#dev 2018-09-10

2018-09-10 UTC
[kevinmarks], KevinMarks_, mblaney, KevinMarks, renem, KartikPrabhu, [iambismark], iasai, cweiske, swentel, barpthewire, jeremych_, ben_thatmustbeme and [grantcodes] joined the channel
#
[grantcodes]
miklb: What part of xray? There is a microformats parser and then there are a whole bunch of meta parsers depending on what you want to do
deathrow1, Amgine, [jgmac1106], [schmarty] and [eddie] joined the channel
#
swentel
[schmarty], any idea what I might be missing when the indiewebring can't validate my links ? they are on the site but the checker seems to miss them ..
miklb_ joined the channel
#
[schmarty]
swentel: sorry for the trouble. let me take a look at the logs.
#
Loqi
[schmarty]: [grantcodes] left you a message 6 days, 21 hours ago: Been using your indieauth lib, but I can't open issues on https://git.schmarty.net/schmarty/indieauth-authentication So 1. It should support scopes, 2. Should be able to get a token 3. Resolve and return redirected user urls. Happy to write pull requests, but don't want to create another separate npm module
#
Loqi
[schmarty]: swentel left you a message 3 days, 21 hours ago: tried verifying my site (realize.be) on indieweb ring, but doesn't seem to find the links .. am I doing something wrong ? :)
#
swentel
oh oops
#
swentel
didn't think loqi would find that as I did it with [] .. :)
#
swentel
s/with/without
#
[schmarty]
loqi knoooooows
#
swentel
apparently
#
[schmarty]
ok, i'm watching the logs now, if you don't mind trying again.
#
[schmarty]
haha well there were no fatal errors so nothing in the logs. 😂
#
[schmarty]
swentel: stepping through the code, i don't see why it's not picking up the links. i'm going to expand this with an admin panel that'll let me run the checks myself, add some debugging, see what i can find.
#
swentel
[schmarty], ok :)
#
swentel
[schmarty], just realized, maybe it's because the site itself is rendered by vue. the content is in the source though, but in a noscript tag
#
swentel
maybe that's the problem
#
swentel
maybe the parser ignores noscript tags, might be a pointer to debug
#
[schmarty]
hmm! great question! i was stepping through the link checking in my devtools, after the page had rendered.
#
[schmarty]
hmm, that doesn't seem to be the issue. disabling JS on page load, i still seem to be able to find the links.
#
[schmarty]
for now i've manually marked you as active 😅
#
swentel
hehe, cool thanks :)
#
[schmarty]
[grantcodes] thanks for the feedback on indieauth-authentication! i was going to wait to put it on github until i had un-hardcoded the tests, but i'll make a github fork now.
#
[schmarty]
!tell [grantcodes]: made a github fork of indieauth-authentication if you're still interested in sending PRs! https://github.com/martymcguire/indieauth-authentication
#
Loqi
Ok, I'll tell them that when I see them next
#
Loqi
[martymcguire] indieauth-authentication: A JavaScript helper class for creating apps that authenticate via IndieAuth.
[grantcodes] joined the channel
#
[grantcodes]
Personally although I know it's technically a separate thing, I really think I single library that could handle auth, scopes and tokens would be ideal
#
[schmarty]
[grantcodes]: agreed that a single library feels best. handling URL normalization / resolution, auth, and scopes/tokens. bonus if it can also get representative h-card info. mega bonus if it can do that while minimizing requests to the user's homepage. :}
#
[grantcodes]
It should be easy enough to get a page once, then just use the returned html / headers for everything else
#
[schmarty]
i believe it's reasonable! actually, a mini-cache for html/headers sounds great because it'd make testing easier.
jgmac1106, KevinMarks, [tantek], [schmarty], [manton], swentel and [keithjgrant] joined the channel
#
[keithjgrant]
Is microformat-node my best bet for parsing am entry on node? (All I really need are links within the main h-entry for sending WMs)
#
sknebel
[keithjgrant]: it's the one javascript-based parser we have, so if you want to parse a page then yes (not sure what you're building?)
