2015-05-29 UTC
# 00:05 tantek KartikPrabhu: why would checking original-post-discovery be any different than checking u-syndication ?
# 00:07 kylewm well many people don't include perma(short)cite/links in their tweets, not to mention indie-reposts or indie-likes where it isn't possible
# 00:07 KartikPrabhu tantek: checking u-syndication is easier since you have permalink of the post that webmentioned and get mf2 from that. For original-post-discovery there is a long-ish non-trivial algorithm to follow
# 00:08 tantek ah, assuming you can't follow your nose in those cases
# 00:09 tantek Ok I have a hypothetical reason - untested - which comes to the same conclusion as you KartikPrabhu re: having the u-syndication already there when a webmention receiver retrieves the original: POSSE reply threading
# 00:10 KartikPrabhu particularly if for some reason Bridgy backfeed hasn't yet sent the backfeed
# 00:11 tantek when indie A replies to indie B and sends B a webmention, when B goes to reply to A, if B's copy of A's post already knows the Ap POSSE copy, then B can immediately reply back both to A, and POSSE reply thread their Bp POSSE copy to Ap.
# 00:12 tantek it makes it easier to do the right thing, and connect all the replies
# 00:13 tantek and yes KartikPrabhu it should make de-duping easier since u-syndication is easier to retrieve than it is to do original-post-discovery
# 00:13 tantek now where to capture this FAQ with answer with 2 solid reasons
# 00:14 KartikPrabhu possibly and link to it from POSSE or some page that has this deduping of comments
tilgovi joined the channel
# 00:25 tantek.com edited /Webmention (+189) "move rel webmention from brainstorming to resolved, id up near discovery, webmention as one of many space separated values" (
view diff )
scoates, nloadholtes and KartikPrabhu joined the channel
scoates and KevinMarks joined the channel
# 01:08 tantek.com edited /Webmention (+1148) "/* How does this solve the spam problem */ note vouch. POSSE or webmention first? Thanks to kylewm and kartikprabhu for reasons." (
view diff )
# 01:11 tantek.com edited /POSSE (+229) "/* FAQ */ POSSE or Send Webmentions First - summary answer an link to longer answer on webmentions page" (
view diff )
callovarne joined the channel
# 01:17 GWG tantek: That is the guy who was watching my development, I think
# 01:26 KartikPrabhu hence I don't think one should put permalinks unless the POSSE copy is truncated. Even then some words like "more at:" would be better
# 01:27 tantek KartikPrabhu: I believe that's pretty close to current recommended practice
# 01:28 KartikPrabhu on that note, you have yourself experienced people being confused about the strange link like letters at the end of a tweet no?
craftycorvid joined the channel
# 01:29 tantek it's not that, but rather that it still looks too much like a link and people wonder why it's not clickable
# 01:29 tantek instead of wondering why when they click it they see no new content
# 01:34 tantek.com edited /Twitter (+587) "/* POSSE abbr note to twitter */ include kartikprabhu's suggestion about More: along with some implications of attempting that approach" (
view diff )
# 01:35 tantek KartikPrabhu: ok I captured your "More (at):" suggestion as it seems reasonable
# 01:40 tantek was that a manual POSSE tweet or via Bridgy Publish?
# 01:44 tantek KartikPrabhu: that's very odd - you're POSSEing a like kind of like a quote?
# 01:46 tantek.com edited /Twitter (+260) "/* POSSE abbr note to twitter */ note KartikPrabhu's like example that has a quotation, and then he POSSEd manually as a note to Twitter" (
view diff )
LCyrin joined the channel
# 01:55 Loqi slack/snarfed: KartikPrabhu out of curiosity, would you use bridgy publish for your current manual posses if you could customize the ellipsizing and link text?
