#dev 2019-01-04

2019-01-04 UTC
KartikPrabhu, [davidmead], [schmarty], [tantek], snarfed, leg, [cleverdevil], cweiske, barpthewire, swentel, [Vincent], krychu_, [relapse], [metbril] and [voss] joined the channel
#
[metbril]
[dgold] You removed support for Twitter from nanopub. What was the rationale behind this?
[Rose] joined the channel
#
dgold
[metbril]: i don't support that organisation, its amplification of nazis, white supremacists and fascists, its general posture towards minority groups in general, its management.
[kevinmarks] joined the channel
#
dgold
if twitter want to pay me to facilitate the delivery of content to its servers, I'll reconsider the situation.
#
dgold
i'm done being complicit with these companies.
#
[metbril]
That's a clear viewpoint. Unfortunately, I think a lot of us do want to syndicate, and this would create a need for a fork that includes the old function, unfortunately.
#
[metbril]
Anyway, I will be creating a Micropub server endpoint plugin for Grav CMS that is inspired by your code, not just forking it.
#
[Rose]
I have a PHP file I can drop in to a Grav install which works great as a micropub server - I can't find the time to make it into a plugin though.
#
[metbril]
If you are willing to share it (gist, github repo, your site or otherwise, I will have a look. We could make it a joint effort.
#
[metbril]
BTW, is it a media endpoint as well? For OwnYourGram etc.?
#
[Rose]
It is. It's basically what Aaron has done a few times at IWCs - building an endpoint quickly. I added a config file for the folders where things should be saved, but that's pretty much the only changes I made.
#
[metbril]
That sounds feasible. I already created an IndieAuth plugin for convenience: https://github.com/metbril/grav-plugin-indieauth
#
[kevinmarks]
[Jacky_See] which CSS framework uses h- ? I've seen u- a lot (which is annoying enough) but h- is extra annoying. (Previously http://www.kevinmarks.com/u-means-style.html )
#
[kevinmarks]
gah, bad autocomplete, slack
#
[kevinmarks]
how do I stop slack being wrong here?
jgmac1106 and [jgmac1106] joined the channel
#
dgold
[metbril]: the code is still there in git history
#
dgold
feel free to grab & use it - I won't be using it, I have no way of testing it, so I cannot support it
#
Zegnat
Hmm, was anyone working on integrating the IndieAuth endpoint into Grav? Did you have plans to go that way, [metbril]?
#
Zegnat
has grown slightly allergic to indieauth.com :P
#
dgold
[metbril]: I don't see any problem with a fork of nanopub. subject to the usual licencing stuff, obvs.
[davidmead] joined the channel
#
Zegnat
Isn’t there already a fork? Who was it by ... skippy? schmarty? something s*?
#
Loqi
[skpy] micropub: a minimal PHP micropub endpoint, with media support
#
Zegnat
That still includes Twitter sync
[metbril] joined the channel
#
[metbril]
@Zegnat I have created an initial IndieAuth plugin that connects to IndieAuth.com. This could evolve to and IndieLogin endpoint later on. However, I wanted to switch to Grav (and Micropub) quickly, and there is no other option atm. I could have added the endpoints manually, but thought this was neater.
#
[metbril]
BTW, for now, I myself would be the only one using it for my personal site, am I not? So no reason for allergic reactions?
#
Zegnat
You never know who finds your plugin. You are pretty high up the search results for "grav indieauth" ;)
#
Zegnat
But for sure, prioritise your own itches first [metbril]!
[Vincent] joined the channel
#
GWG
Morning
#
Loqi
*yawn* ... /stretch/
#
[metbril]
I am no professional developer. How hard would it be to develop the Auth endpoint for Grav?
#
Zegnat
I don’t know what sort of APIs are in Grav, but maybe not too hard?
#
Zegnat
There is a lot of reusable code as well. You can probably just wholesale copy the Mintoken PHP code for a token endpoint
#
Zegnat
What is Mintoken?