#
[keithjgrant]
Well, I use microformat-shiv in Omnibear
#
sknebel
that's the same parser
#
[keithjgrant]
I'm just starting on a github-to-webmention app to send WMs for static sites
#
sknebel
microformat-node is microformat-shiv, just with a DOM implementation added since microformat-shiv uses the browsers
#
@ChrisAldrich
↩️ Your logs or even database may indicate better where it came from, but most likely it's a malformed received webmention or one without data. I've seen one or two in the wild, but unsure what may have caused it. I thought it might be from… http://stream.boffosocko.com/2018/christo37433424-your-logs-or-even-database-may-indicate-better-where
(twitter.com/_/status/1039187525577965572)
#
sknebel
ah. yeah, when you don't know details about the page structure just parsing it is probably the way to go
[schmarty] joined the channel
#
jeremycherfas
Definite noob question: if I want to use rsync to backup a folder on a remote server to my local machine, is it best to run the command on the local server or on the remote server? I can use full paths in either case.
#
[schmarty]
jeremycherfas: i find that what's "best" for that situation depends on several things. e.g. which machine is likely to be up more. i have a couple of local machines that mostly store backups and i have remote servers log into them using restricted accounts to dump their backup data.
#
jeremycherfas
OK. Well, they're both up as much as possible. I would like eventually to make it a cron job, but will do it by hand to begin with.
snarfed joined the channel
#
snarfed.org
edited /Twitter (+327) "can hide retweets from individual accounts"
(view diff)
eli_oat, jgmac1106 and snarfed joined the channel
#
pstuifzand
Once in a while there are people who like to import and export OPML from their microsub server. This is now possible with the `ek` command.
#
Loqi
[Peter Stuifzand] Import OPML files into Microsub servers
#
snarfed
pstuifzand++
#
Loqi
pstuifzand has 6 karma in this channel over the last year (8 in all channels)
snarfed joined the channel
[keithjgrant] joined the channel
#
[keithjgrant]
So, given a page url... Is there a set algorithm for determining which links on the page should have webmentions sent? All links in the main h item? How do I know which h item is the "main" one?
#
sknebel
the one with u-url / u-uid == page url is the main one
#
KartikPrabhu
also all links are mentioned
#
KartikPrabhu
except you might want to avoid sending webmentions to you your own tag-pages or something
#
[keithjgrant]
Yeah, I'm looking to avoid footer links and that sort of thing
#
[keithjgrant]
From what I can tell, Telegraph only lists links in the main h item
#
snarfed
alternatively just send to everything *except* your domain(s) and localhost etc
#
snarfed
limiting to h-item risks omitting comments (necessary for salmentions), other custom areas, etc
#
KartikPrabhu
but sometime you want to mention your own previous posts
#
KartikPrabhu
also comments should be in h-* as p-comment no?
#
[keithjgrant]
Of my page displays comments I've received, I don't want to send WMs back, do I?
#
snarfed
KartikPrabhu: usually, but markup varies
#
[keithjgrant]
So if my post is a reply to post X, I would have send a webmention when I first posted. Then, if mine updates, it would re-send the original webmentions, and post X could see successive comments now displayed on my site. Correct?
#
[keithjgrant]
But if my post displays a comment I've received—post Y–I don't need to send webmentions to Y, do I?
#
snarfed
i haven't followed salmentions in detail, but i think in general we expect wm smarts to be mostly in receivers, not senders. ie pages send webmentions pretty blindly to everything they link to, and receivers interpret semantics, de-dupe, etc
renem joined the channel
#
@rubygems
jekyll-webmention_io (3.0.0): This Gem includes a suite of tools for managing webmentions in Jekyll: * Tags *… https://rubygems.org/gems/jekyll-webmention_io
(twitter.com/_/status/1039250669096652800)
snarfed, [gregorlove] and [manton] joined the channel
#
aaronpk
Yep exactly that. The reason is that if there's a link from one page to another, anyone can send a webmention for that link.
#
aaronpk
Think of sending a webmention as a shortcut for someone stumbling across an existing link from one page to another. It speeds up that process but the link still exists whether or not the webmention is sent
DenSchub and [keithjgrant] joined the channel
#
[keithjgrant]
👍 that totally makes sense
#
[keithjgrant]
So, getting practical, do most endpoint implementations do a good job of ignoring incoming WMs for irrelevant links like that?
mblaney joined the channel