2015-06-16 UTC
j12t joined the channel
KartikPrabhu joined the channel
wolftune joined the channel
tilgovi joined the channel
KevinMarks__, KartikPrabhu, snarfed, renoirb_, rbrt, KevinMarks___, KevinMarks, danlyke_, wolftune, cmhobbs, elf-pavlik, emmak and tilgovi joined the channel
corms_, lukebrooker, evalica, tantek, snarfed and cweiske joined the channel
# 04:25 cweiske 3: client_id ..An authorization endpoint may show the application's icon and title to the user
# 04:25 cweiske 4: It must indicate: The client_id making the request.. Including the name and logo if an h-card is found on the client_id URL
# 04:25 cweiske do we want to force indieauth servers to support MF?
# 04:26 aaronpk technically it says must indicate the client_id, and separately the name and logo *if* an h-card is there
# 04:26 cweiske also another issue: Is there a way for the auth server to verify client_id?
# 04:27 cweiske aaronpk, this still sounds like I have to check for a h-card
# 04:27 aaronpk if you don't want to check for h-card you can just display client_id
# 04:28 aaronpk yeah I would say looking for an h-card is optional. actually i think that is currently implemented as h-x-app
snarfed joined the channel
KartikPrabhu joined the channel
KartikPrabhu and yakker joined the channel
snarfed, elf-pavlik and tantek joined the channel
# 05:55 Loqi Ok, I'll tell them that when I see them next
# 05:56 Loqi Ok, I'll tell them that when I see them next
# 05:57 KartikPrabhu also the permalink should scroll down to the actual post instead of showing the response-context
tilgovi joined the channel
# 06:10 Loqi Ok, I'll tell them that when I see them next
pfefferle joined the channel
edpw and pfefferle_ joined the channel
cweiske joined the channel
Jihaisse joined the channel
j12t, friedcell, Zegnat, csarven and loic_m joined the channel
KevinMarks__ and elf-pavlik joined the channel
# 07:36 Loqi Ok, I'll tell them that when I see them next
j12t, evalica, pfefferle and stream7 joined the channel
eschnou, pfefferle, KevinMarks and KevinMarks___ joined the channel
friedcell and Sebastien-L joined the channel
pfefferle, KevinMarks, glennjones, Sebastien-L, modem, KevinMarks__, j12t, tvn, LauraJ and sammachin joined the channel
# 09:53 cweiske so h-entries are not suceptible for that kind of errors?
mlncn, pfefferle, adactio, tmro and frzn joined the channel
# 10:39 Zegnat cweiske: I see no reason why h-entries wouldn’t be susceptible for wrong date stamps. It all smells like manually-typed timestamps, those are easy to mess up
pfefferle, LanceyWork and j12t joined the channel
petermolnar and friedcell joined the channel
# 11:29 GWG pfefferle: I couldn't sleep, so I sent you a pull request.
# 11:31 petermolnar just by pure curiosity, is anyone here using (open|free)bsd for their servers?
# 11:33 cweiske Zegnat, but some/many arguments agains (rss|atom) feeds have the base that feeds may be invalid
# 11:33 GWG pfefferle: In the PR, reply is p, everything else is u. That should work.
# 11:34 GWG The example shows u-like on the article tag
# 11:34 petermolnar I will always keep my rss feed, even if it's only for the memory of Aaron Swartz
# 11:36 GWG But that is publishing one vs displaying an incoming one
# 11:36 pfefferle it is a parsing thing… for u- you have to parse href, src and date
# 11:36 Zegnat cweiske: RSS requires RFC 822 dates, if you manually write those you can mess up. HTML5 requires “valid date string” / “valid time string” (while ISO 8601 based HTML doesn’t seem to use ISO 8601 ???), if you manually write those you can mess up.