#
Loqi
Mintoken is minimal and self-hostable IndieAuth compatible Token Endpoint in PHP https://indieweb.org/Mintoken
#
[metbril]
For now Micropub and Webmention are top of the list of itches. After that I will have a look.
#
Zegnat
I should maybe look at Grav again at some point. Is it any good for blogging?
#
Loqi
totally
#
Zegnat
Thanks Loqi, but I don’t think you ever used it, so, hush~
#
Loqi
you're welcome
#
Zegnat
Loqi++ Hahahaha
#
Loqi
Loqi has 6 karma in this channel over the last year (50 in all channels)
#
Loqi
Zegnat: lol
#
Zegnat
gives Loqi just a little more sentience
#
[metbril]
You should ask [Rose]. She is the more avid blogger. I find the editing interface in the Admin panel sort of hmmmm. But Micropub would solve that.
#
[Rose]
I don't use the Grav UI except occasionally to install plugins.
#
[Rose]
I use Shortcuts to post with the "Run Script via SSH" command 😄
#
Loqi
agreed.
#
[metbril]
My main reason is that is has plain text / markdown files for editing and is not a static site with its own shortcomings.
#
[Rose]
Zegnat what did you do!?
#
Zegnat
I almost had a conversation with Loqi. I just need to keep feeding the machine, [Rose]!
#
[metbril]
[Rose] another option would be to post through GitHub with Working Copy and the GitSync plugin. (If you need a new pet project)
#
Zegnat
I want to make 2019 my year if blogging, and I wonder if that’ll be easier if I stop also making it the year of developing a blog engine… just need to find one I like now.
#
@mxbck
↩️ Thanks! Yeah its manual now bc I ran out of time, but I plan on automating it. Pushing new notes to twitter maybe via deploy hook or similar. Then pull reactions back via webmentions / bridgy.
(twitter.com/_/status/1081182960026558464)
#
[metbril]
I you like to tinker, now PHP, yaml and markdown, than Grav is your thing 😉
[relapse] joined the channel
#
[relapse]
mmmm, tinkering.
#
[metbril]
@zegnat "manual until it hurts" ?
#
Zegnat
I tried that once, but I couldn’t get into the whole "edit HTML, reupload" mindset often enough to encourage writing
#
Zegnat
Right, Grav comes with YAML, hmmm
#
[metbril]
why edit HTML?
#
[Rose]
I have a script that generates the YAML header for me.
#
[relapse]
Handy.
#
Zegnat
[metbril]: I was manualling until its hurts, that is: I was writing every blog post as an HTML file
#
[metbril]
With Grav you just create/edit a markdown file. Only HTML for difficult stuff.
#
[metbril]
My biggest annoyance with Grav is that it NEEDS a required slug for each post, since that will be the folder name. In most templates that slug will show up as the post title.
#
[metbril]
When adding IndieWeb notes that is difficult. I am still thinking how to properly address this.
#
Zegnat
Sounds like Grav just uses "slug" as a term for URL? Notes have URLs too, so whatever makes sense there?
#
[relapse]
Can you sub in the Post id for the Title if its a non-title-needing post-type?
krychu_ joined the channel
#
[Rose]
I just use the date/time string as my slug for all my "micro" posts.
#
[Rose]
(My currently micropub endpoint is setup to do that)
#
[Rose]
And I customised my theme to not substitute the slug in.
#
[metbril]
zegnat, slug is the same use as in WordPress. It is part of the url. Depends on the route e.g. the place in the folder hierarchy.
#
[metbril]
[relapse] Grav posts don't have an ID like WordPress. In the end, their slug is their unique ID.
#
[metbril]
I was thinking of using a Unix timestamp for slug if one is required. Or a ISO8601 date.
