2016-04-11 UTC
# 00:08 aaronpk i'm surprised there isn't already a mechanism for that in Flask or whatevr you're using # 00:09 bear it's part of the library I created for webmentions, ronkyuu # 00:09 bear not from what i'm seeing in my debug output # 00:10 bear I just added some output for headers and they are all camel case (for this request) # 00:11 bear bah - it must be something i'm doing wrong # 00:11 bear the type of the headers item is requests.structures.CaseInsentiveDict # 00:12 bear but I'm sure as heck not getting a string when I do `r.headers['link']` # 00:13 aaronpk wow python has a class called CaseInsensitiveDict? that's kind of handy # 00:15 bear I don't think it has a language core item, but the Requests library implements one # 00:18 bear i'm using the latest ronkyuu for this, right now i'm hacking in some debug output # 00:18 aaronpk i wonder if i can get it to return a link header like "liNK" # 00:22 bear k, then it's an error in another layer ... # 00:22 bear I love these types of things - they cause me to add more tests # 00:40 miklb aaronpk I totally just cribbed your flexbox use on those test pages. Thanks! :-) # 00:42 miklb now all I need to do is style the webmentions in this thing, write some documentation, and I can share it to get feedback. # 00:47 bear question about the link header for test/1 aaronpk -- shouldn't it be rel="webmention" instead of rel=webmention ? # 00:49 bear I was wondering - my rel finding code was failing # 00:50 bear just like here: link-value = "<" URI-Reference ">" *( ";" link-param ) # 00:50 aaronpk "and MUST be quoted if they contain a semicolon (";")" # 00:51 aaronpk does that mean they don't need to be quoted if there is no semicolon in the rel value? tantek joined the channel
# 00:53 bear ah - I see it now - wow, that sure does read ambiguiously # 00:55 bear | <"> relation-type *( 1*SP relation-type ) <"> # 00:56 aaronpk so rel=webmention *and* rel="webmention" are valid? # 00:56 bear so it could be: rel=webmention OR rel="webmention" # 00:57 bear now to make my regex for rel= discovery work with optional quotes # 00:57 tantek subject lines to not use in comments that get propagated to email: "Your recovery phone number changed", "Your recovery email address changed" # 00:58 tantek every comment you make on a Known instance gets propagated to email notification by default, therefore ... # 00:59 tantek if you start your comment with text like the above, it's likely the recipient will see it in the subject of their email # 00:59 tantek I think WordPress similarly has comment notification by email # 01:00 tantek and since most of these "native" comment systems let you claim to be whoever you want, e.g. claim to be "Google" # 01:03 GWG tantek: Yes. You can do anything with a plugin. mlncn joined the channel
# 01:29 KartikPrabhu aaronpk: might be good to rename the tests according to what they test or something # 01:29 aaronpk yeah i was struggling with coming up for a concise name for each # 01:30 bear oops - sorry aaronpk , I forgot to give feedback on the main page, I got too into fixing ronkyuu to pass test #1 ! # 01:30 aaronpk Test #1: "Webmention endpoint discovery with unquoted HTTP Link header" while descriptive is a little excessive # 01:30 bear don't rename the tests IMO, but just add explainer test # 01:31 tantek aaronpk - have to head to dinner but I'll see if I can come up with pithy shorthand names for these. well done! # 01:31 bear unless you group them: discovery_01 - webmention endpoint discovery with unquoated HTTP ... # 01:31 aaronpk i'm planning on grouping them in different sections on the home page. right now these are all basically various discovery tests # 01:32 tantek came up with an entire shorthand naming system for CSS test suite tests back in day, you may be able to search/dig for it and find it. # 01:32 aaronpk eventually i'm going to add webmention *sending* tests, but that's a whole different story # 01:32 aaronpk where webmention.rocks will send you test mentions # 01:32 kylewm tantek: could you explain your comment about "your password has changed" in an email notification? I'm not following at all, is there a bug? # 01:32 tantek not saying you should, just saying if you were curious about web history trivia etc. # 01:33 tantek kylewm: point was that's a bit of a "scare phrase" (having just seen it in an actual email) and thus I could see it being an attack vector of sorts (people sending/making a comment that is designed to look like a password reset / recovery email message) # 01:34 tantek that's as far as I got - I don't know if there is an actual vulnerability there, just potentially another phishing attack surface # 01:34 tantek for any system that automatically puts 'web comments' into email messages # 01:35 tantek ooh aaronpk I got one I got one (actually make that two :) ) # 01:36 kylewm That kind of thing is really nasty with linked in that sets the person's name as the sender # 01:36 tantek rel="webmention x-some-other-rel-value-whoknows" # 01:36 tantek rel="x-some-other-rel-value-whoknows webmention" # 01:36 tantek rel="x-some-other-before webmention x-some-other-after" # 01:37 tantek yeah - Google Reader failed feed discovery with rel value parsing this way back in the day, I filed a bug, they never fixed it # 01:37 tantek you might say I'm still annoyed they didn't bother with such a simple bug fix # 01:37 tantek so let's put this in the validator / test suite # 01:37 tantek if your serverside consuming code is doing lazy rel value parsing, too bad # 01:37 tantek we could be really evil and test invalid stuff too like: # 01:38 tantek see if anybody errantly tries to send a webmention to those endpoints # 01:38 aaronpk yeah I need to figure out how to show those negative cases # 01:38 tantek you can combine those and put them BEFORE the valid ones # 01:38 tantek so dumb parsers / regexers will get tripped up on the incorrect rel value # 01:39 tantek instead of looking for the correct value later # 01:39 aaronpk there's another one which is when a page advertises an endpoint using all three methods (HTTP Link, html link, a link), the sender is supposed to send to only the first one, so I need to show if it ends up sending to the wrong endpoint # 01:39 tantek heck put a false rel value / endpoint like that both before and after # 01:39 tantek HTTP Link, then whichever a/link comes first in the HTML # 01:40 tantek like if there is a a rel=webmention BEFORE a link rel=webmention, then the a