2015-08-25 UTC
# 00:00 Jeena ha nice, snarfed that is more than good enough, the links are in the log too, even clickable!
cdchapman joined the channel
# 00:02 KartikPrabhu KevinMarks: consider adding it as a lib, the service causes problems to my apache
# 00:03 KevinMarks that's what I meant - adding it to my increasingly random set of tools there
tantek joined the channel
# 00:17 tantek Jeena, you have to just treat the date the webmention was received as the FB RSVP date
# 00:17 Loqi Ok, I'll tell them that when I see them next
# 00:18 Jeena yeah that I'm doing already, I just saw that it didn't match up because I did all of them manually within 2 minutes
lukebrooker joined the channel
# 00:25 snarfed tantek: thanks for starting that discussion! i mostly understand the writeup
mlncn joined the channel
# 00:26 snarfed seems like authoring UI vs how to interpret rendered posts are pretty different use cases. may be worth distinguishing them more loudly
# 00:26 snarfed i'm also with kylewm on the inconsistency issue. the <p>/<br> heuristic is ok; e-content vs p-content is better imho.
# 00:27 snarfed the heuristic *just barely* pushed me over the edge to consider implementing it in bridgy. e- vs p- is more convincing, esp if mf2 parsers expose that reasonably
# 00:39 KevinMarks so how does this auto-whitespace point of view interact with whitespace collapsing in implied name etc?
# 00:46 kylewm KevinMarks: are you thinking like, maybe we'd do the reverse? convert <br> to \n and to space?
# 00:46 kylewm I think that'd make more sense than the current behavior if you have like <span class="p-name">Kyle<br/>Mahan</span>
afrogeek joined the channel
# 00:47 kylewm looks like python replaces nbsp with the unicode non-breaking space character
# 00:51 KevinMarks I'm thinking of the problem where he implied name ends up with all the non-visible whitespace from the source
snarfed and almereyda joined the channel
KevinMarks__, KevinMarks_, sparverius and snarfed joined the channel
# 02:43 bret heroku is shutting off free access to 24/7 free apps. bye bye atom feeds on heroku apps (at least directly to the app)
KevinMarks_ and [snarfed] joined the channel
# 03:00 [snarfed] bret: blatant plug: app engine still has its 24/7 free tier :P
# 03:01 [snarfed] there are definitely also some js apps on gae, most inside the jvm
# 03:02 [snarfed] KartikPrabhu: agreed. heroku is a great platform. sad to see their free tier degrade
# 03:02 bret learning app engine is something I've been interested in a while
# 03:03 bret how does the irc bridge format the names?
# 03:06 bret yeah i get that ;p what mechanism allows it to send messages in that format to the channel without those users being in here?
# 03:44 aaronpk i'm gonna try to launch avatar archiving on webmention.io really quickly
# 03:47 bret aaronpk: how are messages bridged into irc and formatted with the name, and not attributed to a room member?
# 03:48 bret why doesnt it say Loqi: [snarfed]: says this etc
# 03:51 bret ohhhh i see. wow clever. turning on joins/parts explains the mystery :p
# 03:51 aaronpk node is great for that because it makes it easy to have a bunch of IRC clients running from the same process
# 03:56 aaronpk so now it won't accidentally keep web users stuck in IRC, and also the web chat is just post and get requests now, not websockets
# 04:00 bret ive only messed with socket.io which has robust disconnect/reconnect logic
# 04:01 aaronpk yeah i was using plain websockets before, and i thought i had handled all the events but there was some way to close a browser and not trigger the "close" event
# 04:05 bret socket.io does something with heartbeats.. not sure if related
# 04:05 aaronpk this whole process has made me think a lot about chat clients/protocols in general tho
# 04:06 aaronpk i might have a surprising new product launched in the not-so-near future
# 04:09 GWG aaronpk: I always look to you to have done something I have not yet thought about doing, or have thought of, but not gotten to.
