defensively for ease of testing purposes
#
aaronpk ah
#
tantek so without the curl code, http_rels is 38 lines
dybskiy joined the channel
#
tantek screw commandline - I'm just going edit and copy/paste the code in
#
aaronpk hehe
#
aaronpk you should try Gitbox, it's the best git gui i've found
#
aaronpk that's what I use 98% of the time
#
Loqi agreed.
#
tantek ok saved - take a look
#
tantek (and welcome to my classic C coding style ;) )
#
tantek do the tests pass?
#
aaronpk weird, not returning anything
mikeal joined the channel
#
tantek works for me without namespaces
#
tantek aaronpk, e.g. try http_rels('Link: <http://example.org/query?a=b,c>
#
tantek rel="d e", <http://example.org/>
#
tantek rel=f');
#
aaronpk works on that one
#
tantek I can give you more if you like :)
#
aaronpk oh also I was checking for an array with a 'rels' array
#
tantek e.g. this works too: head_http_rels("http://aaronparecki.com/");
#
tantek oh sorry - I simplified what http_rels returns
#
tantek it's head_http_rels that returns an array with 'status' and 'rels' array
#
aaronpk ah
#
aaronpk updates test cases
#
tantek I figure that demonstrates how to build it up into something bigger if you want
#
aaronpk ok now it's only failing on the multi-line ones
#
tantek damn, really?
#
tantek e.g.?
scor joined the channel
#
aaronpk https://github.com/indieweb/link-rel-parser-php/blob/master/tests/BasicTest.php#L37
#
tantek oh I know why - your PHP multiline string literal is not doing \n\r
#
tantek er, \r\n
#
aaronpk oh I thought you relaxed that?
#
tantek only for continuation lines
#
tantek I'll fix it for all
#
aaronpk ah
#
tantek done
#
aaronpk test pass!
#
aaronpk only one change required
#
tantek what was that?
#
aaronpk this: https://github.com/indieweb/link-rel-parser-php/commit/b0373f561a976fee7e9237e2833ae5d1fdedaa67#diff-99fe59200b3e8c25ab7522e07e5a30b5L36
#
tantek oh right - one fix to the tests ;)
#
aaronpk was throwing an error if the key doesn't exist. you must have a lower threshold of error reporting.
#
aaronpk one fix to the tests, one minor fix to the code
#
tantek technically the line-break flexibility is a change in the code to accommodate the not-quite-http-spec multiline test ;)
#
tantek as HTTP requires \r\n linebreaks
#
aaronpk oh yea but the array_key_exists thing
#
aaronpk just submitted to packagist! https://packagist.org/packages/indieweb/link-rel-parser
#
aaronpk ok gotta run!
#
tantek oh is array_key_exists better?
#
tantek aaronpk - did you notice how it parses out the X-Pingback as if it is a Link header too and returns it like a rel?
mikeal, dybskiy, josephboyle, michaelrhodes, shaners, scor, tpinto, acegiak and tantek joined the channel
#
tantek !tell aaronpk re: "must have a lower threshold of error reporting" - not sure - I was just testing it with the default php install on MacOS X 10.7
#
Loqi Ok, I'll tell them that when I see them next
tpinto joined the channel
#
christopheducamp.com edited /WordPress-fr (+3234) "[fr: french translation sync'd - checked welcome]" (view diff)
#
christopheducamp.com edited /WordPress (-6) "/* People using WordPress */ moved blog to other sites list" (view diff)
#
tantek tommorris - did you catch any of anildash's speaking at Mozfest on video?
#
tommorris tantek: no, just saw the tweets about it
#
tommorris tantek: I asked Mozilla if they are going to publish the video
andreypopp joined the channel
#
tantek yeah - saw your request. often this stuff is archived on air.mozilla.org
#
tantek tomorris I'm watching the live stream on air.mozilla.org and don't really understand what is going on
tpinto and barnabywalters joined the channel
#
tantek tommorris - I found an mp3 someone posted: http://s3.amazonaws.com/boos.audioboo.fm/attachments/6142082/anil-dash-anildash-at-mozfest.mp3?AWSAccessKeyId=1K04FNJW0TJXZ6BV2WG2&Expires=1382788042&Signature=Uk5mWwHHwF3fWjCp5wBMEsHRt10%3D&audio_clip_id=1686596
#
tantek it's a fragment
#
tantek the tweets about it seemed a bit high level
#
tantek basic message "open is not enough"
#
tantek "economic value of links"
#
tantek "like spam is coming" - um, it's already here: http://indiewebcamp.com/spam#Like_Spam
#
christopheducamp.com edited /webmention-fr (+12312) "[fr: synchro - french translation to be achieved and checked. still searching french help to test on Jekyll ?)]" (view diff)
#
@xtof_fr @glenux ? en langage clair ? Si tu as du temps, je ch. aide et/ou exemples ds la jungle de #webmention en français : http://indiewebcamp.com/webmention-fr#Impl.C3.A9mentations (twitter.com/_/status/394041607600611328)
LauraJ joined the channel
#
tantek tommoris - does Ferocity (especially the version you're running on your site) accept webmentions? or send? (or only send pingbacks?)
