#dev 2016-12-19

2016-12-19 UTC
rMdes- joined the channel
#
www.boffosocko.com edited /Post_Kinds_Plugin (+554) "Archive pages and RSS feeds" (view diff)
KevinMarks and KevinMarks_ joined the channel
#
aaronpk what is a geo URI?
#
Loqi It looks like we don't have a page for "geo URI" yet. Would you like to create it?
#
aaronpk a geo URI is a URI scheme for describing the latitude and longitude of a specific location, and is used by some Micropub clients as an easy way to specify the location of a post. e.g., <code>geo:45.51533,-122.64653</code>
#
loqi.me created /geo_URI (+251) "prompted by aaronpk and dfn added by aaronpk" (view diff)
#
Loqi ok
#
aaronpk geo URI << http://tools.ietf.org/html/rfc5870
#
loqi.me edited /geo_URI (+54) "/* See Also */ new section" (view diff)
#
Loqi ok, I added "http://tools.ietf.org/html/rfc5870" to the "See Also" section of /geo URI
KevinMarks, kline and gRegorLove joined the channel
#
gregorlove.com edited /StartSSL (+272) "note no longer trusted by Mozilla, Google, Apple" (view diff)
#
aaronparecki.com edited /StartSSL (+4) "stopped trusting *new* certificates. previously issued certificates continue to be trusted." (view diff)
KevinMarks and cweiske joined the channel
#
petermolnar my font size gets f* up in print
#
petermolnar and using inline @media print {} doesn't seem to address it
#
Zegnat That's odd, font sizes should definitely work in print. Is it a specific browser that "fs up"?
#
cweiske you've got a special print stylesheet as I see
#
cweiske https://petermolnar.net/wp-content/themes/petermolnar.eu/print.css
#
petermolnar ( I'm getting rid of less and rewriting my complete css as the bare minimum )
#
petermolnar that is the current site, the trouble is with the one I'm testing :)
#
Zegnat Do you specify different font sizes for browser and print then, petermolnar?
#
petermolnar I'm trying to switch to rem units and specifying font size in html {} or body {}
#
petermolnar so far, it's more or less ok for screen
#
petermolnar but print results in GIGANTINC fonts :)
#
petermolnar I'll pastebin my stuff
#
Zegnat I haven't heard that one before, interesting.
#
KartikPrabhu for print you might want to specif font sizes in pt
#
petermolnar I do
#
petermolnar it's ignored
#
KartikPrabhu that is strange
#
petermolnar http://pastebin.com/WvRTStT0
#
petermolnar full html, w/ inline CSS
#
petermolnar the images won't show though
#
Zegnat That font styling definitely feels intuitive.
#
cweiske chrome shows the print normal - http://p.cweiske.de/388/rev-raw/e1195fa92f9fc19655f86eda94ca76a9fd7e57f5/phork0.html
#
cweiske firefox too
#
petermolnar that's unexpected
#
petermolnar chrome is indeed fine
#
petermolnar ff is still not in my case
#
cweiske which firefox?
#
Zegnat Firefox looks fine too for me
#
cweiske 50.1.0 here
#
petermolnar FF 50.0
#
petermolnar hm
#
petermolnar let me upgrade that
#
cweiske did you try ff with my paste url?
#
petermolnar yes
#
petermolnar ( and I also spotted I need to fix max line width in print )
#
Zegnat I am on 50.0.2 and it looks good in the print preview
#
cweiske screenshot: http://storage3.static.itmages.com/i/16/1219/h_1482143930_2387831_962b3a8b7d.png
#
Zegnat screenshot Firefox 50.0.2 on W10: http://imgur.com/a/kUYVU
#
petermolnar oh, ffs
#
petermolnar PEBKAC
#
petermolnar there is a Scale: in FF print view
#
petermolnar which, for me, for an unknown reason, was set to Custom...
#
petermolnar thanks for the help
#
Zegnat You're welcome. It would've been nice if all problems could be solved with this little help ;)
#
petermolnar right, from now, the only thing I need to figure out, is how to have sharp and crips fonts in prints
j4y_funabashi joined the channel
#
petermolnar whee, we are back in the era of css hacks!
#
petermolnar transform: rotate(0.01deg); - this, apparently, resolves in crispy svg, unlike the one without...
#
petermolnar I'm trying to get rid of external fonts and adding inline svg symbols, referenced for icons in the code
#
petermolnar turns out either chrome or ff will render it blurry when you're making it em/rem sized
#
petermolnar unless you do that rotate trick...
#
petermolnar eh, only solves
#
petermolnar firefox
#
petermolnar not chrome
#
petermolnar ok... so that's width: 1rem; height: 16px; rotate: (0.01deg) and now it's crisp on both ff and chrom
#
petermolnar I feel like it's 2006 again
#
petermolnar o.O http://opensource.perlig.de/rcssmin/#trivia-fun
#
sknebel does the rotate force it to use a different rendering mechanism or what is going on? any theory?