# 11:39 cweiske Zegnat, just read the "Consuming feed files is problematic and fragile." link on the /feed page
# 11:40 Loqi cweiske meant to say: all] of them are format-agnostic problems
# 11:40 GWG pfefferle: Everytime I think I have it...
# 11:42 pfefferle GWG and you don’t have to remove the h-as stuff… add the p- things additionally
# 11:42 Zegnat use p- if the HTML tree doesn’t matter. If there are HTML elements within your container that are important (i.e. they are part of the “value”) use e-
# 11:43 pfefferle GWG Zegnat and there is another possibility ”žp-comment h-entry“ means, that the h-entry is a comment of the h- microformat above
# 11:44 Zegnat p-comment h-entry sounds… odd. h-entry is a root class, so I would guess you have several data elements in the tree below that? So why have p-comment and not e-comment?
# 11:46 pfefferle Zegnat <div class=”žh-entry“>… <div class=”žp-comment h-entry“>… </div> </div>
# 11:48 Zegnat pfefferle, reading it now, and I personally think those should be e-
# 11:48 Zegnat why would a parser ever want to parse those as text?
# 11:59 Zegnat If I look there, the value of a comment parsed by a microformats parser is:
# 11:59 Zegnat "value": "Tim Pietrusky hat dieses Audio-Dokument auf twitter.com ge-\"liked\"\n\n \n \n Tim Pietrusky \n \n 3. Mai 2015 bei 15:15"
# 11:59 Zegnat I *think* if you would mark it up as e-comment that value should include the full “html” as well, which makes more sense to me here.
pfefferle_ joined the channel
# 12:00 Zegnat pfefferle_: did you just get my 3 lines about the value of your comments, or did you timeout for those?
# 12:02 pfefferle We are still not talking about the same… I only wanted to mention that you can define ”žsup-formats“ if you are using p-something with h-something ;)
# 12:03 Zegnat I just hope I am making sense, it is a bit of a weird issue. Overall, I think *all* mf implementations understand that there is a comment there, and they will happily parse the h-entry metadata, BUT if you were to ask an mf parser for the value of the comments on your blogpost you will get plain text back ("value" of "comment" on pin13) rather than the HTML you might expect.
# 12:03 Zegnat Totally getting sub-formats, only doubting the use of p-
# 12:04 Zegnat cweiske: you are right, most of the points made on Brent Simmons’ blog re: parsing feeds apply to HTML and mf. There is no real reason to think they are RSS only and implementers of h-feed readers might have to jump through hoops just as much
# 12:05 tommorris cweiske, Zegnat: sure, but a lot of them flow from invisibility of the data. HTML is a lot more visible than RSS
# 12:06 cweiske When the feed is gone/errored, publisher may still return a 200 OK but send an HTML page instead.
# 12:06 cweiske Using permanent redirects for temporary errors - same
# 12:06 cweiske Very slow responses, or responses that never actually complete - can happen with html, too
# 12:06 cweiske Infinitely long responses. eg, feed server has an error and prints an error message in a infinite loop until something stops it - same
# 12:07 tommorris the point is more if someone uses a fucked up date format on their HTML pages, you see it. if they do so in their RSS feed, the only person who sees is it is the grumbly programmer who writes rssdateformats.tumblr.com ;-)
# 12:07 cweiske sometimes they misidentify RSS as something else (eg, text/plain). - same
# 12:07 Zegnat You don’t, date formats are often under <time datetime>
# 12:07 cweiske Returning an HTML page containing HTML/Javascript redirects instead of using HTTP redirects. - same problem with html
# 12:07 tommorris header/meta stuff is invisble regardless of format, but stuff in human-readable HTML surfaces a lot more than stuff in machine-readable XML
# 12:08 tommorris because strangely enough humans consume the HTML and not XML
# 12:08 Zegnat tommorris: your own timestamps are invisible data, just sayin’ ;)
# 12:09 Zegnat rssdateformats.tumblr.com - has a point, people don’t understand timestamps
# 12:09 tommorris it has a broader point - errors lurk for longer in RSS feeds than they do in HTML
# 12:09 Zegnat cweiske is mostly talking about the Brent Simmons link, which makes good points that are also valid for HTML h-feeds
# 12:10 Zegnat there is no proof of that. There might be blogs right now using illegal values for datetime attributes but we don’t know yet
# 12:10 cweiske tommorris, nobody notices if you screw up your h-feed and h-entry classes and timestamps and other hidden metadata
# 12:10 cweiske not a single of your users looking at your page in the browser
# 12:11 Zegnat It is a bit of a moot discussion though, until h-feed gets half the adoption of RSS we simply can’t tell how people might mess up
