#dev 2017-07-31

2017-07-31 UTC
[miklb], eli_oat, deep-book-gk, AngeloGladding, tantek, gRegorLove, GK1wmSU, cweiske and barpthewire joined the channel
#
voxpelli First time I've seen Webmentions criticized as to "mainly serve advertisers and spammers" https://github.com/jcs/lobsters/issues/383#issuecomment-318916502
#
Loqi [crertel] I'm not sure that having folks meandering over to Lobsters because somebody linked their blog is a great idea--especially if it's just to fuel the outrage machine. Additionally, I'm not sure how I feel about having yet another way of opening up our d...
#
tantek wat
#
tantek mmm lobsters
[kevinmarks] and tantek joined the channel
#
Zegnat voxpelli++ for your well balanced reply in that lobsters issue
#
Loqi voxpelli has 10 karma in this channel (107 overall)
#
voxpelli Zegnat: thanks :)
tantek joined the channel
#
Zegnat That moment when someone asks you if the HTML validates, and you have to tell them it does not because validators are broken. *sigh*
AngeloGladding joined the channel
#
tantek Zegnat: oh? what problem with validation?
#
tantek.com edited /webactions (+843) "/* How */ Wrap multiple silo post actions, without JS!" (view diff)
#
voxpelli tantek: you updating that reminds me – has a comparison to the Web Share API been made?
#
tantek that one is thanks to fastgood.cheap which requires JS to actually work (per https://twitter.com/philpennock/status/891734286864453632 via kevinmarks), but ironically includes tweet/share links that DO NOT require JS to function. amazing.
#
@philpennock Today I noticed that http://fastgood.cheap/ lets you select all three together as long as JavaScript is disabled. There's a moral here. (twitter.com/_/status/891734286864453632)
#
tantek voxpelli: any API would be complementary with webactions which are markup
#
voxpelli tantek: true, but I guess some deeper aspects would be shared with webactions being the declarative way of interacting with it and web share api possibly the programmatic way
#
voxpelli https://github.com/WICG/web-share-target eg. possibly being a way to register a share target
#
Loqi [WICG] web-share-target: Web API proposal for receiving shared data
#
voxpelli an interesting polyfill mechanism at least, less hacky than indieconfig :)
#
voxpelli haven't yet himself looked closely on it though
#
voxpelli huh, will apparently be active by default in Chrome on Android starting with next stable version (61)
#
tantek voxpelli: if you mean the Google Web Share API proposal, independent of any comparison, I think the Web Share API is a poor design, it's based more around enabling large silos, instead of arbitrary sites becoming share providers
#
tantek I'll have to take a longer look at it later - I've been so far disappointed by Google efforts in this area, I kinda wish the folks working on it would start with their own indieweb sites first
#
tantek voxpelli: what would be more interesting is a Web Extension (that could work in FF / Chrome / Safari / Edge) to support <indie-action> natively
#
tantek (without the indieconfig polyfills)
#
voxpelli tantek: looks like there is public support from both Firefox and Safari for Web Share
#
voxpelli https://www.chromestatus.com/feature/5668769141620736
#
voxpelli But agree that the key is to get indie sites in there, and https://github.com/WICG/web-share-target is a bit interesting there as it wants to integrate with native apps as wlel
#
Loqi [WICG] web-share-target: Web API proposal for receiving shared data
#
tantek sigh - "public support" is a big exaggeration for a link to a bugzilla bug
KevinMarks joined the channel
#
tantek that's outright misleading.