#
[Rose]
I do yyyy-MM-dd-hhmm
#
aaronpk
can you make a slug that has slashes? Like /2019/01/04/9876?
eli_oat joined the channel
#
[Rose]
Readable to humans, works on the machine no problem.
krychu_ joined the channel
#
[Rose]
The slug is a folder name, I think slashes are forbidden,
#
[metbril]
[aaronpk] You can in various ways. That would require nesting posts deep in hierarchy. But each level needs a markdown file
#
[relapse]
Could be worth reading the URL design page on indie web. YYYY/MM/DD/Iterator sounds close. https://indieweb.org/URL_design
#
aaronpk
Interesting
#
aaronpk
Grav looks neat, I should look at it more closely
#
[metbril]
I have resolved this by overwriting the generated permalink (/blog/some-slug) with (/2019/01/04/some-slug)
#
[Rose]
I''m not sure nesting all those folders is a great idea on Grav. Mine already takes forever to recache with all the images I have.
#
[metbril]
I have all posts under Blog and overwrite the default route.
#
[Rose]
(I'm exaggerating, but it's long enough that statuscake tells me all my sites on that server are down)
[pfefferle] joined the channel
#
[metbril]
Code snippet, from an alpha version of a plugin to rewrite permalinks
#
[metbril]
public function onPageProcessed(Event $e)
#
[metbril]
// only parse pages with 'item' template (blog entries)
#
[metbril]
$page = $e['page'];
#
[metbril]
if (!($template == 'item')) {
#
[metbril]
return;
#
[metbril]
$header = $page->header();
#
[metbril]
$template = $page->template();
#
[metbril]
$date = isset($header->date) ? date('Y-m-d', strtotime($header->date)) : date('Y-m-d');
#
[metbril]
$date = substr(preg_replace('/-/', '/', $date), 0, 10);
#
[metbril]
$slug = $page->slug();
eli_oat joined the channel
#
[Rose]
For anyone following from the main channel: There's nothing in my laravel or sql log files that indicate talking to watchtower is failing
#
[Rose]
And watchtower itself has no log files, which I think implies something is wring
#
aaronpk
[Rose]: when you run the watchtower script it should be writing log files and showing you output on the console
#
[Rose]
It's running via cron, theoretically.
#
aaronpk
watchtower needs that background script running
#
[relapse]
[metbril] Have you thought of using 'Ymd' as the date format string rather than 'Y-m-d' and then preg_replacing?
#
aaronpk
I don't think it works via cron
eli_oat1 joined the channel
#
[relapse]
Can you run the script manually as a test?
#
[Rose]
This said to run a script via cron: https://ascraeus.org/websub-part-ii-aperture/
#
aaronpk
Oh right it needs both
#
aaronpk
the cron queues up the fetching tasks
#
[Rose]
Which is the other half of "both"?
#
[metbril]
[relapse] As said, I'm not a pro coder. Just doing this in my free time. So my coding skills are limited.
#
aaronpk
but you have to run the worker too
#
aaronpk
in watchtower
#
[Rose]
No sign of a worker in that link, so I won't have done it.
#
[Rose]
I'm not familiar with Laravel things
#
[Rose]
(Something I want to improve!)
#
[Rose]
How do I start the worker?
#
Zegnat
[Rose]: see https://ascraeus.org/websub-part-i-watchtower/ under “Watchtower as a Service”
#
Zegnat
That service is the worker and needs to be constantly running
#
Zegnat
dgold++ for having IDs on his headers
#
Loqi
dgold has 16 karma in this channel over the last year (25 in all channels)
#
[Rose]
Pretty sure I did that, but I may have missed it
[schmarty] joined the channel
#
Zegnat
Someone in here can probably tell you how to check systemctl for added services and whether they are running. sknebel?
#
[Rose]
I'm checking the systemctl already
#
Zegnat
Good :)
#
Zegnat
I do not use it enough to know by heart where to check
#
Zegnat
needs one of those on the go RPi server setups for practice
#
[Rose]
watchtower.service is enabled
#
sknebel
is it running?
#
[Rose]
State=active? It seems not
#
[Rose]
Just as well, there's a typo in the script path!