should win # 01:40 aaronpk "the first HTTP Link header takes precedence, followed by the first <link> element, and finally the first <a> element." # 01:40 aaronpk i don't know *why* it's worded that way, but it's been that way for a long long time # 01:41 tantek I wonder if anyone actually did that literally # 01:42 tantek dangit well I'm looking for the first with rel~=webmention # 01:42 aaronpk it does seem reasonable to look for the first a or link with rel=webmention # 01:43 tantek it takes more code to look once for link, then again for a # 01:43 bear that is what ronkyuu does, check for the presence of a link header, if not look for an href # 01:44 bear and the href scan will return <link> before <a href 's # 01:44 aaronpk i should add one that has img src="*" rel="webmention" # 01:44 tantek bear really? just because you expect <link> to be in head? # 01:45 bear dude, i'm just now getting used to HTML let alone HTML5 wolftune joined the channel
# 01:56 bear odd - test/2 returned a status code of 200 but it's not showing # 02:05 tantek aaronpk: please consider pure HTML source order for rel discovery rather than preferring LINK over A. I'm fairly certain that other (previous?) rel discovery mechanisms did that. # 02:05 tantek could file an issue to track it - got to run to dinner though # 02:14 aaronpk bear: yay! also hooray for finding fun edge cases in my markup # 02:15 bear i'm sending a test webmention for test/2 and the result is 200 - but i'm not seeing it # 02:15 aaronpk this is where my plan to show a log on the site of all the webmentions it receives (including invalid ones) would be helpful # 02:15 aaronpk and just to double check, you discovered the endpoint for test/2 separately? # 02:15 bear yea, this is why i'm adding a debug=True to ronkyuu :) # 02:16 bear my code does the discovery,send for each url found # 02:16 bear (at least that is what it should be doing, let me check) # 02:17 aaronpk (you shouldn't have gotten a 200 if you sent to the wrong endpoint) # 02:17 bear yep - discoverEndpoint() and sendWebmention() are for each url that qualifies # 02:34 bear aaronpk - is there a repo I can file issues against for webmention.rocks ? # 02:35 bear ah - I see it at the bottom of the main page now snarfed joined the channel
# 02:52 aaronpk bear: it returned HTML in response to your POST request? # 02:54 aaronpk it's advertising the webmention endpoint as http, but it should be https # 02:55 aaronpk yeah when I switched on https, I made it redirect all http requests # 02:55 aaronpk so you're actually getting back an HTTP 301 for the post to the webmention endpoint # 02:55 bear hmm, my send webmention checks for a 405 with location header but not for a 301 # 02:56 bear ohhh - it does! but it's a human refactoring error that is causing the error # 02:58 aaronpk according to HTTP you're not supposde to POST to the location returned by the 301 tho # 03:08 bear oh I see my confusion, the result became a GET but I check only for a 200 - I don't confirm that the *POST* generated the 200 # 03:53 bear KartikPrabhu - i'm pushing 0.3.6 to pypi now j12t joined the channel
mblaney and gRegorLove joined the channel
# 05:00 gRegorLove Regarding wm discovery, my code doesn't differentiate between 'a rel' and 'link rel' currently, so I think it would just use whichever one it found first in the HTML. # 05:22 miklb thanks. still need to document how someone unfamiliar with GitHub can use it, but needed to make it "real" heh # 05:35 bear aaronpk++ just totally stole the webmention.rocks html and headers to mock out tests for ronkyuu # 05:37 bear wrote a bit of code to run thru each link and store the html and headers # 05:37 bear later after we help you fill out the tests I'll have to write something that parses the main page for tests :) # 05:39 bear i'll have to give that a scan later - i'm approaching max bear time for a sunday :/ tantek joined the channel
# 05:44 tantek good to know gRegorLove - perhaps we can convince aaronpk :) # 05:45 gRegorLove tantek: if I'm understanding you correctly, you said the 'a rel' should take precedence over 'link rel' right? # 05:45 tantek no I'm saying the first a OR link rel should take precence # 05:45 Loqi tantek meant to say: no I'm saying the first a OR link rel should take precedence # 05:46 tantek that document order of something linky with rel webmention takes precedence, not specific tag name # 05:46 bear I'm wondering if the html body should ever override the links in header tho # 05:47 tantek bear, unlikely if we're following first in source order rules jedahan joined the channel
# 05:47 bear true, then I'm going to not change my implementation just now as I first check head for links and then proceed to html # 05:50 bear ok, I think I confused myself just now # 05:50 tantek I also just check http header first, and if no link rel webmention there, then check html # 05:50 bear my order of processing is this: HTTP headers, html <head>, html <body> # 05:53 bear k, I think ronkyuu is good - when it's scanning html it is looking for anything with rel=[webmention,...] # 06:00 bear cool - heading to bed - with v0.3.6 of ronkyuu all 8 tests are passing - going to try and keep up with the pace of webmention.rocks # 06:00 tantek good news, looks like I no longer need to twice-webmention to Known for comments to show up there # 06:01 tantek going to see if Falcon can match that shortly ;) cweiske joined the channel
# 06:30 cweiske tweet about all incoming and outgoing webmentions worldwide # 06:30 aaronpk maybe congratulate people when they get their first comment on the site # 06:30 aaronpk cweiske: that sounds like a job for @allthewebmentions nitot_ and jihaisse joined the channel
[kevinmarks] joined the channel
# 07:16 [kevinmarks] Next, register trackback.sucks and make it display all trackbacks. It will rapidly live up to its name. friedcell joined the channel
# 07:30 miklb voxpelli thanks. as I said earlier, merciless criticism encouraged loic_m, pfefferle, nitot, friedcell, adactio and Pierre-O joined the channel
# 09:02 pfefferle acegiak I will add the changes also to my fork... perhaps we should discuss how to handle/implement the salmention in a good way, to unify both versions. # 09:25 acegiak Pfefferle, yes! Except I'm super busy all the time at the moment. Maybe over email? It just !tell me your thoughts and I'll get back to you when I can? # 09:25 acegiak Really I'm all about the hack that gets things working to start with. I'm not so much the person that implements things properly for the long term j12t, sanduhrs and nitot joined the channel
# 10:00 pfefferle acegiak perhaps we start it as a independent plugin and merge it, when it is finished... # 10:09 acegiak Pfefferle would require a lot of interhooking with the semantic linkbacks plugin j12t joined the channel
Nissyen, friedcell, almereyda, Pierre-O, squeakytoy and j12t joined the channel