#
tantek if you read the bugzilla bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1312422
#
tantek you can see that that it's a "Let's see" and "maybe something for next year." at best
#
voxpelli yeah, Chromes self-enforced requirement of having public support from other browsers before shipping new features probably pushes them into such interpretations :/
#
tantek more like "public looking at it"
#
tantek which Firefox does with lots of technologies / proposals
#
voxpelli a bit premature of Chrome to ship it publicly in just a few weeks, I guess the other browser vendors could have some good feedback
#
tantek in speaking with annevk, at last June's Mozilla allhands, he was also skeptical / critical of it
#
voxpelli so in a way Web Intents all over again :/ a standard pushed without wide support
#
tantek even the Safari "public support" is just a link to a bug considering it: https://bugs.webkit.org/show_bug.cgi?id=171100 :P
#
tantek no, not even considering, just tracking it
#
tantek "This is interesting"
#
tantek well the bug name is " Consider implementing Web Share API"
#
tantek still, that's a far cry from "public support"
#
tantek so much for trusting Chrome status pages :P
#
tantek voxpelli: at least it's not as wide-ranging as web intents
#
voxpelli The "Intent to Ship" was somewhat more clear: https://groups.google.com/a/chromium.org/forum/#!searchin/blink-dev/web$20share/blink-dev/nwmTLZKFFIA/8bslfL3JAwAJ "Public support (expressed interest)"
KevinMarks and tantek joined the channel
#
calumryan.com edited /Planning (+27) "/* Planned */" (view diff)
#
kodfabrik.se edited /Planning (+26) "/* Planned */" (view diff)
#
kodfabrik.se edited /Planning (+15) "/* Brighton */" (view diff)
[kevinmarks] joined the channel
#
[kevinmarks] Licensing https://meshedinsights.com/2017/07/27/5-reasons-facebooks-react-license-was-a-mistake/amp/
#
cweiske the amp version of this page is really better to read
#
cweiske without all that tag cloud and calendar stuff on the right
KevinMarks, davidmead, KevinMarks_, [jeremycherfas], barpthewire and jeremycherfas joined the channel
#
Zegnat tantek, re HTML validation: WHATWG changed MAIN so it points at the main content of the current sectioning item (e.g. of an ARTICLE), but W3C says there can only be a single MAIN element on the entire page.
#
Zegnat I tend to follow WHATWG’s living HTML spec, which means HTML I produced now throws errors in validators that follow W3C.
#
cweiske that's what you get from using "living specs"
#
cweiske which I would not call "specs" at all
KevinMarks joined the channel
#
sknebel (I was under the impression that everything around sectioning/document outline is pretty academic anyways, since it has no implementations outside of validators?
#
sknebel )
#
Zegnat sknebel, well, some of them do imply “ARIA roles” for screenreaders
jeremycherfas joined the channel
#
voxpelli one could argue that "living specs" are the only specs that actually describe the real world – other specs are just a snapshot of a distant past
#
voxpelli I'm much +1 on the practice of living specs
#
cweiske livingspecs--
#
Loqi livingspecs has -1 karma
#
cweiske you can never say "implements $specname" because tomorrow it's can be broken
#
cweiske because the spec changed
#
cweiske "plays mp3 files"
#
cweiske erm, no, because we changed the mp3 format last week
#
cweiske sorry
#
voxpelli that's not true - living specs, unlike versioned specs, comes with the promise of backwards compatibility
#
cweiske [citation needed]
#
sknebel backwards compatibility is orthogonal to that
#
voxpelli https://wiki.whatwg.org/wiki/FAQ#Does_that_mean_the_specifications_can_change_at_any_time.3F
#
voxpelli some problem with versioned specs are: 1. they describe a reality that might not match the actual implemented one 2. they imply a possibility of introducing breaking changes
#
voxpelli living specs on the other hand describes what's actually implemented
#
cweiske "we are extremely careful!", and still 9 months later a baby was born
#
voxpelli if that's the case, then the baby would have been born anyhow, only difference is whether it would have been acknowledged or not
#
voxpelli (we don't need to compare HTML as a living standard to XHTML 2, right?)
#
voxpelli (or bring up how difficult the move from Python 2 to Python 3 was?)
#
cweiske "we fix the spec to match what the browsers do"
#
cweiske which means that I can implement the spec only to see that the two other implementers chose to implement it differently
#
cweiske and now my implementation is not following the spec
#
voxpelli cweiske: isn't that a rather sensible thing to do? to have specifications that's actually specify how things work rather than to describe some dream world that no one in practice actually will ever follow?