# 12:12 cweiske that's the correct answer: not *if*, but *how* they will mess up :)
pfefferle_, pfefferle, sanduhrs, modem and jacus joined the channel
# 12:51 GWG pfefferle: I'll revise submission then tonight. Having the -like makes the test parsers happy
# 12:53 GWG I also have to switch from p-content the
# 12:56 GWG That reminds me, this is still a bit rough but how does the mention/comments presentation look on this?
modem and tvn joined the channel
# 13:08 GWG I think I got the presentation to a nice level of simplicity
pfefferle joined the channel
j12t and pfefferle joined the channel
# 14:01 Zegnat GWG: it is a clear presentation, but I feel a little weird about the error. Wouldn’t at/on/@ be better? “Scott Eckers @ facebook.com: [comment]”
KevinMarks, fourtonfish and cmhobbs joined the channel
# 14:33 jonnybarnes anyone here running a push server? what value for lease_seconds do you use?
# 14:35 cweiske to make it possible to run a cronjob every 7 days
# 14:36 cweiske hubs could implement ISO 8601, which specifies durations like P1Y, P2W, P2D
loic_m and chalettu joined the channel
# 14:51 aaronpk jonnybarnes: I do 3 days right now. I like cweiske's idea of adding on an extra hour tho
# 14:54 aaronpk i mean ultimately subscribers should proactively refresh feeds before they expire, so like checking on some more frequent cron like every hour and refreshing ones that are expiring within an hour
Pierre-O joined the channel
# 15:00 cweiske aaronpk, when will indieauth.com send the client_id to indieauth servers?
pfefferle, KartikPrabhu, snarfed and j12t joined the channel
# 15:14 cweiske doesn't sound like it's on top of your priority list
# 15:14 aaronpk if I could spend all my time on indieweb things I would :)
wolftune joined the channel
# 15:21 rhiaro I'm working on making indieweb enough of my PhD that I can work on it all the time :)
# 15:22 rhiaro aka I'm only doing indieweb/socialwg stuff anyway so I have to summon a thesis out of it or I won't have anything
benborges and afrogeek1 joined the channel
# 15:30 jonnybarnes yeah, thats what the spec says right? If a subscription doesn't renew by the time luease_seconds has expired, then forget about that subscription?
# 15:31 aaronpk yeah, the hub stops delivering to the susbcriber after it expires
snarfed and tantek joined the channel
# 15:49 KevinMarks So apple now supports SVG favicons but still wants multiple bitmaps as well?
# 15:50 aaronpk maybe that's a sign they'll be moving to SVG for app icons?
# 15:58 Zegnat KevinMarks, do you have a link about that for me?
# 16:02 tantek !tell benwerd for that Tumblr sprint, I suppose we need to better document all the different Tumblr post types and indieweb equivalents then right? so that export Tumblr, import to IndieWeb, and then POSSE IndieWeb -> Tumblr are work with high fidelity. :)
# 16:02 Loqi Ok, I'll tell them that when I see them next
# 16:09 kylewm tantek__: the small icon is because it's a copy of a copy... and my old design used very small avatar icons. perhaps I should mark up avatars with a data tag that has the full-size image!