#
petermolnar I'd guess it forces sw rendering
#
petermolnar but that's a guess
#
sknebel hm... I would have expected that to be always active for printing
#
petermolnar oh, this is on screen
#
petermolnar not just for printing
#
sknebel did you check actual prints (prints to PDF) or the preview?
#
sknebel oh, ok
#
petermolnar so: inline css, inline svg icons, inline everything, except images: min 29k, max 142k HTML file size ( the 142k contains a _lot_ of configuration text )
#
petermolnar getting ipfs ready :)
indiescripter_ joined the channel
#
petermolnar how do mf2 parsers react to base64 inline images for author images?
#
martymcguire[m] hahaha
#
martymcguire[m] http://reductress.com/post/wow-this-woman-3d-printed-her-own-menstrual-cup-and-it-was-a-fucking-nightmare/
#
Loqi hehe
#
martymcguire[m] dang it again. wrong matrix channel. sorry, folks. ;[
#
martymcguire[m] i wish the riot.im client had a way to color the chat window based on the room. it is far too easy to forget which room i had focus on last.
#
martymcguire[m] oh look, they do. okay! hopefully this will help stop me pasting non-indieweb links in indieweb rooms.
#
aaronpk petermolnar: i would imagine the base64 data would come through as the value of the property
#
Zegnat A data uri sounds just as valid as an http uri for u-photo, imho, petermolnar. I don’t think the problem will be with parsers, there might be a problem for consumers of the parser output
#
ben_thatmustbeme may overflow some fields
#
Zegnat Well, previous HTTP1.1 standard allowed URLs of unlimited length, right? So those fields could overflow for normal image URLs too, ben_thatmustbeme ;)
#
ben_thatmustbeme not sure, realistically i don't see anyone making URLs longer than what browsers support
#
sknebel pretty sure the major browser don't really have a limit, or if they do it's really large (>50k). search engines and other non-browser clients are probably more worrysome
#
sknebel ... of course there is a stack overflow question for that: http://stackoverflow.com/questions/417142/what-is-the-maximum-length-of-a-url-in-different-browsers
#
Zegnat If you have dropped old IE support, according to this research, that would still put URLs on over 4000 chars.
#
Zegnat See, sknebel’s link, didn’t need to link it again
#
Zegnat Measured IE9+ to 4043
#
ben_thatmustbeme oh wow
#
ben_thatmustbeme much larger than i thought it would be
#
ben_thatmustbeme it used to be only 2000 or so
#
j4y_funabashi /quit
#
aaronpk got sidetracked working on Quill UI improvements
#
aaronpk adding reply context preview to quill \o/
#
ben_thatmustbeme cool
tantek joined the channel
#
ben_thatmustbeme i want to start adding weight tracking to inkstone and my site
#
aaronpk nice
#
ben_thatmustbeme the photo posting is working in inkstone now (with media endpoints at least)
#
aaronpk woo
#
ben_thatmustbeme but it doesn't work on mobile
#
ben_thatmustbeme i suspect because its timing out
#
aaronpk i'm also making Quill look less like a test utility now that micropub.rocks exists
KartikPrabhu joined the channel
#
ben_thatmustbeme dang, failed to parse https://ben.thatmustbe.me/note/2016/12/19/2/
#
Loqi [Ben Roberts] testing out hubzilla's permalink parsability
#
ben_thatmustbeme probably need some PR to update hubzilla / friendica mf2 markup. but as i remember they layout sort of prevented it
#
tantek ben_thatmustbeme: when you track down which object or property that had trouble with the layout - I'm definitely interested in analyzing that
#
Loqi tantek: gRegorLove left you a message 23 hours, 19 minutes ago: Re archiving planning notes I added an "Initial Planning" subhead to the IWC page, like https://indieweb.org/2016/LA/Planning#Initial_Planning
#
Loqi tantek: gRegorLove left you a message 23 hours, 19 minutes ago: Linked from https://indieweb.org/Planning#Completed
#
tantek whoa
#
ben_thatmustbeme tantek: its that my software only looks at the first h-* on the page, the permalinks in hubzilla / friendica have an h-card first, then the h-entry
#
tantek oh yeah that makes sense
#
tantek !tell gRegorLove not seeing the link to LA/Planning from /Planning#Completed but that approach sounds good
#
Loqi Ok, I'll tell them that when I see them next
#
tantek ben_thatmustbeme: ah as a method of determining what "is" the page?
#
aaronpk ben_thatmustbeme: one trick i've found helpful for that is looking for an h-* that has a url property that matches the URL I fetched
#
tantek ben_thatmustbeme: is it possible to put the class=h-entry on the <body> of their post permalinks?
#
ben_thatmustbeme yeah, i think i'm going to have to do that
#
ben_thatmustbeme tantek: as i remember it is not, at least not with some deeper changes that I would not want to fiddle with
#
ben_thatmustbeme it had a fairly stick idea of "whatever is part of the post, goes in this box" but the user is always shown on the left bar, which is marked up an an h-card
#
ben_thatmustbeme so breaking out of that would take some new functionality
#
ben_thatmustbeme not the type of thing i would want to do unless i really got involved deeper in the project or got help from a developer
#
tantek are there any divs or such between body and that sidebar that you could put the class="h-entry" on?