#
cweiske "living spec" means "never done"
#
voxpelli never and always
#
cweiske sign of the you-need-updates-everyday world view
#
cweiske which I disagree with
#
cweiske welcome, TCP/IP living spec
#
cweiske in which you cannot connect to the network tomorrow because the majority chose to "fix" the spec
#
voxpelli the IndieWeb specs are a great example of living specs – we dogfood, we document, we iterate some more, we document some more – the specs always reflect what's actually going on right now and is open to input from everyone
#
cweiske those are no finished spec
#
cweiske s
#
voxpelli [citation needed]
#
voxpelli have you actually seen a spec change in such a way such as the TCP/IP spec would not allow someone to connect?
#
cweiske I had to fix my indieauth-openidproxy lately because it relied on an implementation detail that was deemed unnecessary now
#
voxpelli as you said – those standards are not finished, so then maybe you just need to make a case for that feature to be needed
#
voxpelli but if no one implements it – then its also a case where it doesn't matter what the spec say, the reality will be the same anyway
#
voxpelli there's also this priority from the HTML spec world:
#
voxpelli > In case of conflict, consider users over authors over implementors over specifiers over theoretical purity.
#
cweiske .. which led to the huge how-to-handle-broken-html algorithm. no thanks.
#
voxpelli so XHTML 2 would have been the right move rather than to acknowledge and iterate on the current implementation?
#
voxpelli prioritize a good spec over something that's actually implemented and useful?
#
cweiske no browser did implement the broken-html-handling the same way, so you can't say it was actually implemented
#
cweiske today we have no library that is able to parse html5 correctly, apart from browsers
#
cweiske not a single one that implements this broken-html-algo
#
voxpelli cweiske: it wasn't implemented, because it wasn't acknowledged, HTML5/HTML Living Standard fixed that by acknowledging that broken behavior is something that browsers needs to handle and therefore something that a spec should deal with
#
voxpelli so now it is at least possible to implement such handling without reverse engineering four different browser engines
#
cweiske and there are exactly 4 organizations that have the resources to implement that
#
cweiske so who else uses living specs?
#
cweiske apart from whatwg
#
cweiske ?
#
voxpelli There are tools that claim that they do: https://www.npmjs.com/package/parse5
[kevinmarks] joined the channel
#
[kevinmarks] I'm pretty sure that html5lib implements the broken html spec
#
cweiske html5lib 0.999999999
#
cweiske hah
#
cweiske read the changelog
#
cweiske I guess they will have to add more 9 after the . in the next years
#
cweiske because there will never be a 1.0
#
cweiske https://pypi.python.org/pypi/html5lib
#
voxpelli cweiske: we at least stick to our opinions: https://github.com/pubsubhubbub/PubSubHubbub/issues/42 ;)
#
Loqi [aaronpk] #42 Make PuSH a "living" spec
singpolyma joined the channel
#
[kevinmarks] ttk.me/t4CH2
#
[kevinmarks] Hm, no preview, loqi?
#
[kevinmarks] http://tantek.com/2011/163/t2/microformats-specs-edit-button-schemaorg-foo
#
Loqi [Tantek Çelik] "Our [@microformats] specs have an edit button, theirs don't even have an edit history." - @KevinMarks #schemaorg #foo
#
voxpelli !tell mblaney fyi, just tested a new indie-config setup on a fresh Firefox on my new MacBook – worked great :)
#
Loqi Ok, I'll tell them that when I see them next
barpthewire, KevinMarks, brobertson, myfreeweb, [kevinmarks], eli_oat[m], zoglesby, TheGillies, plindner, schmarty, mindB, AlanPearce[m], eli_oat, gRegorLove, eli_oat1, [davidmead], [jeremycherfas], [miklb] and tantek joined the channel
#
jeremycherfas !tell petermolnar When you looked at the description of Grav’s honey pot field https://learn.getgrav.org/forms/forms/fields-available#the-honeypot-field did you get the impression that no further processing is actually necessary?
#
Loqi Ok, I'll tell them that when I see them next
#
jeremycherfas “The error prevents that form from being properly submitted.”