snarfed joined the channel
# 16:13 kylewm heh, my salmentions implementation needs some work... it bypasses the normal de-duplication it seems
friedcell and j12t joined the channel
# 16:19 kylewm I very frequently have this problem with mf2 markup where i want two <div> columns | avatar | name and content |
# 16:19 kylewm very hard to markup the "p-author h-card" that way
gRegorLove joined the channel
j12t joined the channel
# 16:42 kylewm KevinMarks_: is it possible to clear the column below the avatar then?
# 16:43 kylewm I think you can do something like that with overflow: off
# 16:44 KevinMarks_ there's probably a way with giving the image a negative offset so it goes into the column
# 16:48 KartikPrabhu kylewm: you shouldn't need to use 2 divs to make such columns must be doable with CSS. here is a way
# 16:49 KartikPrabhu put everything in the "content" column including avatar, make it have a right margin of the size of avatar. use position: relative on avatar to shift it to the left
# 16:51 kylewm got it, OK thanks I will try that. Should work better on mobile too
# 16:52 KartikPrabhu kylewm: this also worked in FF dev tools. I moved the image into "content-column". content-column has margin-left:3em and then the image has margin-left:-3em;
# 17:03 kylewm KartikPrabhu: do i still need to float: left on the image?
# 17:04 KartikPrabhu i don't think so let me check. So I just moved the <img> tag which has no float and it worked
# 17:07 KevinMarks_ 'cos I have seen weird crap like one word a line when people do fixed margins
mlncn joined the channel
# 17:08 kylewm feels bad for people who don't have quantum physicists to help fix their css
# 17:09 KartikPrabhu if the stupid writers at Big Bang Theory use this exchange somewhere, we can sue them now :)
# 17:11 kylewm thanks y'all, that fixed at least two problems i was having
# 17:12 KartikPrabhu I still don't understand flex-box tbh. Really complicated stuff for CSS
# 17:17 KevinMarks_ pinterest has fixed width columns horizontally, with variable height items vertically that tuck in under each other
# 17:18 KartikPrabhu KevinMarks_ I think flexbox is more about getting uniform heights to adjacent things
# 17:19 KartikPrabhu though I have tried that layout before (with JS) and it fails miserably if the "boxes" have wildy different heights
benwerd, danlyke, KevinMarks__, snarfed1, bret-raspi, snarfed, KevinMarks, bret1, nxd4n, gavinc and cweiske joined the channel
# 18:23 cweiske aaronpk, indieauth.com has to pass indieauth.com as client_id because auth servers use it to verify the redirect_uri parameter
snarfed joined the channel
# 18:24 aaronpk oh yeah and the redirect_uri is always gonna be indieauth.com/something huh
# 18:24 cweiske so unless you want to require apps using indieauth.com to know and to link to the indieauth.com redirect uri, you cannot use their URL as client_id
# 18:25 aaronpk hm well that isn't the worst idea actually, because if an app is using indieauth.com it's because it has the trusted relationship with it
bret1 joined the channel
# 18:26 aaronpk KartikPrabhu: no this is when an app uses indieauth.com rather than implementing indieauth protocol on its own
# 18:27 aaronpk i could make this whole thing more explicit by requiring people sign up for an account before their app can use indieauth.com or something
# 18:28 Loqi aaronpk meant to say: i could make this relationship more explicit by requiring people sign up for an account before their app can use indieauth.com or something
# 18:31 GWG The more I think I get things, the less I do
# 18:32 aaronpk it might be useful to me, to make sure I have contact info for people using it in case I need to get in touch
eschnou joined the channel
# 18:33 aaronpk or if I had a legal department it would give me a chance to make people click the "I agree to the terms and conditions" box before using it
# 18:34 ben_thatmustbeme the only thing i could see is the ability to man in the middle and change the redirect_uri, many oauth endpoints make you specify where you will redirect beforehand
uhhyeahbret joined the channel
# 18:35 aaronpk which is handled in indieauth by publishing your valid redirect URIs on the client_id page
cmhobbs joined the channel
# 18:37 ben_thatmustbeme yeah sorry i missed a bunch of the earlier stuff, i was distracted watching an entire orchestra play 4'33"
uhhyeahbret joined the channel
# 18:37 Loqi ben_thatmustbeme meant to say: yeah sorry i missed a bunch of the earlier stuff, i was distracted watching an entire orchestra perform 4'33"
bret and Sebastien-L joined the channel
davidmead joined the channel
# 18:47 snarfed and i'm happy to (try to) answer any app engine q's
# 18:47 snarfed the GA label is mostly just marketing/messaging - the PHP support has been available and solid for a while - but still
# 18:48 aaronpk most important: is the free tier per app or per account? :D
# 18:49 cweiske " All applications are free within a usage quota that is reset daily. Paid applications can use more at the prices below. "