#
tantek.com edited /webmention (+501) "/* Discovery in PHP */ HTTP HEAD request and rel parsing library now implemented and shared on github! brief notes about how to use in this context" (view diff)
Acidnerd, glennjones, Acidnerd1 and xtof joined the channel
#
tantek hey xtof - what are you running on your main site now?
#
xtof bonjour tantek : currently studying jekyll on a local instance. Think I should be ready next monday.
#
xtof Will start at the root name on christopheducamp.com and will describe indiemark first steps learning in french...
#
tantek wonderful!
bnvk, Acidnerd, xtof, andreypopp, bnvk_, bnvk__, ozten, tpinto and glennjones joined the channel
#
@OpenITP RT @OTI: @cznweb @OpenITP @indiewebcamp @redecentralize Thanks! (twitter.com/_/status/394126204632309760)
#
@OpenITP RT @cznweb: #FF @OpenITP @OTI @IndieWebCamp @Redecentralize (twitter.com/_/status/394126230788009985)
xtof, LauraJ and singpolyma joined the channel
#
aaronpk I need to display "comments" under my blog posts now
#
Loqi aaronpk: tantek left you a message 7 hours, 22 minutes ago: re: "must have a lower threshold of error reporting" - not sure - I was just testing it with the default php install on MacOS X 10.7
#
aaronpk just had someone email me a link to their blog post they wrote because they "didn't find a trackback feature" on my site!
#
christopheducamp.com created /spam-fr (+9164) "[fr: french translation draft]" (view diff)
tpinto, glennjones, caseorganic, tilgovi, mikeal and barnabywalters joined the channel
#
barnabywalters web dataflow UI experimentation: http://waterpigs.co.uk/intertubes/
#
barnabywalters doesn’t do much yet but you can wire the boxes together
#
aaronpk barnabywalters: hello!
#
barnabywalters mafternoon aaronpk!
caseorganic, melvster, tpinto and andreypopp joined the channel
#
tobilehman.com edited /OpenID (+64) "/* Consuming Sites */ add cufp.org" (view diff)
#
tobilehman.com edited /OpenID (+7) "/* Consuming Sites */" (view diff)
bnvk, xtof, KevinMarks, LauraJ, Slyphoria, paulcp, caseorganic, tpinto, josephboyle and earplugs joined the channel
#
@kevinmarks @kwerb interoperability, composability antifragility http://epeus.blogspot.com/2008/07/open-source-and-social-cloud-computing.html http://epeus.blogspot.com/2012/12/the-antifragility-of-web.html #indieweb (twitter.com/_/status/394189057195909120)
#
@nchenga @pfefferle how do I add Webmention to my WP blog? http://chiperoni.ch/wordpress/b/jo (twitter.com/_/status/394192935735939074)
caseorganic, glennjones and paulcp joined the channel
#
@nchenga IndieAuth is part of the Indie Web movement to take back control of your online identity. http://chiperoni.ch/wordpress/2013/10/indieauth-is-part-of-the-indie-web-movement-to-take-back-control-of-your-online-identity/ (twitter.com/_/status/394202036016660480)
aaronpk_, tpinto, jalbertbowdenii, Jeena and tobiastom joined the channel
#
tobiastom hey.
#
barnabywalters greetings tobiastom
ttepasse joined the channel
#
tobiastom I have q question about indie auth.
#
tobiastom on the website you say that persona and SMS are supported providers. How does that work?
#
bret aaronpk: I wanted to ask you about how you do flat file storage for your site... thinking of building off of my existing content that works in jekyll, and I know you had similar roots for your site
#
bret tobiastom: for SMS, you list a phone number on your site, and it sends you a one time authentication token via sms to that number
#
bret tobiastom: for persona, you list an email on your site and lets indieauth if you own it or not
#
tobiastom bret: you could use something like http://getkirby.com/
#
bret tobiastom: the best way to test out those auth methods is to try it out yourself!
#
tobiastom will do, thanks.
#
bret kirby looks cool! i was looking to try and roll my own thing though in node/express
#
tobiastom there are more solutions like kirby, but I prefer that one.
#
bret yeah it looks nice
#
bret it costs money though, can you hack on it?
#
tobiastom yeah, you can. it's completely available on github: https://github.com/bastianallgeier/kirbycms
#
tobiastom the price is just a license thing which supports the continued development of it.
aaronpk_ joined the channel
#
bret ah is ee
#
bret see
#
aaronpk what who is this aaronpk_ imposter
#
aaronpk bret: my file storage looks very similar to jekyll's markdown files
#
bret aaronpk: do you have any notes on how you go from the flat files the live site?
#
bret maybe next time we meet and have time we can talk about it
tpinto and tantek joined the channel
#
tantek good afternoon #indiewebcamp!