#
gregorlove.com edited /Planning (+78) "/* Brighton */ MozFest" (view diff)
#
jeremycherfas So maybe there’s no need to test anything?
j12t, dougbeal|iOS and eli_oat joined the channel
#
Zegnat jeremycherfas, if you use all of Gravs internal form functions then you wouldn’t need to do anything else, just include it
#
jeremycherfas That’s what I thought. But kind of hard to test …
#
jeremycherfas Maybe I’ll just let it out into the wild and see what happens. Tomorrow.
#
jeremycherfas Thanks Martijn
#
Zegnat The form should be in your HTML.
#
Zegnat So just use the browser inspector to make it visible, change its contents, and try to submit the form
#
Zegnat Should be able to do that locally :)
#
jeremycherfas How to fill in the field?
#
Zegnat If you can make it visible in the browser you can just write in it like in any other form field
#
jeremycherfas You can’t. It is a very special kind of field that is always hidden, always empty and throws an error if filled.
#
jeremycherfas I’m just going to let it loose, I think.
#
Zegnat It should still be in the HTML as a field, for bots to find, so you should be able to make it visible in the browser as well
#
Zegnat But should be fine to “let loose” ;)
#
jeremycherfas Hhhm. I haven’t looked to see whether it is hidden by CSS. Running out of time today. :)
tantek and eli_oat1 joined the channel
#
sknebel aaronpk: getting gateway timeouts on indieauth.com
#
sknebel oh, github is down
#
sknebel that might be it (happened while trying to login via github)
#
aaronpk whoa
#
Zegnat sknebel, get on indieauth already ;)
#
sknebel selfauth :P
#
Zegnat Yeah, if you don’t want to write your own – uhm – Python? – implementation, go with selfauth.
#
Zegnat I should push my libsodium fork public at some point this week, hmm.
#
Zegnat libsodium-using fork of selfauth, that is
#
Loqi yea
#
sebsel lol at Githubs fail-unicorn https://seblog.nl/temp/media-endpoint/e41f8c-github-down.jpg
calumryan, KevinMarks, j12t, KartikPrabhu, KevinMarks_, [kevinmarks] and eli_oat joined the channel
#
tantek.com edited /Planning (+546) "/* 2017 */ sort by number of confirmed co-organizers, and put Brighton at end of 2017 possible previous sites due to "not happening any time soon", add possible dates for NYC" (view diff)
#
aaronpk github is down and everyone briefly gets excited about decentralizing their git workflows
#
tantek and hopefully makes incremental progress then captures it in a more persistent state than email/IRC/twitter
#
tantek I wonder how many orgs tried to create an issue on github for the problem of them depending on github
#
aaronpk lol
#
tantek aaronpk, feel free to add something: https://www.w3.org/wiki/Github :)
#
aaronpk makes me want to work on indieweb issue tracking with webmention comments
#
aaronpk i *think* it's not a huge stretch from what we have working now
#
tantek yup
#
tantek what is an issue?
#
Loqi issue is a special kind of article post that is a reply to typically some source code, though potentially anything at a source control repository https://indieweb.org/issue
#
tantek and be sure to blog about it *while* github is down, so you can get all the HN kids on it
#
tantek even if all you do is blog your progress to date, and just a bit of forward thinking brainstorming
#
loqi.me created /commentpara (+27) "prompted by tantek and dfn added by tantek" (view diff)
#
loqi.me created /commentparade (+27) "prompted by sebsel and dfn added by sebsel" (view diff)
#
aaronparecki.com edited /reply (+1103) "/* Issues */" (view diff)
#
aaronparecki.com edited /comments (+7) "/* Issues */ rename header" (view diff)
KevinMarks joined the channel
#
tantek.com edited /reply (+361) "/* replies showing up on your main website */ who? citations? contrary prior art, note also Twitter profiles do not show replies by default (in separate view)" (view diff)
#
loqi.me created /sticker (+165) "prompted by sebsel and dfn added by sebsel" (view diff)
eli_oat joined the channel
#
loqi.me edited /sticker (+35) "/* See Also */ new section" (view diff)
#
loqi.me created /FB_Messenger (+31) "prompted by tantek and dfn added by tantek" (view diff)
#
loqi.me edited /Facebook (+21) "tantek added "Facebook Messenger" to "See Also"" (view diff)
#
seblog.nl edited /sticker (+1069) "Added Swarm and Facebook, and one IndieWeb example" (view diff)
#
seblog.nl edited /sticker (+210) "/* Swarm */ add screenshot" (view diff)
#
seblog.nl edited /sticker (+379) "/* Facebook */" (view diff)
#
sebsel All the screenshots.