# 18:50 aaronpk I wonder if I should rewrite my PuSH hub to work on appengine
# 19:06 aaronpk "You are receiving this message because we noticed you were having trouble logging in to your account."
squeakytoy joined the channel
tantek joined the channel
# 19:12 aaronpk hm apparently I can't use a google voice phone number for login verifications on twitter, and twitter won't let me use the app-based verification without adding a phone number which is dumb
# 19:15 cweiske do you still have them or shall I try to login again?
# 19:15 aaronpk it's scrolled way back, i need you to try to log in again
davidmead and benwerd joined the channel
# 19:23 cweiske aaronpk, I don't have a ruby setup to test it. can you roll that patch out on a test server or even live?
# 19:27 Jeena so the homebrew club in portland is tomorrow?
# 19:28 aaronpk oh I think the redirect took you back to a different server :P
# 19:31 cweiske which means I can dump the github link from my homepage now.
# 19:35 cweiske aaronpk, indieauth.com does not link the redirect_uri in its html head
# 19:38 aaronpk more ruby was required for your last fix than this one :P
# 19:43 KevinMarks_ does PHP on appengine need big rewrites to make things work? Given how appengine discourages filesystem access, SQL etc
indie-visitor joined the channel
# 19:46 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
# 19:46 snarfed mysql is definitely supported (on goog cloud sql)
# 19:46 snarfed afaik they have wordpress, drupal, and mediawiki running
# 19:47 snarfed hah yes, original app engine was in 08 after all :P
# 19:47 KevinMarks_ true, but you know what I mean - the python and java stuff expected you to use datastore
# 19:49 aaronpk yeah i'm impressed and pdx needs to step up somehow
KartikPrabhu joined the channel
# 20:01 cweiske now I can add redirect_uri verification to the indieauth-openid proxy
davidmead joined the channel
eschnou and tvn joined the channel
# 21:02 gRegorLove I know tantek suggested it, but has anyone started reaching out to /2014 Portland attendees?
# 21:03 gRegorLove I can send some tweets. Don't want to pester people if they've already been asked though.
csarven, lukebrooker, tilgovi, KartikPrabhu, KevinMarks__ and KevinMarks joined the channel
shiflett, almereyda, LCyrin, edpw, mlncn and KartikPrabhu joined the channel
# 23:27 aaronpk oh great I got added to a giant twitter group message that is spam
snarfed1 joined the channel
# 23:36 Loqi slack/tantek: How? Did you enable anyone to DM you?
# 23:36 aaronpk yeah, and then after that they launched group messaging
# 23:47 Loqi slack/tantek: Hey Gregor could you start a /invitation#Prototypes section like this one and start documenting your "text only" invitation notes/tweets?
# 23:55 Loqi slack/tantek: Even though "Are you going to ?" Is not quite an invitation, would still be helpful to capture the practice
# 23:55 Loqi slack/tantek: Or what you came up with as a text equivalent when aiming for a Twitter based invitation message.
shaunguice joined the channel