#
barnabywalters afternoon tantek
#
tantek catches up on logs
#
tantek barnabywalters - what do you think of the link rel parser I posted (with aaronpk's help) ?
#
tantek https://github.com/indieweb/link-rel-parser-php/blob/master/src/IndieWeb/link_rel_parser.php
#
barnabywalters tantek: looks good
#
tantek thanks!
#
barnabywalters did you see sandeep’s one? he wrote a state machine instead of using regexes: https://github.com/phpish/link_header/blob/master/link_header.php
#
barnabywalters oh interesting normalisation of custom pingback headers
#
tantek yeah I did :)
KevinMarks joined the channel
#
tantek if you think of any other similar-to-rel HTTP header fields that are worth converting to rel equivalents, let meknow
#
tantek barnabywalters how does this work if $link_values is just a string? >>> $link_values = array($link_values)
#
tantek also - mine accepts a raw HTTP header block as one big string
#
barnabywalters looks like it normalises it into a list of one item instead of demanding a list
#
barnabywalters tan
#
barnabywalters tantek: yep, that’s really useful for PHP curl handling
#
tantek exactly - hence I also included the function to *do* a PHP curl HEAD request to a URL for the Link headers if that's all you want to do: head_http_rels()
#
barnabywalters so crazy that PHP doesn’t have string HTTP headers -> normalised array built in
#
tantek barnabywalters - I don't see how I'd use sandeepshetty's function to parse HTTP headers - looks like you have to do some preprocessing first
#
barnabywalters tantek: not advocating its use (haven’t used it myself), just making sure you knew it existed :)
#
tantek yeah I think I saw it because it's linked from the Libraries section: http://indiewebcamp.com/webmention#Libraries
#
tantek but it's not useful on its own and it was easier to write a *38 line* function that did *all of it*
#
tantek I like functions that fit on a single screenful on my MacBook Air 11"
#
tantek plus it handles x-pingback etc.
#
tantek doesn't understand why most code/functions takes *pages and pages*
#
KevinMarks java
#
tantek KevinMarks - do you mean "java" as a programming culture problem? As in, "java" programmers writing code in other languages write bloated code?
#
barnabywalters there are tradeoffs, as with anything. I find cassis extremely difficult to understand because of all the single letter variables
#
tantek barnabywalters - yeah - no argument there - the usability of cassis.js needs work
#
tantek which is why I decided to write this function as just PHP rather than cassis
#
tantek though the http_rels($h) function could easily be converted to cassis.js
#
KevinMarks yes, java trains people into insane verbosity. Also spawning needles entities
#
barnabywalters also, 10-level deep code folder structures
#
tantek so, I think selfdogfood kind of cures that
caseorganic joined the channel
#
tantek you'll never get anything done on your own site if you have to write pages and pages of code with 10-level deep code folder structures.
#
tantek no one in the indieweb community is building their own site/code with Java AFAIK.
#
KevinMarks good point
josephboyle joined the channel
#
KevinMarks perl is the opposite problem...
#
tantek I've been fairly convinced by Crockford's JSLint coding guidelines
#
tantek.com edited /webmention (+331) "add link_rel_parser to webmention endpoint discovery section in libraries" (view diff)
#
tantek barnabywalters - how do you iterate through the hyperlinks in your posts when doing webmentions? do you iterate through the HTML output of your system, or do you iterate through some sort of pre-output like markdown etc.?
#
barnabywalters IIRC I use the cassis URL regex to find all the URLs in my HTML
#
tantek wow - even if they're not hyperlinked then
#
tantek just plain text mentions like ttk.me
#
barnabywalters they will be by then, because I use cassis auto_link
#
barnabywalters lots of cassis in my note posting flow
#
tantek sure - but then you'll get double mentions right? for the URL in the href and for the URL in the plain text?
#
barnabywalters nope, I filter out duplicates
#
barnabywalters I should probably just do a DOM query for all elements with href
#
tantek also - it won't handle relative URls
#
Loqi yea!
#
tantek e.g. if you reply to yourself with a relative URL
#
barnabywalters usually I’m not that fussed about sending webmentions to relative URLs on my own site, but it’s a good point
#
tantek so you can comment on your own stuff
#
barnabywalters I try not to use too many relative URLs, having seen how awkward it is to implement spec-compliant software which consumes them correctly
#
tantek so true
#
barnabywalters woo! got typed connections working for web dataflow environment: http://waterpigs.co.uk/intertubes/
#
barnabywalters try dragging the output of the URL input to the inport of the output module
tpinto joined the channel
#
tantek whoa
#
barnabywalters so tomorrow I just need to build the backend
#
barnabywalters yahoo pipes v2
#
tantek heh
#
tantek yahoo pipes that you can run on your own domain
#
tantek indie pipes
#
barnabywalters anyone know who’s maintaining yahoo pipes?
#
barnabywalters sounds like most of the original creators moved on to other stuff/away from yahoo
#
tantek it's probably on autopilot
Acidnerd joined the channel
#
barnabywalters goodnight all
#
Loqi buenas noches