#
loqi.me created /sticker_reply (+20) "prompted by sebsel and dfn added by sebsel" (view diff)
#
loqi.me created /Facebook_Messenger (+87) "prompted by sebsel and dfn added by sebsel" (view diff)
[jeremycherfas], eli_oat and eli_oat1 joined the channel
#
seblog.nl edited /sticker (+348) "note that stickers often have limited availability" (view diff)
eli_oat joined the channel
#
seblog.nl edited /photo_reply (+14) "/* See Also */ add sticker" (view diff)
eli_oat1 joined the channel
#
@rubygems jekyll-webmention_io (2.6.1): This Gem includes a suite of tools for managing webmentions in Jekyll: * Tags -… https://rubygems.org/gems/jekyll-webmention_io (twitter.com/_/status/892128052846657540)
davidmead joined the channel
#
aaronparecki.com edited /hwcpdx (+68) (view diff)
[chrisaldrich] and KartikPrabhu joined the channel
#
eli_oat1 Potentially stupid PHP question, but I can't figure it out for the life of me. I'm attempting to parse some XML and assign the content of various tags to variables. I, however, cannot figure out how to do this.
#
eli_oat1 Example:
#
eli_oat1 <Inventory>
#
eli_oat1 <InventoryItem>
#
eli_oat1 <Name>Banana</Name>
#
eli_oat1 <ID>10</ID>
#
eli_oat1 <Value>
#
eli_oat1 <Text>
#
eli_oat1 …
#
eli_oat1 </Text>
#
eli_oat1 </Value>
#
eli_oat1 </InventoryItem>
#
eli_oat1 <InventoryItem>
#
eli_oat1 <Name>Banana</Name>
#
eli_oat1 <ID>10</ID>
#
eli_oat1 <Value>
#
eli_oat1 <Text>
#
eli_oat1 …
#
eli_oat1 Potentially stupid PHP question, but I can't figure it out for the life of me. I'm attempting to parse some XML and assign the content of various tags to variables. I, however, cannot figure out how to do this.
eli_oat joined the channel
#
Zegnat eli_oat you probably want to look at DOMDocument, which can parse XML. You can then walk the tree, or use methods like “getelementbytagname”, or use XPath to find the right element. Once you have the element’s object you can set its value to whatever you like.
#
Zegnat (I would not use simplexml, it sounds simple, but you will find it coming back to haunt you at some point in the future)
#
eli_oat awesome! Thanks, I've been using simplexml, which may be part of my issue!
#
Zegnat I have always found simplexml to become part of my issues ;)
#
eli_oat (also sorry if I sent my initial message twice. My client got grumpy when I tried to send xml)
#
Zegnat I have to get to bed, but I am happy to help later if you get stuck.
#
eli_oat Thanks so much!
#
eli_oat Have a good night
#
eli_oat hopefully you've set me down the right path
KevinMarks joined the channel
#
@nhoizey @AaronGustafson Strange, `bundle outdated` tells me there’s a 2.6.1, but `bundle update --source jekyll-webmention_… https://twitter.com/i/web/status/892157525898010628 (twitter.com/_/status/892157525898010628)
#
@nhoizey @AaronGustafson I have gem 'jekyll-webmention_io' in my Gemfile. (twitter.com/_/status/892157429399597056)
#
@nhoizey @AaronGustafson `bundle install jekyll-webmention_io` gives me version 2.0.1 instead of 2.6.1, how can I get the latest? (twitter.com/_/status/892155618441797633)
dougbeal|mb1 and KevinMarks joined the channel
#
@nhoizey @AaronGustafson It looks like jekyll-webmention_io depends on json 2.0, while algoliasearch-jekyll depends on json… https://twitter.com/i/web/status/892161953937936385 (twitter.com/_/status/892161953937936385)
singpolyma and leg joined the channel