KartikPrabhu joined the channel
# 01:58 KartikPrabhu tantek: err yes. didn't know how else to do it since there is no Twitter copy
# 01:59 KartikPrabhu snarfed: I use Bridgy Publish for most cases, except when heavy editing of the text is needed to make it read nicer on Twitter
# 01:59 tantek KartikPrabhu: well it's odd that you're liking and quoting at the same time
# 01:59 tantek you've almost written your own note post there - since the actual quote is not continuous in the IRC channel archive you link to
# 02:00 KartikPrabhu true. I also add my own text to likes sometimes to keep note of why I like it
jciv joined the channel
# 02:08 tantek KartikPrabhu: that sounds more like a bookmark
# 02:08 GWG That reminders me. I keep thinking about learning Javascript
# 02:17 tantek email so f'ing depressing. it's like 90% growth hacking bullshit *after* you wade through the random unicode spam.
indie-visitor joined the channel
snarfed, n8fr8, callovarne, KevinMarks_, KevinMarks and j12t_ joined the channel
# 04:09 tantek aaronpk: perhaps not quite as productive a wiki day as yesterday, yet answering that FAQ I think is useful for helping implementations that do POSSE & send webmentions
LCyrin joined the channel
# 04:22 GWG You know, I still don't feel I have POSSE down.
# 04:30 tantek GWG, it's taken a while and many iterations to get to where we are with understanding good POSSE practices
# 04:31 tantek if you can formulate a specific question, perhaps we can clarify further
gRegorLove joined the channel
# 04:31 GWG I feel, like many things, I have not made it as easy as I can.
# 04:33 GWG My biggest problems come from trying to make things simpler and easier.
# 04:33 tantek yes, POSSEing should be completely automatic for the author
# 04:35 GWG I keep punting switching to Bridgy Publish down the road.
# 04:35 GWG Because to test it well, I feel like I'd be spamming my acount
# 04:36 GWG I also always had trouble POSSEing to Twitter
# 04:47 tantek GWG, definitely setup a test Twitter account, e.g. gwg_test
# 04:47 tantek and then just use that until you're happy with the results
# 04:47 tantek you can even make it private so that you don't spam search results (AKA Loqi searches :) )
# 04:48 GWG They still have private accounts? Good to know.
# 04:49 GWG But, either way, we'll see what I can do over the course of the summer.
# 04:49 GWG To get Bridgy Publish working, I need to fix my theme, because Bridgy Publish was picking it up incorrectly. Which means I'm marking up incorrectly.
# 04:51 tantek GWG - go get your private test account - that's an easy thing you can take care of in a few minutes
KartikPrabhu joined the channel
# 05:10 tantek GWG, harder as in fixing the markup in your theme
# 05:10 GWG tantek: I've fixed a lot in the unstyled version.
# 05:10 GWG I have to double check then restyle.
# 05:13 tantek GWG, right, that's more work than creating a new private twitter account :)
# 05:14 GWG Right now, I have to finish getting the dev setup on the computer I'm bringing on the plane next Wednesday finishd.
# 05:14 GWG But, either way, I will set one up.
KevinMarks, snarfed, KevinMarks_, Jihaisse, glennjones, fkooman and tantek joined the channel
# 06:52 acegiak So here's an interesting observation: Since I've started consciously making an effort to take and post selfies occasionally people reply to those posts with compliments. The ability to "like" that reply post is a really nice subtle way of sayinga polite "thanks" without having to up it into a foofaraw of an actual third reply post
KartikPrabhu, craftycorvid, nloadholtes, tmro, fkooman, eschnou and davidpeach joined the channel
KevinMarks, friedcell, Sebastien-L and interactivist joined the channel
elf-pavlik, adactio, stream7, tvn, friedcell, KevinMarks, petermolnar, Pierre-O, evalica, yakker, interactivist, nloadholtes and LanceyWork joined the channel
frzn and friedcell joined the channel
# 12:39 frzn good morning, GWG
n8fr8 joined the channel
# 12:45 GWG I'm trying to pick a fun name for a Twitter test account.
# 12:47 GWG I want to fix my Twitter POSSE without spamming.
glennjones joined the channel
# 12:47 GWG And develop ny long delayed Bridgy for WordPress plug in
elf-pavlik and tvn joined the channel
# 12:59 rhiaro I'd just spam twitter and delete them afterwards, but I don't care about my followers ;)
# 13:00 rhiaro if people are getting notifications for individual tweets, they deserve it
mlncn joined the channel
KartikPrabhu joined the channel
Deledrius joined the channel
snarfed joined the channel
fourtonfish joined the channel
nloadholtes joined the channel
Guest99613 joined the channel
# 14:08 tommorris "C'est á Portland, repère de hipsters américains au Nord-Ouest des Etat-Unis que le projet Indie Web naît en 2011." bwahaha.
# 14:09 tommorris KartikPrabhu: hey, it's nobody's error - you can generate new DH keypairs from the certificate.
# 14:12 Guest99613 yeah I just went through generating new moduli files on every machine I SSH into. it was less than ideal.