# 04:10 aaronpk I am hoping to at least document some designs for this on the sooner side
snarfed joined the channel
gRegorLove joined the channel
# 04:43 GWG aaronpk: I don't know what to say to that.
# 04:44 aaronpk it means anyone using webmention.io is going to be getting new permanent author photo URLs that won't break when people change their profile images
[snarfed] joined the channel
# 04:56 aaronpk that may also be a new record for inception to launch for an idea i've had :D
mblaney joined the channel
lukebrooker joined the channel
# 05:04 mblaney so after getting sending webmentions working, I've been looking at syndicating posts
# 05:04 mblaney started with twitter, but quickly realised brid.gy was the way to go
# 05:06 mblaney so I signed up to brid.gy, which looks great but is telling me I need webmention support.
# 05:06 KartikPrabhu mblaney: you need webmention support to get backfeed of replies from syndicated posts
# 05:06 mblaney I guess that means I need to be able to receive webmentions?
snarfed joined the channel
# 05:13 mblaney I entered a post to check out the preview, and the webmention command line it provides is what I was after. very helpful!
# 05:22 mblaney snarfed if I publish a post with a p-name, brid.gy selects that instead of e-content is that correct? (for sending to twitter anyway...)
# 05:23 snarfed you can avoid that by omitting p-name for notes (ie tweets)
# 05:24 mblaney ah, thanks. I have read that page a few times now, but there's a lot to take in....
# 05:25 mblaney yes it's optional from my posting interface, but I guess I need to work on when I should POSSE to twitter or not.
# 05:49 mblaney hmmm it seems brid.gy is using any test inside an h-entry rather than looking for e-content?
# 05:49 Loqi mblaney meant to say: hmmm it seems brid.gy is using any text inside an h-entry rather than looking for e-content?
# 05:55 snarfed yup, and also that it found that content property
# 05:55 snarfed mf2 parsers have some extra logic about implying property values (among other things)...might be that
# 05:57 mblaney under properties it gets content right, but I don't understand what the "name" field is.
# 05:58 snarfed ah. that may be extracted from text when not explicitly set
# 06:00 mblaney but if I set it, won't bridgy use the p-name instead of the e-content?
# 06:01 snarfed yes. just use class="p-name e-content" to set both to the same thing
DePhraser, jciv, voxpelli, tommorris, KevinMarks__, lukebrooker, eschnou, tantek and pfefferle joined the channel
sanduhrs joined the channel
# 07:26 tantek considers line-by-line processing in auto_link
squeakytoy2 joined the channel
# 07:41 tantek !tell aaronpk,KartikPrabhu would you have any interest in using an auto_space function that turned line-breaks into <br/>s and both leading and multiple whitespace characters into to mimic white-space:pre-wrap but with markup?
# 07:41 Loqi Ok, I'll tell them that when I see them next
elf-pavlik and eschnou joined the channel
# 07:52 KartikPrabhu I haven't found any instance of my responses that cause me concern about white space
friedcell joined the channel
# 08:08 tantek KartikPrabhu: yes that's the point - for people (i.e. your friends) reading your posts in a /reader
# 08:09 KartikPrabhu yes possibly. But afaik no one but kylewm reads my posts in a reader :)
# 08:12 KartikPrabhu tantek: but in general I agree it would be good to have a consensus about how to read white-spaced posts
# 08:12 tantek KartikPrabhu: my current thinking is to suggest treating p-content (plain value) and e-content (HTML) differently
indie-visitor joined the channel
# 08:13 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
# 08:13 tantek that is, preserving whitespace in display for p-content (plain value), but not e-content (HTML)
# 08:13 tantek (which would break display of my current notes, hence I'm looking to fix that first)
# 08:13 KartikPrabhu distinguish depending on using the content property directly vs content>value?
# 08:14 tantek content property either just has a plain text string, or has that in a "value" and "HTML"
# 08:16 tantek whoa just found an aaronpk note POSSEd to an Amazon review!
# 08:17 Loqi Amazon is a company that sells various products and services, in particular several hosting services like S3 and EC2 that may be used for various indieweb site features https://indiewebcamp.com/amazon
