#dev 2017-12-02

2017-12-02 UTC
#
gRegorLove
GWG: quite a few bug fixes and improved backcompat parsing, since May 27: https://github.com/indieweb/php-mf2/pulls?q=is%3Apr+is%3Aclosed
#
gRegorLove
Oops, backcompat is already in.
[kevinmarks], dougbeal|mb1 and tantek joined the channel
renem, KartikPrabhu, tantek, [miklb], j12t_, [jeremycherfas] and [kevinmarks] joined the channel
#
loqi.me
edited /static_site (+97) "[kevinmarks] added "minimal python static site generator https://l4ur3n.github.io/blog/simple-blogging-engine.html" to "See Also""
(view diff)
KartikPrabhu and gRegorLove joined the channel
#
islandinthenet.com
created /User:Islandinthenet.com (+422) "khurt williams"
(view diff)
John__ joined the channel
#
snarfed.org
edited /Atom (+242) "RFCs"
(view diff)
snarfed joined the channel
#
snarfed
hey aaronpk, just checking, you're not parsing mf2 out of granary's atom content, are you?
#
snarfed
probably not, since there's not much there...but if so, definitely ask it for mf2 explicitly instead, with format=mf2-json
#
aaronpk
no I was planning on parsing the HTML version
aaronpk_ joined the channel
#
Zegnat
Does anyone know if you can subscribe to a tag on a GitHub issue tracker?
snarfed joined the channel
#
snarfed
er yeah, i meant s/atom/html/
#
snarfed
definitely try the json mf2 first, it'll be a bit cleaner
#
aaronpk
oh are you generating that separately rather than parsing the html?
#
snarfed
yeah, the html is actually generated from the json
#
aaronpk
ironically I don't have any code to consume the mf2 json directly, right now everything assumes it'll be parsing the HTML
#
snarfed
oh wow ok
#
aaronpk
I mean it shouldn't be hard to bypass the parsing step to use the JSON directly, but kind of funny
#
snarfed
heh very
#
aaronpk
actually wait...
#
aaronpk
if i'm consuming this JSON directly, I have to have a lot more error checking since I can't be sure it conforms to the mf2 json structure
#
aaronpk
whereas if I'm using the mf2 parser I know that the resulting JSON will always follow the structure output by the parser
#
aaronpk
that p-location doesn't belong in "type" and that would never have been created by the mf2 parser
#
snarfed
plus any big deviations btw granary's json and html are arguably all bugs
#
snarfed
so go ahead w/html, and i'll look at the bugs you find
#
GWG
snarfed: Writing up the Semantic Linkbacks changelog for the PR so pfefferle can go stable. You were very busy
#
snarfed
GWG: huh i thought i captured all the facepile changes in the changelog
#
snarfed
gotta run, back later
tantek, [kevinmarks], aaronpk and [miklb] joined the channel
#
GWG
Does anyone have any thoughts on the possible dangers of parsing all URLs in a page looking for video/audio extensions?
[cleverdevil] and [miklb] joined the channel
#
jjuran
GWG: If parsing a URL is considered dangerous, you have a serious problem.
#
aaronparecki.com
edited /MediaWiki_upgrade (-499) "/* Plugins */ update based on extensions actually in use"
(view diff)
#
aaronparecki.com
edited /p3k (-136) "remove use of embedded github issue extension"
(view diff)
#
aaronparecki.com
edited /wiki (-22) "remove use of embedded github issue extension"
(view diff)
#
aaronparecki.com
edited /MediaWiki_upgrade (-101) "removed github extension"
(view diff)
#
GWG
We had a Github extension?
#
aaronpk
it was only on two pages, but embedded a list of github issues in the wiki page
#
www.svenknebel.de
created /User:Kaja.sknebel.net/Events (+50429) "Created page with "<span style="float:right;height:192px;font-size:128px;margin-top:1em; margin-bottom:-100px">📆</span> <span class="p-summary">'''Welcome to upcoming IndieWeb <dfn>events</dfn>!...""
(view diff)
KartikPrabhu joined the channel
#
Zegnat
GWG, what does “parsing all URLs” mean? Run a regex looking for URLs? Parsing HTML?
#
GWG
Yes. And I'm trying to think if there is a security issue for a user.
#
Zegnat
There shouldn’t be any security issue from just looking for URLs. There can be security issues depending on parsing though. E.g. if you parse HTML with libxml (e.g. PHP’s DOMDocument) some XML bugs could have it crash.
KartikPrabhu joined the channel
#
Zegnat
Automatically loading video files could theoretically have security issues, because the video lib on the user’s side may have issues. Libav I think has certain security problems in their video decoding. So you do not ever want to start video without letting the user decide.
#
doubleloop.net
edited /bookmark (+95) "/* Additional Indieweb Discussion */ Add link to IWC Berlin 2017 session"
(view diff)
sebsel joined the channel
#
GWG
aaronpk: Why do you have two different locations on your check-in posts?
#
aaronpk
one is the venue i'm checking in to, the other is where I am when I posted the checkin
#
GWG
I'm trying to look at my markup.
#
GWG
And how I do location in general.
#
GWG
I have check-ins and location coming from two different pieces of code
#
aaronpk
me too. the location property comes from my GPS database
#
aaronpk
the checkin comes from the micropub request
#
GWG
Well, in my case, it is different code looking at the same stuff. But being as the checkin is independent of the location and vice versa, I need to think about it.
#
Zegnat
That’s why my location and checkin properties are set to the same h-card on my blog: I was at said venue when I checked into it.
[kevinmarks] joined the channel
#
[kevinmarks]
Foursquare keeps the user's geopoint and the venue separately (that's part of how they get good at guessing locations)
#
GWG
I may have to do that
#
GWG
I still haven't done venues.
#
aaronpk
I cheated on venues by not having a venue database. venue data is stored in each post
KartikPrabhu and leg joined the channel
#
GWG
I am not implementing till I have a place to store them.