# 14:13 Guest99613 wel, it won't let me change nick. I'll restart this client. - endi
endi_ and mlncn joined the channel
# 14:26 endi_ strange the server bugs out and I go into reconnect loop if I try to use 'endi' as nick
chalettu, jansauer, shiflett, gRegorLove, frzn, stream7 and tilgovi joined the channel
# 16:22 bear the logjam issue is solvable by generating a dhparam file that is > 1024 bits and having your web server use it
jansauer and KevinMarks__ joined the channel
# 16:26 kylewm iirc KartikPrabhu is on Django-specific PaaS hosting?
# 16:31 bear it's whoever is in charge of the web server (nginx/apache)
# 16:31 bear apps or proxied apps have no control over it (well, unless they are doing the ssl termination)
snarfed, interactivist, tilgovi, afrogeek1, afrogeek, wolftune, benwerd, tantek and eschnou joined the channel
# 17:54 tantek aaronpk - interesting - does that update mean the entire photo property is replaced at that post with the new (single) value?
# 17:55 aaronpk basically i took the parsed microformats json of the edit example we made, and trimmed out anything that could be implied by the micropub endpoint
# 17:55 bret that would update the photo url right?
# 17:55 aaronpk basically the same process as coming up with the original form-encoded micropub request for creating things as documented on /micropub
# 17:56 aaronpk technically the "properties" object isn't needed, but that makes it match the parsed JSON
# 17:56 tantek aaronpk: any chance we can develop the form encoded version first? simpler design first? and then go all trendy JSON?
# 17:57 aaronpk kiiinda... except once you require nested properties, form-encoding becomes ugly/hacky
# 17:57 tantek I like this design process for such protocols: HTML+mf -> form encoded -> trendy JSON
# 17:57 tantek that way we can swap out the trendy next year for whatever
# 17:57 aaronpk we don't need nesting for *creating* things because we can just assume every property specified in the create is part of the tthing being created
# 17:57 tantek aaronpk I'm convinced we can figure out flat & simple for form encoded
# 17:57 aaronpk but for editing things, we first need to specify the URL of the thing being edited, and then also need to specify the changes
# 17:58 aaronpk that's what led to the "mp-" prefix spec'd out earlier
# 17:58 bret aaronpk: how do we handle properties that are deleted?
# 17:58 bret or is it a coplete overwrite rather than an inherent type merge
nloadholtes joined the channel
# 17:59 aaronpk the problem with the mp- prefix is it's pretty arbitrary
# 17:59 tantek aaronpk: nah, it follows a decent prefix naming pattern
gavinc joined the channel
# 18:00 tantek when it comes down to it, all vocabularies are pretty arbitrary
# 18:00 tantek only diff is if you can point to your methodology for coming up with them
# 18:02 aaronpk deleting values and properties starts to get hairy tho
# 18:05 aaronpk being able to use a top-level property to specify the difference is nice
# 18:06 aaronpk that would look like add[category]=indieweb in form-encoding
# 18:07 aaronpk except that form-encoding itself doesn't have the concept of nested properties, but most parsers have implemented it with [] despite it not being in the spec
# 18:12 aaronpk and because it's not in the spec, you get the difference between languages of how they handle things like foo=1&foo=2 vs foo[]=1&foo[]=2
# 18:12 tantek you could add the photo update example there too
tilgovi joined the channel
interactivist joined the channel
# 18:22 bret but there are good ones out there that do both
# 18:23 aaronpk bret: creating the form-encoded post should be part of any HTTP library you're using.. node doesn't do that?
# 18:26 tantek the burden of proof is on inclusion of extra text
# 18:26 tantek also "list" implies an ordering which I don't think you want
# 18:27 tantek if that's what you're trying to say, yet there is no need for a new term/concept "list" to express that
# 18:28 tantek hence you can omit "to a list" and nothing is lost.
# 18:29 tantek example: the term "collection" is often used in various Social Web WG discussions, and when you use a specific term "list", then you incite debates about "list" vs "collection" etc.