# 08:17 KartikPrabhu yes. so if using content as plain string do some auto_space or if using ceontent as HTML don't?
# 08:18 tantek KartikPrabhu: no auto_space is what you would need to generate e-content as HTML
# 08:18 tantek if all you serve is p-content, then you're indicating to readers that they should preserve your whitespace with e.g. white-space:pre-wrap
# 08:18 tantek just as I (we?) currently use auto_link on the publishing side
# 08:21 KartikPrabhu tantek: re auto_space on publishing side, it might make authoring more tricky as one has to remember the mapping while writing a note. I'd rather my note be displayed using pre-wrap
# 08:22 tantek no the point is you should not have to remember any mapping while authoring
# 08:22 tantek it should just "work" to preserve what you type when you're authoring
# 08:23 tantek that's a reasonable expectation of an auto_space implementation
# 08:23 KartikPrabhu hmm in that case some one should code it and iterate according to authoring experience :)
_fran joined the channel
# 08:23 KevinMarks also, if tantek does add it to cassis we can implement live preview in client-side js
# 08:24 tantek currently I'm thinking of adding it as a separate function
# 08:24 tantek rather than adding it as an optional feature in auto_link
# 08:25 KartikPrabhu hmm I don't use cassis.js in any case. But I am in favor of having separate functions that others can iterate upon
# 08:35 tantek I definitely get the feeling I'm building a markdown replacement piece by piece in CASSIS
# 08:37 KartikPrabhu tantek: I am mostly not a fan of "simplified markup" syntaxes. Plaintext or HTML has been my POV; except when I have to resort ot Latex
# 08:38 tantek KartikPrabhu: I too am mostly not a fan of "simplified markup" syntaxes
# 08:38 tantek hence why I'm approaching it from a very "looks good in its plain text form regardless" perspective
# 08:40 KartikPrabhu progressively enhanced text! I like it as a concept but I think markdown was also started with the same goal
pfefferle joined the channel
# 08:41 KartikPrabhu " The idea is that a Markdown-formatted document should be publishable as-is, as plain text, without looking like it’s been marked up with tags or formatting instructions."
# 08:43 KartikPrabhu of course not criticising Markdown principles. but somewhere down the line it got out of Gruber's control
# 08:43 tantek I'm pretty sure I did the proper "go back to first principles" thing and came to a different conclusion that daringfireball
# 08:44 tantek I think Gruber himself sacrificed that principle even with just markdown hyperlinking syntax
# 08:44 KartikPrabhu but if a nice plain text to HTML convention was made, I'd certainly give it a serious thought
# 08:45 tantek that alone is sufficient justification to dump markdown
# 08:46 KartikPrabhu it was a conscious decision on my part to not use markdown for authoring notes. plaintext for notes (with auto link) and HTML for articles
# 08:54 KartikPrabhu actually that is mostly how I distinguish whether to post a note or an article. Is is writeable in plain text
# 08:59 tantek KartikPrabhu: my goal is to preserve that "writeable in plain text" aspect
# 09:02 KartikPrabhu I have been thinking about good footnotes markup myself but haven't figured out a solution yet
pfefferle joined the channel
# 09:18 tantek 4 is not hard - already implemented in auto_link
# 09:19 tantek 8 I just realized I'm not sure about because what about nested lists
# 09:21 KartikPrabhu that's why white-spacing is complicated. lots of conventions use white -space for different thing
tvn and j12t joined the channel
adactio joined the channel
j12t, stream7 and glennjones joined the channel
LanceyWork, frzn, j12t and nedorito joined the channel
pfefferle and j12t joined the channel
friedcell, pfefferle, tantek, cmhobbs, j12t, benwerd, elf-pavlik, fourtonfish, snarfed, mlncn and wolftune joined the channel
# 14:24 tantek.com edited /Falcon (+1659) "/* event posts */ POSSE events use-cases with specific per use-case with conditionals POSSE destinations" (
view diff )