# 11:42 voxpelli pfefferle: will you come to any of the upcoming german IWC:s? # 11:43 voxpelli saw you mentioning Salmention and Salmentions feels like a great group exercise :) loic_m joined the channel
j12t, hs0ucy and friedcell joined the channel
# 13:14 GWG pfefferle, I saw your note. I realized the case sensitive mistake after I did it. I thought I fixed them all. snarfed and prtksxna joined the channel
snarfed1, singpolyma, hs0ucy, snarfed, j12t, Pierre-O and shiflett joined the channel
# 15:17 GWG petermolnar: The next time you are around, would like to talk. # 15:18 GWG !tell petermolnar I want to talk about Webmention Again # 15:18 Loqi Ok, I'll tell them that when I see them next tantek, hs0ucy and KartikPrabhu joined the channel
# 15:57 miklb would it be prudent to have an h-card on every page? my project has one on home page, but wondering if I should add one to footer on non-home pages. # 15:57 singpolyma miklb: I do on non-home pages, but that doesn't necessarily meant you need to ;) # 15:58 tantek miklb - "h-card on every page" is a plumbing-first way of looking at it # 15:58 tantek you have a page, from a user perspective that page was likely created on a certain date (dtpublished), it likely has a name (title, p-name), and content, thus it makes sense to mark it up with h-entry # 15:59 tantek the page likely has an author as well (p-author), who likely has their own URL (even if another page on the same site), thus it makes sense to hyperlink the author name from a UX perspective and then once you have that, it makes sense to mark *that* up with h-card # 15:59 miklb tantek I have all of that marked up. So that would be sufficient? # 16:00 tantek so yes (in general) to original question "an h-card on every page?" but not for the reasoning you may have been thinking # 16:01 miklb so use the relevant data to build the h-card # 16:01 tantek right! and the data is relevant for UX reasons, not for arbitrary plumbing reasons # 16:02 miklb no, I'm only interested in indieweb best practices # 16:03 tantek right, and our best practices come from UX/design first # 16:06 tantek miklb it would be as the "p-author h-card" inside the h-entry for the page # 16:07 tantek do you already have something like <html class="h-entry"> or something like that/ # 16:08 tantek see above where I said start from the page down, as an h-entry # 16:08 tantek it doesn't make sense to jump straight into hiding an h-card deep in a footer somewhere nitot joined the channel
# 16:09 miklb yes, the article is marked up as h-entry. So that is an implied h-card? # 16:09 tantek if you have h-entry markup already, then you should have a p-author property for that # 16:09 tantek and that's your hook for the h-card for the page, as the p-author of the page # 16:10 tantek the h-entry markup I quoted above was just one way of doing it - whatever makes sense for your CMS / publishing tool # 16:11 GWG So, if you have information about yourself on every page... Isn't this the minimal hcard linking to full scenario that is coming up? # 16:12 miklb jekyll is just HTML so I can markup however I want. Currently I'm adding the author info in the article footer, but if I understand correctly, I should consider moving that above the e-content # 16:12 tantek GWG see above where I said "the page likely has an author as well (p-author), who likely has their own URL (even if another page on the same site)" # 16:12 KartikPrabhu miklb: above/below does not matter. It should be a child of the h-entry element # 16:13 GWG tantek, I was just pointing out that there was a discussion of it in the past # 16:14 tantek GWG, can you link to that discussion of it in the past? # 16:17 KartikPrabhu i don't think there is any consuming code that does minimal vs full h-card thing, so right now it is ignorable # 16:18 GWG KartikPrabhu, my point was only to add to the conversation that the issue had been discussed prior. emmak joined the channel
# 16:18 GWG tantek, it was classified as one. # 16:19 voxpelli miklb: I think my "theme" has a h-card for every entry # 16:19 aaronpk miklb: it doesn't even need to be a child of the h-entry, since /authorship *should* find it if it's outside too # 16:19 GWG The page should probably have that moved to brainstorm # 16:20 aaronpk my site has my h-card in the footer of every post for example. tho i'm going to add my h-card inside the post for a few kinds of posts soon. # 16:21 tantek aaronpk: hmm that wasn't the point of authorship # 16:21 miklb aaronpk that's honestly why I brought up the question, was seeing how you are doing it # 16:21 tantek authorship aims to capture existing common cases, not to express best practices for publishing snarfed joined the channel
# 16:21 kylewm it would be an interesting challenge if the minimal h-card included only name and url (not photo) ... # 16:22 aaronpk tantek: hm, but it works in reverse, since once you've defined how to find author info, someone can look at that as what to expect consumers to be able to find # 16:22 KartikPrabhu aaronpk: people not familiar with /authorship. The easier thing to understand is that the h-card is child of the h-entry. If you know /authorship inside-out you could do many other things # 16:23 miklb just to be clear, I'm only questioning for my own edification and to be able to share with other Jekyll users (or anyone else looking for markup examples) # 16:23 KartikPrabhu kylewm: I don't know why it is a challenge. my webmention endpoint just ignores it if there is no photo # 16:23 tantek aaronpk: no it does not. just because browsers support tables for layout does not mean you should seriously consider tables for layout as an option # 16:23 tantek that's the wrong place to go for publishing advice / guidelines / best practices # 16:24 tantek looks like I'm seeing all the webmention endpoints except test 7, hmmm... # 16:25 tantek kinda surprised by that actually, thought the code I wrote (which is open sourced) for link headers was case insensitive # 16:26 aaronpk is on the train to Seattle, with a less than great wifi connection, and