# 18:29 tantek it's like honey to the architecture astronauts
# 18:30 tantek there is no need to call it anything - just "add a value" is fine
# 18:30 aaronpk i'm trying to explicitly say it doesn't affect anything else in that... array/list/whatev
# 18:32 aaronpk if i can refer to everything as "properties have one or more values" then I can avoid terms like "array" and "list"
# 18:32 tantek yes, and "If the property does not exist already, it is created."
# 18:33 tantek similarly, for removing, "If no values remain, the property is removed."
# 18:33 tantek (because that's what'll happen in the HTML result)
# 18:34 tantek aaronpk: would you be up for keeping the form-encoded version side-by-side with these?
# 18:34 tantek I still think it's better to keep the form-encoded design up to date first, then have the JSON follow from that
# 18:35 tantek to continue encouraging the HTML+mf -> form-encoded -> trendy format (JSON, XML) etc.
yakker and snarfed joined the channel
indie-visitor joined the channel
# 18:45 Loqi Welcome, indie-visitor! Set your nickname by typing /nick yourname
# 18:48 tantek sorry that was too much unfiltered noise - kicked/banned
# 19:32 bret aaronpk: sorry was pulled away from irc. no, node's http server does not automatically decode the body, and even express requires that you set that middle ware up as of today. requires libs to do the encoding and decoding. using the JSON.stringify or JSON.parse on the body of the req or res is easier in this case. but doing form encoded isnt hard either,
# 19:32 bret puts it in the funny looking form encoded on wire encoding until whatever on the other side turns it back into whatever native object the language uses
# 19:33 bret it felt hard at first when i wrote gitpub, but its actually pretty trivial as Im a better JS dev now
# 19:34 bret im tempted to actually accept form or json encoded micropub
# 19:35 tantek bret - where is that on your itch list priorities?
# 19:36 bret aaronpk: yeah that loosk great, even if it gets transmitted as form
# 19:37 bret tantek: im in itch gridlock right now :[
# 19:38 tantek bret - you can escape the gridlock by braindumping them to your User: page Itches section
# 19:39 tantek the reason braindumping them there (externalizing them from your brain) works, is because it's a distinctly differnt cognitive task to recognize/discover/document itches, than it is to prioritize them
# 19:40 tantek so by default just braindump them to the end of your existing list, until no more come to mind
# 19:40 aaronpk i've been using Wunderlist for that recently. although I think I need to make a separate list for "my site" vs "indieweb things"
# 19:40 tantek then take a mental break (walk, drink, eat, etc.) and come back and look at the whole list and see what "hits you"
# 19:41 bret been distracted by writing dev tooling stuff
# 19:41 tantek then move that to the top (acknowledging that realization), save, and work on it
# 19:41 tantek bret - it's ok to be distracted by other stuff, just add it all
# 19:41 tantek or if you need to separate "your site" vs. "other" you can keep multiple lists for that
# 19:43 bret aaronpk: im back on the taskwarrior train
# 19:44 tantek we have a bunch of big anniversaries coming up this year
# 19:44 tantek 2015-06-20 - 10 year anniversary of launch of microformats.org
# 19:46 tantek 2015-07-18 - 5 year anniversary of the first Federated Social Web Summit
# 19:46 tantek aaronpk we should kind of make a big deal of that
# 19:47 tantek especially assuming we get major things built / launched at IWC 2015 - which is the weekend before
# 19:47 aaronpk begs the question, what kind of major things would be good to build during IWC?
# 19:49 tantek that almost sounds like you're asking for a VISION or something ;)
# 19:49 benwerd I'm thinking about building flat-file storage for Known so you don't need a db at all
# 19:49 tantek running Known off a github io backend would be amazing
# 19:50 tantek nah, git bindings is a decent enough approach / portable etc. with a default of github io
# 19:51 aaronpk benwerd: you're thinking about making the Known app generate static HTML files that can be hosted somewhere?
# 19:51 tantek whereas the others are all snowflakes (Dropbox, Cozy, etc.)
# 19:51 Pierre-O benwerd: I have a quick question, can I host a version of Convoy (I believe it is based on known), and can I make a convoy-IH plugin?