# 14:27 GWG pfefferle: Hate to be a bother, but asking about Semantic Linkbacks again.
# 14:33 snarfed GWG: pfefferle: ooh me too! can't wait to see the repos merge :P
# 14:40 GWG pfefferle: I just wanted an updated WordPress.org. snarfed wants salmentions
glennjones joined the channel
# 14:49 pfefferle oh, ok the update is no problem the salmention is one… the code of acegiak is far away from beeing productive… a lot of var_dumps and other debug stuff
j12t, tvn_, pfefferle, jciv, benborges, scoates and indie-visitor joined the channel
sparverius, KevinMarks_, j12t, shiflett and sanduhrs joined the channel
sparverius joined the channel
# 15:59 tantek hey jonnybarnes - no guys here just people ;)
niamu joined the channel
adactio, snarfed, squeakytoy, tantek and elf-pavlik joined the channel
j12t, davidpeach, sparverius, gRegorLove, almereyda, jonnybarnes and tantek joined the channel
# 17:43 gRegorLove I still need to do some re-processing of wm on there. Aaron's should be a like.
[snarfed] joined the channel
# 17:45 [snarfed] Jeena: did you add the second link to that FB event after initially creating it? bridgy may have fetched and cached it in between
# 17:46 [snarfed] Jeena: rsvps after you added the second link should be sent to both links (i think)
# 17:54 Jeena and no the rsvp after I added the second link didn't send a webmention to it
# 17:54 Jeena but I will test it in two weeks with the link already there from the beginning :)
# 17:57 davidpeach really need to tidy up my post UI tonight though. I've found myself adding check boxes for this n that and it's starting to get out of control.
# 18:00 Jeena yeah I already did that, no problem, and don't be sorry for putting out an awesome service which I'm using all the time for free, hm I guess I should look for the donate button
# 18:01 Jeena oh damn, there is none, ok then I will check out what ASPCA and/or EFF are
scor joined the channel
# 18:03 Jeena hehe yeah, money is always the easy answer for those who have it :-p
[kylewm] joined the channel
# 18:07 [kylewm] snarfed: there does seem to be something wrong there, it thinks it finds two original posts, but only sends wm's to one of them
# 18:09 Jeena ok now I donated a small amount of money and perhaps in the future I will be able to make some pr
# 18:14 [kylewm] snarfed: Jeena: ok I think I get it. Because the Response["invited"] was created when the event was first processed, the Response object still has the one original URL. then when it changes from invited to Response["rsvp-maybe"], it doesn't check again for new original URLs
MylesBraithwait- joined the channel
# 18:21 GWG I've been thinking about this since yesterday. Shiu
# 18:22 GWG Should I build display functionality with no built in post UI
# 18:23 [snarfed] GWG: sure! i author a decent amount of indieweb/mf2 stuff as raw html, with no dedicated ui
# 18:23 GWG snarfed, I am thinking of post types where the post UI could be Micropub
# 18:24 GWG For example, Teacup. I don't have a built in UI for food
# 18:24 Jeena I always do: rails generate scaffold events title:string from:datetime to:datetime description:text geo:string address:string
# 18:25 Jeena and the database and the model, etc. too obviously
# 18:28 GWG I'm just seeing a bunch of Micropub clients that could do the work for me.
# 18:29 GWG So, I suppose that the question is more built in UI
# 18:34 GWG The Independent part of me hates to rely on third parties 100%
# 18:34 kylewm GWG: I think the next iteration of blog thing I write will not have its own post UI (and I know aaronpk is moving that way too)
# 18:35 aaronpk yeah, i'm not building a posting UI for my new site
# 18:35 Loqi aaronpk: tantek left you a message 10 hours, 54 minutes ago: would you have any interest in using an auto_space function that turned line-breaks into <br/>s and both leading and multiple whitespace characters into to mimic white-space:pre-wrap but with markup? http://indiewebcamp.com/irc/2015-08-25/line/1440488504381
# 18:37 GWG I suppose that anyone can self host.
# 18:38 aaronpk and if you don't like quill you can build your own. the point is I think there is a benefit of separating those components even if you build both yourself
# 18:39 GWG aaronpk, I am thinking of more than myself. People already find Indieweb and WordPress confusing
# 18:39 aaronpk i'm rebuilding my site from scratch, but the cool thing is I won't need to rebuild Teacup or Quill to work with my new site