is taking this opportunity to build an authorization endpoint into p3k so that he doesn't have to rely on indieauth.com when signing in to local apps squeakytoy2 joined the channel
# 16:27 GWG aaronpk, if you used WordPress I'd have everything I wanted built already. # 16:29 aaronpk i mean I do have an offline version of indieauth.com too, but it's annoying to have to run that app when i just want to sign in to my dev copy of quill # 16:29 snarfed aaronpk: an alternative is to have apps skip auth (or however you want) when running in local/dev/etc mode # 16:29 tantek And hey RFC7230 "header field consists of a case-insensitive field name" there we go # 16:30 GWG I want an authorization endpoint that uses the WordPress login. # 16:30 Loqi gives GWG an authorization endpoint that uses the WordPress login # 16:30 GWG Loqi, I will look forward to your pull request # 16:31 aaronpk i'd still need a way to create a session with Quill to tell it where to post to yakker joined the channel
# 16:31 snarfed eh, depends on how many separate services you want to be able to talk to each other locally # 16:31 voxpelli miklb: it's a bit odd btw that the rel-author and p-author h-card has different URL:s? # 16:32 tantek aaronpk: have you considered having Micropub officially supersede RFC 5023 "The Atom Publishing Protocol"? Pretty sure Micropub handles everything better # 16:32 voxpelli +1 on Snarfed, my local apps, like my WebMention endpoint skips auth locally # 16:32 voxpelli although a local auth endpoint sounds like a good thing as well (as do a WP-based one) # 16:32 snarfed aaronpk: yup, i definitely haven't tried following it to its extreme like you'd need to. probably still doable though. but the local indieauth alternative is also good and maybe easier for you # 16:33 aaronpk tantek: interesting. I should read through that to see if there's anything in there not yet covered by Micropub. I suspect it has a few more things, especially around collections/lists. # 16:35 GWG snarfed, I promise I'll get back to Micropub eventually. # 16:36 GWG snarfed, I am currently burying pfefferle in refactoring webmentions for new specification parameters. # 16:37 miklb voxpelli I think I need to revisit that part after this conversation # 16:38 voxpelli I realize that embarrassingly enough I have forgotten rel-author (which I have been a fairly vocal advocate for and built some tools that consumes :P) # 16:39 GWG snarfed, I think the WordPress community, with all the webmention articles of late, needs to see those plugins being updated. # 16:40 kylewm aaronpk: I always have to bypass auth for unit tests, wouldn't imagine it's too much extra work beyond that # 16:40 voxpelli miklb: btw, to keep the schema focused and minimal as well, perhaps remove the Schema.org stuff? so that one can get a good example of a pure IndieWeb-theme? # 16:40 snarfed GWG: maybe? i dunno. i'm a bit skeptical of the trending idea that recent updates are such an important measure of a project's health, esp for small projects. # 16:41 snarfed i mean, i get that it's *a* measure, but especially for smaller things that work and are well tested (which wp-micropub isn't, granted), it doesn't make as much sense to condemn projects that haven't changed much in a year, or three # 16:41 miklb voxpelli I debated that myself. When I was going through the html 5 spec, it was using schema markup. That was a bit self serving as this will eventually be a starting point for my own redesign and wanted to use it. # 16:41 snarfed (if they're *abandoned*, then maybe, sure, but that's a bit different) # 16:42 GWG snarfed, there have been a lot of changes in webmentions since the plugin was written # 16:43 aaronpk kylewm: it's more like how do I then get a session on Quill that posts to my micropub endpoint # 16:43 aaronpk it just sounds like more work and more extra code paths to maintain # 16:44 GWG For example, the current plugin ignores a 404 on a source link. # 16:44 voxpelli miklb: can maybe be a bit confusing to see mixed metadata formats – also not sure the Schema.org parts add much, I haven't really felt an urge to add them to my theme # 16:45 GWG So a lot of little things that will set up for features like Vouch, Salmentions, delete, etc. # 16:45 miklb voxpelli I could certainly remove them now and archive the markup for my own uses later. # 16:46 aaronpk GWG: I would prioritize supporting updates and deletes over extensions like vouch, if that helps at all # 16:46 GWG I didn't write an asynchronous handler, but I changed the code so someone could drop one in more easily. # 16:46 GWG aaronpk, that is part of the refactoring. # 16:47 GWG I'm not adding features till the new code is merged in to support it. # 16:47 GWG Basically, the hooks I can add new functionality into without disrupting existing functionality tantek joined the channel
# 16:50 miklb voxpelli I'm impatient by nature so wanted to get this out there, plus, this is my my first true attempt at a solo project intended for others vs just throwing some of my own code up on GitHub # 16:50 GWG The same applies to Semantic Linkbacks. I'm trying to fix update functionality there. # 16:51 miklb a wise man once told me "release early, release often" :-) # 16:51 voxpelli miklb: +1, release early, release often is a good thing :) will try to give it a better look and some more feedback later on + look at whether our themes can converge somewhat # 16:51 GWG I broke it trying to get the webmention enhanced before the notification went out about it. snarfed joined the channel
# 16:52 miklb voxpelli I waited on the micropub stuff since I'm not 100% sure of how it all connects. Figured once I could test it personally, would make more sense to me to add. # 16:52 miklb voxpelli also, you might be happy to know that GitHub is working https # 16:52 tantek miklb not sure how the schema nonsense made it into that HTML5 draft - I'll look into it from a W3C perspective, as schema is a Google thing (with some MS/Y!