# 19:51 aaronpk or using files in place of a database, where the PHP app still serves the requests but reads from disk instead of a DB
# 19:52 benwerd Pierre-O: No, but you can host the social media plugins, which are all on GitHub
# 19:52 Pierre-O I think I could reverse engineer the convoy plugin but well, no time for that, was just curious, thanks :)
# 19:53 benwerd Pierre-O: we have to leave a few ways for us to pay the rent / eat
# 19:53 benwerd But yes, alternative services are definitely possible
# 19:53 benwerd And we've made sure to open source all the actual functionality
# 19:56 bret benwerd: leveldb is nice because it defaults to a library use, vs service oriented client/server
# 19:57 rhiaro kylewm: I'm retiring my .so in September because registrar are awful and won't let me transfer away, so that's okay
# 19:58 kylewm I'm surprised so many new startups are relying on .ly for cute domain hacks ... seems very risky
# 20:00 GWG rhiaro: I'm tempted to say, .so what?
# 20:01 GWG I think I'll stick to .us . Reasonably priced and non-controversial.
# 20:02 kylewm actually I assumed optimizely was .ly but they are just optimizely.com
# 20:02 benwerd Pierre-O: We do license Convoy to universities and so on, so could do something like that with IH
# 20:03 GWG tantek: I'm on that page elsewhere, aren't I?
# 20:03 tantek GWG, don't know - if you have a .us domain that you use with your site - you should add it to the .us section
frzn joined the channel
# 20:09 aaronpk have you bought a .ro before? it was a pita for me
# 20:09 aaronpk it involved either 1) faxing a copy of my credit card and drivers' license or 2) wire transfer
# 20:10 aaronpk but it's a one-time registration fee, haven't had to renew it
# 20:11 bret i've read that .io money goes and funds some pretty awful stuff
# 20:11 GWG Unfortunately for me, .ke doesn't seem to do top level, so I can't get my last name.
snarfed joined the channel
# 20:14 aaronpk oh i remember the other weird thing about sending payment for the .ro
# 20:15 aaronpk they say that wire transfer fees for sending *and*receiving are paid by the customer
# 20:15 aaronpk but then they say they can't tell me how much their bank will charge to receive
# 20:15 aaronpk so I had to call their bank to ask if they charge a fee to receive USD, which they did not
Lancey and KartikPrabhu joined the channel
# 20:21 GWG Not sure what I'll do with it, but...
KartikPrabhu joined the channel
# 20:51 tantek anyone here intersted in an early Octoberish IWC Austin?
# 20:52 tantek I ask because there's an AEA (An Event Apart) Austin early October, and it may make sense logistically and in terms of crossover to have an IWC before/after
# 20:52 rhiaro ooh my favourite couchsurfer so far is in austin!
# 20:53 rhiaro (probably not financially viable though really)
# 20:53 gRegorLove I do love Austin, but probably won't be able to travel there in October
# 20:53 rhiaro there's little chance I'll be able to blag funding for TPAC, so I can't do two expensive trips in october
interactivist joined the channel
snarfed joined the channel
# 21:06 aaronpk yes, but trying to figure out the best venue, I want to do it on the east side this week
# 21:07 kylewm I would totally consider Austin in October. I have a friend who just moved to Corpus Christi
# 21:08 kylewm (who i would like to visit, not that those two places are particularly close)
# 21:11 tantek yes - Austin is a fun place and I've missed it since bailing on SXSW since 2013
# 21:14 kylewm does React really render UI components in JS Canvas?
# 21:17 bear react native may, but react renders straight to it's virtual dom
# 21:17 bret that is used for mvc stuff for actually creating the real dom
# 21:17 tantek goes to manually update his upcoming events box on his index.html
# 21:19 snarfed virtual canvas shadow dom layout rendering mwahahahahaha!!!
# 21:19 bret one insight i had... whats the alternative to virtualdom (eg, representing your dom as a js object)? not having one and tying all logic to the real dom api
# 21:20 bret virtualdom gives you access to your dom, but with a different api
# 21:20 bear the reason they use virtualdom vs native is to control refresh/updates
# 21:20 bear when you touch the real dom you cannot control when updates happen
# 21:20 bear but with the vdom you can have active/inactive "pages" which don't force a dom update even if they are offscreen
# 21:21 tantek bear - not true - hence requestAnimationFrame
# 21:21 bear i'm talking about updating a div that may have 1000 sub elements and then the dom redrawing all of them
# 21:22 tantek bear - screen updates of any kind are animations of a sort - nevermind the name of the webapi
# 21:23 bret aparently there is some advantage with the vdom in terms of diffing between updates
# 21:23 tantek dang, squarespace templates are now apparently js;dr :P
# 21:25 kylewm bear: thanks for explaining that. it maps pretty well to Java Swing actually which I'm much more familiar with :P
# 21:31 tantek realizes he's only indie RSVPd to 1 of the 3 upcoming events he has listed, so still can't automate generation of his "Speaking & Events" box