# 18:39 GWG On the other hand, I can get more functionality faster by not building a post UI now.
# 18:40 GWG Nothing says that I can't do it later.
sparverius joined the channel
shiflett, cmhobbs_ and frzn joined the channel
pfefferle joined the channel
# 19:24 aaronpk avatar URLs from webmention.io are now on that domain!
eschnou joined the channel
# 19:30 aaronpk I could go back and try to archive old ones, but i haven't yet
# 19:35 aaronpk finally found an appropriate use for content-addressable storage and Lambda!
friedcell and tvn joined the channel
friedcell joined the channel
cleverdevil joined the channel
Unifex joined the channel
j12t joined the channel
j12t and KevinMarks joined the channel
# 20:49 GWG pfefferle: Not a problem. Thank you though. I've been eagerly awaiting this. It allows me to do several things
tantek joined the channel
# 21:17 aaronpk kylewm: what do you think about adding twitter support to silo.pub?
# 21:18 kylewm aaronpk: I'm into it, just need to figure out how to make a fake profile page that points to the right authorization_endpoint and stuff
frzn, sparverius and tantek joined the channel
# 21:27 KevinMarks hm, isn't this the "if there's native auth, don't do the rel-me auth dance" issue?
# 21:30 tantek KartikPrabhu: when did you notice it? (the class name :) )
# 21:37 kylewm aaronpk: do you have a use case for silo.pub / twitter?
# 21:38 aaronpk it came up just now from rhiaro but then I realized I would totally use it too
# 21:39 aaronpk i mean it doesn't really have to be micropub, but might as well be
# 21:41 pfefferle GWG I am also working on my local sendmail to check your pull requests
lukebrooker and tridnguyen joined the channel
# 21:50 tridnguyen I have a question about indieauth.com
# 21:50 tridnguyen does it matter if the user's domain is http or https?
# 21:51 aaronpk but if the page is served from both then they are treated as different URLs
# 21:52 aaronpk the full URL, not just the domain, since some people sign in with a URL that includes a path
# 21:53 tridnguyen makes sense :)
# 21:53 tridnguyen heh - thanks a lot
# 21:53 tridnguyen I just tried out indieauth today for myself - really like it alot
# 21:54 tridnguyen didn't know about it before
# 21:54 tridnguyen I saw it on Kickball/awesome-selfhosted
# 21:55 tridnguyen btw, the code that is returned from indieauth in step 3, that is only to verify in step 4? should it be stored as session token?
# 21:56 tridnguyen so the user would need to log in every time they come to a site that uses indieauth?
# 21:56 aaronpk no, you can set your own server-side cookie at that point
# 21:56 tridnguyen with a session token?
# 21:58 aaronpk it would be like if they had entered a username and password, how would you "log them in"
lukebrooker joined the channel
Zegnat and djwesto joined the channel
cdchapman joined the channel
# 22:37 tantek it's like someone's personal indieweb tools wiki page?
# 22:37 aaronpk looks like one page in a series of "awesome" tools of various sorts
mlncn, benwerd and almereyda joined the channel
# 23:03 KevinMarks did he really mean to suggest replacing DNS with js-based lookup on a single site?
# 23:03 KevinMarks it's one of those posts where I should respond but it woudl take ages
# 23:07 aaronpk omg this script has been importing 2000 avatars per hour for the past 3 hours
typhoo and cdchapman joined the channel
# 23:33 kylewm hehe, you could probably roast marshmallows over its cpu
# 23:34 kylewm also, any chance Lambda will notice that you are doing a whole bunch of SHA1 calculations and think you are mining bitcoins?
[snarfed] joined the channel
# 23:38 [snarfed] i wonder if there's some side channel to determine when a cpu is doing specific kinds of computation like hashing... but probably very difficult
[aaronpk] joined the channel
# 23:44 [aaronpk] Well I'm only making one request at a time, and each takes 1-2 seconds. Not sure if that's because of the sha calculation or just the network requests
# 23:45 [aaronpk] Hah I wonder how many bitcoins you could mine on their free tier. They probably thought of that already tho.
# 23:52 GWG I'm going to keep working on that, but anyone have comments on the Why?