/Yandex input) - not really a "standard" in terms of openly developed as we typically expect # 16:53 voxpelli miklb: sounds like a good idea, the micropub endpoint should probably be my main focus for at least the first iwc, I should probably try to make it a simple service to sign up to # 16:53 aaronpk i've gotten better about writing more complete tests since I wrote that library # 16:54 tantek aaronpk - these are the two functions I wrote http_rels and head_http_rels snarfed joined the channel
# 16:54 tantek fixed two bugs in http_rels, and now working on head_http_rels which has to handle relative hrefs # 16:55 miklb tantek to be clear, I think the example in the draft uses a slight variation from schema.org. This was simply my attempt at incorporating. Perhaps better documentation there would help novices like me # 16:56 tantek miklb indeed - it goes to show the power of examples, even if they are informative / "non-normative" # 16:57 tantek I think KevinMarks was working on adding an h-entry example right next to it (before it preferably) so not too many people would be led astray by the schema stuff # 16:57 voxpelli tantek: couldn't find a page on the IWC wiki regarding schema.org, surely there must be such a page, right? # 16:57 tantek voxpelli: no because it has no relevance for the indieweb # 16:58 voxpelli and perhaps an explanation why mf2 is the better choice # 16:58 tantek voxpelli: if you think it would help - e.g. like if get more folks asking same kinds of questions as miklb, then go for it! # 16:59 tantek KartikPrabhu has tried both and can help with the comparison aspects # 16:59 voxpelli feels he is so not the right person to create such a page though :P # 17:02 miklb doh. didn't see the prior reference. I have a friend visiting Iceland and that was one of her concerns. # 17:03 Loqi [Ryan Barrett] Field trip inside PrÃhnúkagÃgur volcano near ReykjavÃk, Iceland. (And yes, the base camp has a fax number.) # 17:03 Loqi [Ryan Barrett] Field trip inside PrÃhnúkagÃgur volcano near ReykjavÃk, Iceland. (And yes, the base camp has a fax number.) # 17:06 kylewm imho, wiki pages that are mostly critical of some other technology tend to just get us in trouble # 17:07 tantek is really glad he already has a CASSIS function for making a relative URL absolute begriffs joined the channel
# 17:07 tantek (I think that's likely why we had not created a page yet) # 17:08 tantek (This page intentionally left blank because we couldn't think of anything nice to say and/or this thing is not relevant to the indieweb) # 17:08 bear unless the tech has items that people will run into afoul of, a negative page should be avoided # 17:10 Loqi bear meant to say: unless the tech has items that people have run into afoul of, a negative page should be avoided # 17:11 tantek right, wait for someone to canary themselves into an anti-pattern and then document it # 17:12 tantek that way when theoreticians get defensive about their favorite thing being anti-patterned, we can cite the real world example # 17:13 bear ^^ that shortens a lot of arguments very quickly # 17:14 tantek joins bear in the webmention.rocks 8 for 8 club :D # 17:15 kylewm tantek: would you plz delete the volcano article :) # 17:16 kylewm the pages that have "IndieWeb Examples" with no entries are loud-and-clear :) snarfed1 joined the channel
# 17:18 tantek gah - really don't like merging pull requests that do nothing but change spaces and comment *formatting* (not actual comments) # 17:18 aaronpk as the maintainer of a project, you have full authority to reject those :P # 17:19 tantek aaronpk: on IndieWeb projects I think nearly everyone here has authority to accept / reject # 17:19 aaronpk what's worse though is when someone submits a PR with both whitespace and functionality change, cause then it makesi t super hard to figure out what the actual change is # 17:19 snarfed1 aaronpk: ?w=1 # 17:20 snarfed1 (agreed, regardless) # 17:20 tantek but if just a whitespace change without subsequent code change it's like really? # 17:22 tantek * @return array $rels rel values as indices to arrays of URLs # 17:22 tantek does anyone recognize this comment syntax as something worthy of maintaining? does it help auto-generate something? # 17:22 kylewm tantek: lol which thing were you looking in my direction for? # 17:23 tantek ok people, you got like 60 seconds to defend this @keyword nonsense in PHP function header comments before it gets reverted # 17:23 kylewm oh, funny. i thought that was a great pull request # 17:23 tantek (if it's not obvious enough for one of the PHP folks here to quickly cite why it matters, then it's not obvious enough to keep) # 17:24 aaronpk some editors use that to provide autocomplete and inline docs about functions # 17:24 tantek kylewm: I kind of see that kind of GitHub formatting-only commit as GitHub activity inflation # 17:25 bear if the language has a standard lint setup, and the project is open source -- IMO it is worthwhile to document that the project follows (or not) the standard # 17:26 tantek I did a lot of work to get cassis.js to pass jslint wolftune joined the channel
# 17:26 bear then you help defuse the format-wars by being able to point to the lint spec # 17:26 tantek ok phpdoc looks like a reasonable enough thing to support (since someone went to the trouble) begriffs_ joined the channel
# 17:28 bear warns about speaking to the why/reason-behind someone submitted a PR and focus on the reason it's not a good reason to merge # 17:29 tantek bear, given how many more steps it takes to "save" something in github rather than just "Save", I'm suspicious of such activity # 17:30 tantek in my experience, the more you code, the more you get allergic to "ceremony" # 17:30 tantek (which is what all the command line extra mumbojumbo is) # 17:30 kylewm there is a guy on wikipedia that just changes every instance of "comprised of" to "composed of", people are irritated by surprising things. mix tabs and spaces is not that weird of a peeve # 17:31 miklb voxpelli ah, I see now what you meant about different URLs. I'm still grokking how GitHub pages handles urls, i.e. site.github.url vs site.baseurl/site.url # 17:31 bear my point is this - sometimes a PR like that is nothing but "oh, hey, easy github juice" but often it's someone who is trying to become an opensource dev and thinks "hey, this helped me grok the project, let me share" # 17:32 aaronpk (except when that one company ran a promotion where they'd mail you a shirt if you made N pull requests to open source projects in a month) # 17:32 bear some folks *live* by that activity graph # 17:33 kylewm oh man, as someone looking for a job, i think github points are a kind of big deal # 17:33 aaronpk huh, i've never paid attention to the activity graph of a person # 17:33 bear yea, some hiring folks now take a look at a persons github profile to gauge "10x dev worthyness" # 17:34 aaronpk i mean i've looked at candidates' github profiles, but i actually go look