#
tantek.com edited /Template:IndieWebCamp (+24) "2016 IWCs done, 2017 is being planned!" (view diff)
KevinMarks joined the channel
#
tantek.com edited /Events (+2535) "/* Brainstorming */ add problem statement, replacement requirements stubs" (view diff)
KevinMarks, rMdes- and KevinMarks_ joined the channel
#
loqi.me created /safety (+26) "prompted by tantek and dfn added by tantek" (view diff)
#
loqi.me created /marked_as_safe (+24) "prompted by tantek and dfn added by tantek" (view diff)
#
loqi.me created /disaster (+25) "prompted by tantek and dfn added by tantek" (view diff)
KevinMarks joined the channel
#
KartikPrabhu any one with experience with encodeURI method in JS? It seems FF and Chrome do it differently with is causing fragmantion troubles!
#
KartikPrabhu fragmention*
#
tantek oh dear
#
tantek what do the docs say?
#
KartikPrabhu MDN docs don't seem to say anything about this difference
#
tantek but does it say what *should* happen given whatever input you're giving the function?
#
KartikPrabhu yes, special characters should be UTF-8 encoded
#
KartikPrabhu will have to do more stringent testing to deduce exactly what is going wrong
#
KartikPrabhu I was wondeirng if someone had come across this already
#
tantek do different browsers have different support of which characters are "special"?
#
KartikPrabhu I thought "no". doing some console debugging now
#
KartikPrabhu OK seems that FF replaces non-break space character with a %20 which is a normal space, but Chrome uses %C2%A0 which is... something else?
#
KartikPrabhu I can't tell if this "replacing non-break space with space" is happening before encodeURI though
#
Zegnat 00A0 is non-break space. So I guess %C2 means 00?
#
KartikPrabhu so chrome is double encoding it?
#
bear does that depend on what the content-type is of the page? I know with forms, the encoding used changes with the content-type of the page where the form lives
#
KartikPrabhu i am using the same file on both FF and Chrome to test this
#
KartikPrabhu with encoding utf-8 declared in the head
#
tantek sounds like one or the other has a bug, or the spec is ambiguous
#
tantek considers asking "dumb" questions in #whatwg to see what happens
#
tantek KartikPrabhu: did you already check Bugzilla for encodeURI bugs?
#
KartikPrabhu looking, didn't know that existed
#
tantek wait what
#
tantek looking where?
#
KartikPrabhu on Bugzilla
#
tantek where Firefox bugs are reported etc.
#
KartikPrabhu no search results for encodeURI
#
tantek ok that's a good start - thanks
#
tantek KartikPrabhu: for some dry reading: https://tc39.github.io/ecma262/#sec-encodeuri-uri
#
KartikPrabhu yup, that is what MDN told me too. I am simply confused as to why FF and Chrome are doing different things
#
tantek KartikPrabhu: check https://github.com/tc39/test262 for encodeURI tests
#
tantek and see if there is one for what you're finding
#
KartikPrabhu in fact fragmention.js also replaces all space characters with normal spaces. So on FF (which converts non-break spaces to normal space) fragmention.js works; but on Chrome (which encodes to UTF-8) fragmention.js does not work
#
tantek KartikPrabhu: is the whitespace normalization something your script does or you depend on the browser for?
#
KartikPrabhu fragmention.js does it explicitly. My script generates a fragmention from some given text and uses encodeURI to encode it. FF and Chrome differ on the encodeURI bit
#
tantek and you're sure that in both cases the input to encodeURI is the same?
#
KartikPrabhu yes, also checked that
#
KartikPrabhu this was also reported by adactio: https://adactio.com/journal/11632#confuses%20the%20fragmention
#
Loqi [Jeremy Keith] Deep linking with fragmentions
#
KartikPrabhu in fact that is how I came up on it
#
KartikPrabhu tantek: I can upload a demo which generates the console logs
#
tantek KartikPrabhu: we can start with a js snippet that does an alert of an encodeURI string that produces a result different in FF vs Chrome
#
KartikPrabhu ok will try to condense into that
#
tantek then we can argue what *should* be the right output per spec, or that the spec doesn't say (spec bug)
#
tantek and then what *should* be the right output per good design / compat
#
KartikPrabhu yup, got it
#
tantek then edit spec if necessary and file bug against whichever browser(s) do it differently
#
tantek that's kind of how this stuff gets resolved
#
tantek so let's start with that javascript: one-liner - always a good way to demonstrate web (in)compat
#
KartikPrabhu OK wow! it does seem that the problem is in the input to encodeURI...
#
KartikPrabhu somehow the text obtained from a selection is messing up
#
tantek so the selection API is producing different results?
#
KartikPrabhu writing test to check that
#
KartikPrabhu when used the text from some HTML elment both FF and Chrome gave the same answer
#
KartikPrabhu tantek: here is the demo https://kartikprabhu.com/static/demo/selection-encodeURI.html