# 21:32 tantek would anyone else be interested in Bridgy Publish RSVP POSSE to Lanyrd?
# 21:33 tantek goes to create another Bridgy Publish issue specifically for RSVP
# 21:41 tantek wonders if he should dare to ask the Lanyrd folks to "just" support receiving webmentions of Indie RSVPs.
# 21:42 tantek is there magic text to auto-link an issue # reference in github issue comments?
# 21:47 KevinMarks can we nag simon into adding rsvp support in lanyrd directly?
# 21:53 tantek anyone else here going to YxYY this year? (June 12-15)
# 21:54 KartikPrabhu tommorris: bear: thanks for the info on Logjam thingie. I am on a shared hosting so I suppose I should ask those guys to fix it
# 21:56 tantek if we use jargon here (#indiewebcamp) we should document them
# 21:56 tantek general idea is to keep this channel and its logs more approachable
# 21:56 tantek which means minimize jargon usage, and when necessary, make it so people can ask "what is"
# 21:57 KartikPrabhu and that my site is getting a B instead of A+ on SSL labs because of it
KevinMarks_ joined the channel
# 22:00 tantek hmmm - do any mf2 parsers support backcompat for h-event (i.e. vevent) ?
# 22:03 KartikPrabhu kylewm: do you know why hevent -> h-event added to mf2py when it should be vevent -> h-event?
# 22:04 KartikPrabhu tantek: oh mf2py says "ported and adapted from php-mf2" so likely php-mf2 also have hevent
j12t joined the channel
XgF joined the channel
# 22:12 tantek KartikPrabhu: while you're in there - can you check the other backcompat support classes?
# 22:34 KartikPrabhu I could for starters just duplicate the hreview backcompat to hreview-aggregate but both of those are lacking in the recommended "item" property translation
# 22:34 kylewm KartikPrabhu: no it doesn't :/ and the transformation for 'fn' looks totally wrong
# 22:35 kylewm could add a hack fix for just this specific case in the code (instead of in the big block of X -> Y translations)
# 22:36 kylewm or represent it generally like {'item': {'vcard': ['p-item', 'h-card'], etc. }}
# 22:36 kylewm KartikPrabhu: if it's literally the only case where we have to do this kind of conditional behavior, I think it's fine to hack it in. when there are two or three such cases, it'll be worth it to generalize
# 22:37 KartikPrabhu hmm. i think the "item" thing in hreview and hreview-aggregate is the only case so far
# 22:38 KartikPrabhu ok kylewm my mf2py repo is updated with the hevent -> vevent fix. feel free to pull from that and add these hacks :)
# 22:43 kylewm um, I think I ruined your world with aggressive rebasing last year
# 22:49 GWG zachdonovan has a conflict on the 17th. I have one on the 3rd
# 22:50 kylewm I hate to ask, but can anybody help me and Kartik get back in sync on github? does he need to "git reset --hard orign/master"?
# 22:51 KartikPrabhu kylewm: i think all the code is synced correctly but there are these weird commit things floaitng around
# 22:55 gRegorLove kylewm, KartikPrabhu syncing always confuses me and I have to look it up each time, heh.
# 22:57 gRegorLove I think the method I've settled on is git checkout upstream/master -b mydevbranch
# 22:59 kylewm gRegorLove: do you run git fetch upstream first?
# 23:00 gRegorLove I think it's not necessary. I think it creates a new branch from the upstream HEAD
# 23:01 kylewm the problem is, all our files are the same, but the commit hashes are all different
# 23:02 kylewm because i split up one giant pull request into a bunch of smaller requests (that was the aggressive rebasing)
# 23:02 gRegorLove Might be easiest for one of you to create a new branch off that rebased branch and re-apply your changes there.
gavinc and nloadholtes joined the channel
# 23:07 gRegorLove I'm using the "PS" there when contributing to php-mf2, for example.
# 23:08 gRegorLove delete the new branch after each PR, the forked master can easily be synced from upstream then
wilfredh joined the channel
# 23:14 kylewm KartikPrabhu: I think #B from gRegorLove's link would work nicely
# 23:15 kylewm put your current master into a new branch called "backup", then move the "master" branch to upstream/master
tantek joined the channel
j12t joined the channel
j12t_, KevinMarks_ and KevinMarks__ joined the channel