at the contents of projects # 17:34 tantek I'd do the opposite, and random sample their commits to see "substantive" they were # 17:34 tantek again, to see if they're more about the code or more about the ceremony # 17:34 tantek lots of little github edits means you have way too much time on your hands and not enough impatience to put up with github's ceremony # 17:34 bear I didn't mean to distract the conversation, was just saying that reasons for a PR are not always obvious # 17:35 tantek lol no thanks. makes it harder to find substantive commits # 17:36 tantek for the same reason you shouldn't upgrade for every little software upgrade # 17:36 kylewm and why some projects ask you to squash your commits, but tha'ts more ceremony # 17:36 bear now your talking about the process of change/pull-requests # 17:36 bear PRs and commit messages have become the indirect way of measuring a dev's velocity by some (IMO inferior) managers # 17:41 tantek ok, changes merged except for my own debugging code snarfed joined the channel
# 17:45 tantek huh, I think this exercise actually helped me track down what was causing odd tab vs space mixing in my editor. # 17:52 tantek ok, as I was saying, now that I've tested the fixes in production, I'm going to commit them to github :P # 17:58 tantek but yes, with those changes, Falcon passes 8 of 8 @webmentionrocks tests # 17:59 miklb voxpelli btw it seems themes as gems are going to be a real thing in jekyll (at least from a gh-pages perspective) # 18:00 voxpelli miklb: oh? I've avoided gems so far because they don't work with GH Pages # 18:01 voxpelli really wishes Jekyll would get an alternative non-ruby implementation, would benefit static sites a lot # 18:01 miklb voxpelli I pinged benbalter on a question about code review of a jekyll gem/plugin (archives) and stated themes as gems & https were current priorities KartikPrabhu joined the channel
# 18:10 miklb voxpelli what exactly do you mean by "non-ruby" # 18:11 aaronpk oh another benefit of PRs on github is that Github will run unit tests on the PR branch so you can see if your tests pass before merging # 18:11 voxpelli miklb: like libsass was for Sass, it "solved" Sass' problem of many using gem extensions when it wasn't needed # 18:16 voxpelli aaronpk: and if one has linting, ensure that the linting passes :) # 18:17 voxpelli Really nice when people can get that feedback upfront for themselves before one gets time to comment on their PR # 18:18 tantek voxpelli: interesting, I wonder if I can set that up for cassis.js and jslint somehow # 18:18 tantek hopes the test case that aaronpk added passed with the code changes >:) # 18:19 voxpelli tantek: should be doable, I can have a quick look at setting up such a Travis CI flow for your repo # 18:26 tantek aaronpk: one of the communities I'm in is discusisng pros/cons of FB private group vs. having a community Slack # 18:26 voxpelli tantek: with what settings? I get a couple of failures # 18:27 tantek hmm, it may have regressed since I last tested it (a few years ago?) # 18:27 aaronpk interesting. the style of FB groups and Slack is very different # 18:27 tantek aaronpk: the use-case is similar though, that's the point # 18:27 tantek the tool specifics are less relevant than the desire for a "private forum" at a high lvel # 18:27 Loqi tantek meant to say: the tool specifics are less relevant than the desire for a "private forum" at a high level # 18:28 aaronpk yes, I just think that which one is more appropriate depends on the kind of interaction the group wants # 18:29 tantek hence the discussion of pros and cons - there's no consensus on that # 18:32 miklb if one were to posse to slack, in theory could posse to this channel, yes? # 18:35 voxpelli tantek: then you install eg. jslint as an additional dev dependency and add it to the test command of your package.json, then it will run on each commit and PR when you activate the repo on the Travis site # 18:41 tantek will come back when there is a more obvious path forward # 18:41 bear point me to the repo and I will help setup linting per PR/merge # 18:42 singpolyma tantek: backfeed sounds harder, since it's hard to tell what is in reply to a post in a chatroom context # 18:43 bear tantek - do you use TravisCI or CircleCI currently? # 18:43 voxpelli bear: I guess I should try out CircleCI as well some day :) Adding Travis everywhere now as I have that for 10 or so repos already :) # 18:43 bear k, I know that brid.gy uses CircleCi already, trying to gauge which (obviously biased to CircleCI as I work there) # 18:45 tantek hey that sounds good - better tech support ;) # 18:46 bear it's also good for both to be used so we have examples to point at # 18:46 bear i've forked cassis, i'll set it up using circleci for now and submit a PR # 18:49 bear CircleCI has the same policy basically as travis - everyone can get a single build container and open source projects can apply for free 4 build containers # 18:50 bear our messaging has been ... well... weak in that area # 18:51 bear a lot of that is changing with OS X support now, different Ubuntu containers and soon another support that I can't talk about until tomorrow Pierre-O joined the channel
# 18:57 tantek snarfed++ for passing @webmentionrocks 8 for 8! # 18:59 tantek or rather, let's see if webmentionsrock supports webmention updates ... hs0ucy joined the channel
gRegorLove, KartikPrabhu and snarfed joined the channel
# 19:09 snarfed pfefferle++ for the wordpress webmention passing all of the webmentionsrock tests # 19:09 tantek still, pretty happy to have tested updates - though as a validator shouldn't it somehow give me validation for sending a successful update?!? # 19:10 GWG snarfed, now I don't have to do that. Good. # 19:11 aaronpk thinks he should quickly add more tests since people seem to be passing these very quickly # 19:11 bear hmm, what is the tool used for php linting? I see references to a PHP CodeSniffer but not being a PHP dev... # 19:12 bear ( I have tests running for both php and js for cassis ) # 19:13 tantek aaronpk definitely tests for updates and deletes would be awesome snarfed joined the channel
# 19:22 GWG aaronpk, as in... I didn't make it hard enough? # 19:25 GWG tantek, I could use such tests shortly. squeakytoy2, squeakytoy and bear joined the channel
# 19:33 bear my server is offline so I lost my irc connections :/ bear_web1 joined the channel
# 19:36 bear_web1 goes to look for slack info # 19:43 bear_web voxpelli - phpcs requires an xml file to configure what it checks for - is there a default one someplace? # 19:45 bear_web I have no idea why i'm online as bear_web* also # 19:45 bear_web ah, so I can just add phpcs to the composer json # 19:45 voxpelli bear_web: one can see a list by running "phpcs -i" and the most modern standard of them are the "PSR2" # 19:46 voxpelli bear_web: yep, installing it through composer is simplest, doing a pear dance or something is having PHP showing it's no that very nice side of things :P # 19:47 bear_web yea, i've lived the pain of pear before # 19:47 tantek going to try a Bridgy Publish Facebook photo tagging test # 19:47 tantek I wonder if supplying http/https URLs makes any difference, or if supplying www.facebook.com vs facebook.com makes any difference # 19:48 voxpelli bear_web: "phpcs --standard=PSR2 ." should then do the trick, but I'm not sure whether that's actually what tantek is using :P # 19:49 tantek bridgypublish++ (already did it in #indiechat so this may have no effect) # 19:49 tantek I didn't know you could double-karma across channels like that # 19:50 aaronpk hmm how should a webmention receiver handle http vs https URLs when the contents are the same? # 19:51 voxpelli aaronpk: I think I normalize them so that the two are treated as the same URL # 19:51 tantek aaronpk, better file it as an issue, that's a pretty big general question in W3C circles # 19:52 tantek might even CC @timbl because I think he has a specific opinion on that - a site serving over both http and https (not redirecting) # 19:52 tantek I'd say de-duping is fine, and prefer the https version # 19:52 voxpelli aaronpk: I never print a normalized URL anywhere though, it's just for deduplication # 19:53 aaronpk yeah the image cache thing is what prevented me from doing that on my old site # 19:53 tantek aaronpk: I mean if the URL *domain+path*s are the same, then prefer the https # 19:53 voxpelli in my case I really needed it for the target of a webmention as else webmentions would be fragmented when sent to a site like gRegorLove's # 19:54 aaronpk tantek: source URL only? before getting to the step of parsing mf2? # 19:55 tantek consider the attack scenario, someone somehow fakes up an http domain to pretend to be you, but can't mimic your https # 19:55 voxpelli I think eg. some people sent a webmention to "kevinmarks.com" rather than "www.kevinmarks.com", and that didn't do well before normalization # 19:55 bear_web tantek - what command do you run for jslint for cassis? # 19:55 bear_web :stuck_out_tongue: # 19:55 voxpelli bear_web: tantek: I couldn't find any that worked, I think that doing a new one based on eslint would be the easier approach # 19:56 tantek I kind of liked the jslint Crockford branding but any js/eslint would be an improvement over none # 19:56 aaronpk now i kind of wish it would keep them around forever # 19:57 Loqi [gRegor Morrill] Publish on your Andesite, Syndicate Elsewhere # 19:57 tantek aaronpk - you could archive the first 48hours ;) # 19:59 tantek or first 48 hours after you announce it on the telcon tomorrow # 20:00 bear_web wow - phpcs --standard=PSR2 is *slow* # 20:00 tantek perhaps the webmentionrocks copy should suggest commenting with what publishing tool and/or any HTTP/HTML link header/rel libraries people are using # 20:03 bear_web grabs some food while phpcs chews on things # 20:11 bear_web goes to lunch (will add eslint to the mix after lunch) # 20:24 aaronpk okay now i'm at least storing the response data on disk for archival purposes bear joined the channel
# 20:31 aaronpk oh dear. embedly thinks bear's photo is the main photo for the page tantek and [shaners] joined the channel
# 20:52 aaronpk hmm. why use a different CoC than the IndieWebCamp one? # 20:52 [shaners] I didn’t think about that. I used this one because lots of projects are moving to it. # 20:53 tantek I believe that's called "cargo cult" behavior # 20:53 aaronpk probably better to link to the indiewebcamp one instead of duplicating the copy also # 20:53 singpolyma cargo cult CoC: copies file into repo, doesn't read it, doesn't abide by it ;) # 20:54 Loqi IndieWebCamp code of conduct tl;dr: Be respectful of other people, respectfully ask people to stop if you are bothered, and if you can't resolve an issue contact staff. If you are being a problem, it will be apparent and you'll be asked to leave https://indiewebcamp.com/code-of-conduct # 20:54 tantek akin to rel="license" but for a code of conduct instead of copyright? # 20:56 [shaners] It’s not cargo culting. (I forgot about the IWC wiki CoC.) I’ve read the Contributor Covenant. I agree with it and think its’ good. # 20:56 [shaners] Me using the Contributor Covenant is akin to using a well tested open source library that has lots of eyeballs on it. # 20:57 singpolyma [shaners]: I didn't necessarily mean you specifically. just the hordes adding a CoC file blindly to their repo # 20:57 tantek shaners, the reasoning you gave "because lots of projects are" is cargo cult reasoning # 20:57 Loqi tantek meant to say: shaners, the reasoning you initially gave "because lots of projects are" is cargo cult reasoning # 20:57 tantek "Convenant" sounds either too religiousy or too Halo-alieny IMO # 20:57 Loqi tantek meant to say: "Covenant" sounds either too religiousy or too Halo-alieny IMO # 20:59 [shaners] "lots of projects are moving to it” => list all of the reasons why this one is good. good enough for lots of projects big and small by people I respect and agree with. # 21:00 [shaners] aaronpk: I forget your opinion about OSS licenses. Is MIT intentional or just the boilerplate one? (vs CC0) # 21:00 aaronpk i have a hard time justfying using cc0 for libraries since it's not generally well accepted by lawyer types # 21:01 aaronpk I forget the differences between Apache and MIT, but I think I've been using Apache more lately # 21:03 [shaners] aaronpk: You’re right. I swapped MIT/Apache when typing. My bad. # 21:06 aaronpk hm yes. it was first written for the events, so the event organizer was the contact in person. # 21:08 bret because its the default on npm init -y bear joined the channel
# 21:11 [shaners] aaronpk: To be clear, I don’t have a strong opinion about which OSS license you should use on your repos. I was just curious. And had forgotten if we had talked about it. # 21:12 bear ISC is yet another variant of an open source license # 21:12 bret it clarifies some points that MIT is criticized for being vague on iirc # 21:13 bear it's a rewording of the MIT and the BSD 3 clause to be very precise/brief # 21:14 [shaners] bear: that’s not a very helpful definition of yet another acronym # 21:14 [shaners] the sarcasm / snark definitions of things on the wiki are really frustrating # 21:15 bear yea, that was an accident - working to fix it # 21:15 [shaners] I understand that this was an accident. But there are plenty of intentional snark pages. # 21:16 bear yes, i'm trying to reduce the number of them but it's a slow/idle-time project tantek joined the channel
# 21:22 aaronpk is CODE_OF_CONDUCT.md a github supported thing now? emmak joined the channel
# 21:26 tantek aaronpk: wish I'd heard about it then, I could have perhaps helped with pointing to some lawyerly resources, e.g. wiki.mozilla.org/Standards/licensing # 21:26 tantek shaners++ for pointing out stuff we need to explicitly add to the IndieWebCamp code-of-conduct # 21:27 tantek shaners - good point about the snarkinitions, I'll work on suggesting some explicit paths for improvement so anyone that comes across them can take a crack # 21:27 aaronpk I know of CONTRIBUTING.md but hadn't heard of the code of conduct one yet # 21:29 tantek I have to run now but want to make sure they're explicitly captured / handled # 21:29 aaronpk that post only talks about contributing, not code of conduct. so if you know of another post please update! # 21:31 [shaners] Gosh. I’m not reading well today. (…Cmon brain, wake up!) # 21:32 kylewm I've been converted to CC0 for writing Known plugins, with the hope that they could be integrated into core without any additional effort/attribution # 21:33 aaronpk kylewm: is that sometihng Known recommends for writing plugins? # 21:34 aaronpk that would be interesting if they requested a specific license # 21:35 kylewm (whereas I think it's cool that Mf2 et al are in that list) # 21:36 [shaners] aaronpk tantek: I think I was conflating CONTRIBUTING.md being an actual feature that GH recognizes and CODE_OF_CONDUCT.md just being an emerging convention. # 21:40 kylewm hahaha I just caught up to gRegorLove's rock puns # 21:43 aaronpk ooh i should make webmention.rocks keep the comments visible if they have likes from other domains, otherwise they drop off after 48 hours # 22:01 kylewm wonders how long before tantek adds .rocks to cassis Lancey joined the channel
# 22:09 bear (/me flips back to this screen after an impromptu work meeting) # 22:13 aaronpk it has to be right? Cause his webmention from his note worked # 22:18 aaronpk oh maybe not! his note text didn't include the link, just the in-reply-to URLs Nissyen1 joined the channel
# 22:29 aaronpk it reminds me of a regex that validates email addresses # 22:29 [shaners] Might be a case for something like this (but in php, not ruby) Jay- joined the channel
# 22:49 bear submits PR to tantek to add circleci build support for unit tests and linting tantek joined the channel
# 23:08 aaronpk Oh yeah, that sounds like a resonable request for telegraph # 23:12 kylewm aaronpk: do you have any intention of publishing XRay as a library? (as opposed to as a service). like I'm looking at XRay/Formats/Mf2, seems useful # 23:14 aaronpk I just copy/pasted that file into webmention.rocks # 23:14 aaronpk just need to think through the interface and how to tie in an http client to it # 23:14 bear eyes his "really want to do" list and wonders where the rich patrons are # 23:16 kylewm calling 2015, we want our unlimited supply of VC funding back # 23:17 bear heck, I would settle for a tech commune - give me a basic room, common area, community kitchen Jay- joined the channel
# 23:20 bear true, location isn't all that important - but I think my wife would not like living outside of the US :/ [shaners] joined the channel
# 23:28 [shaners] aaronpk: curious when you named it link-rel-parser and not rel-link-parser? # 23:29 [shaners] just realizing that (as far as I know), we always refer to rels as rel-*. # 23:32 aaronpk Well the * in rel-* refers to the rel value, not the name of the thing the rel is on tantek joined the channel
KartikPrabhu bear: does requests not normalise it in some way?
bear yea, I see <"> in the https://tools.ietf.org/html/rfc5988#section-5
aaronpk okay there's a little explantation on the home page now https://webmention.rocks/
gRegorLove Regarding wm discovery, my code doesn't differentiate between 'a rel' and 'link rel' currently, so I think it would just use whichever one it found first in the HTML.
Loqi gRegorLove: aaronpk left you a message 21 hours, 7 minutes ago: interesting point. however the authorship algorithm results in an author h-card, not an author URL, so the xmpp URL in that test case could still be used to match that h-card. http://indiewebcamp.com/irc/2016-04-10/line/1460274782358
miklb Project officially on GitHub https://github.com/miklb/jekyll-indieweb merciless criticism appreciated
[kevinmarks] Next, register trackback.sucks and make it display all trackbacks. It will rapidly live up to its name.
pfefferle acegiak is it only this part https://github.com/acegiak/Semantic-Linkbacks/blob/master/handler/semantic-linkbacks-microformats-handler.php#L196 ???
singpolyma miklb: I do on non-home pages, but that doesn't necessarily meant you need to ;)
tantek KartikPrabhu: right, publishers should really not have to care nor know about /authorship
kylewm.com edited /2016 (+26) "/* Assistance */ update with aaronpk's email instead of benwerd" (view diff)
kylewm volcano is a Landform with a faxNumber according to https://schema.org/Volcano.
loqi.me created /volcano (+106) "prompted by kylewm https://indiewebcamp.com/irc/2016-04-11/line/1460393961036 and dfn added by kylewm" (view diff)
snarfed miklb: heh, great minds. https://snarfed.org/2015-10-02_15607
tantek.com deleted /volcano "content was not relevant to the indieweb"
voxpelli tantek: regarding Travis CI, basically you just add a file like https://github.com/maxogden/dependency-check/blob/master/.travis.yml
kodfabrik.se edited /Planning (+29) "/* Brighton, UK */ Noting my interest in another european IWC, maybe in time for IW Summit?" (view diff)
[shaners] https://indiewebcamp.com/code-of-conduct needs contact information for people who are experiencing abuse or harassment.
bret its basically MIT: https://opensource.org/licenses/ISC
veganstraightedge.com edited /code-of-conduct-feedback (+249) "/* Open Discussions */" (view diff)
notes.whatthefuck.computer edited /Events (+147) "/* June */ add Decentralized Web Summit" (view diff)