#
dgold
those are placeholder paths (on my site)
#
[Rose]
Yes, I had wathtower.rosemaryorchard.com instead of the correct url there
#
[metbril]
[relapse] I took that challenge and came up with this:
#
[metbril]
$date = isset($header->date) ? date('Ymd', strtotime($header->date)) : date('Ymd');
#
[metbril]
$date = preg_replace('/(\d{4})(\d{2})(\d{2})/', '\1/\2/\3', $date);
#
[metbril]
That's probably what you were refering to.
#
[relapse]
OK. I'm wondering why you need a preg_replace after, since it's run through date it'll be dddddddd or null if strtotime fails.
#
[relapse]
I don't use Grav but I tried trimming it to the below. Be interesting to see if it works.
#
[relapse]
`public function onPageProcessed(Event $e)
#
[relapse]
{
#
[relapse]
// only parse pages with 'item' template (blog entries)
#
[relapse]
$page = $e['page'];
#
[relapse]
if ($page->template() != 'item') {
#
[relapse]
return;
#
[relapse]
}
#
[relapse]
$date = date('Ymd', isset($header->date) ? strtotime($header->date) : time());
#
[relapse]
$header = $page->header();
#
[relapse]
$route = '/'.$date.'/'.$page->slug();
#
[relapse]
$header->routes['default'] = $route;
#
[relapse]
$page->route($route);
#
[Rose]
Uh oh, uncaught Pheanstalk Connection Exception
[smerrill] joined the channel
#
[smerrill]
aloha.
#
[metbril]
I see what you mean. Just
#
[metbril]
``` $date = isset($header->date) ? date('Y/m/d', strtotime($header->date)) : date('Y/m/d');
#
[metbril]
that should have been shift-enter 😞
#
[relapse]
heh.
#
[relapse]
yup anyway.
#
[metbril]
$date = isset($header->date) ? date('Y/m/d', strtotime($header->date)) : date('Y/m/d');
#
[metbril]
does work, too. I've learned something today. 👏
#
[relapse]
Winning!
#
[relapse]
I shunted the date('Y/m/d'... to the start so you don't have to specify the format twice, and have the opportunity to miss setting one of them. Which I do far too often.
#
[relapse]
Hi [smerrill]
#
[metbril]
In your sample you omitted the forward slashes, which would be required for the permalink path I need.
#
[Rose]
Hi Smerrill 🙂
#
[relapse]
Indeed, I missed that, sorry.
#
[smerrill]
lots of micropub talk. that’s exciting. do feel free to use anything from my attempt for your own. 🙂
#
[Rose]
The service itself won't run, it says "Start request repeated too quickly"
#
sknebel
that probably means it isn't starting properly
#
sknebel
any error shown?
#
sknebel
in systemctl status or in a log written by the service?
#
[Rose]
When I run the file manually I get the Pheanstalk Connection Exception mentioned above.
#
[metbril]
I've given the ordinal number in the permalink a thought, but that needs some thought. The `date:` header allows for just a date without a time. If not present it defaults to the date/time of last save of the file. So in order to create an ordinal number, I would first need to sort all posts with knowing that the order will never change again.
#
sknebel
[Rose]: what's the precise error?
#
[metbril]
Hope I'm making sence. English is not my native language.
#
[Rose]
It's really long, uploading it.
#
sknebel
is beanstalkd running?
#
[Rose]
Not that I know of
#
[relapse]
[metbril] Perfect sense. Yes, if Grav doesn't use a numeric key for posts and relies on unique directories, then ordinal would take a lot of calculating to ensure it never changed.
#
sknebel
if you installed it through the system package manager it should have a systemd unit already, check that?
#
[metbril]
As long as I manually create posts through the interface, I could keep track of this, but with Micropub clients, OwnYourGram, OwnYourSwarm etc. this would be very difficult if not impossible.
#
[Rose]
systemd is running for lots of other things
#
sknebel
and start and enable that if it isn't yet
#
[relapse]
Agreed.
#
[Rose]
It is, if I do systemctl status watchtower I get "Result: stat-limit'hit"
#
[Rose]
*start limit hit
#
sknebel
no, for beanstalkd
#
sknebel
systemctl status beanstalkd.service ?
#
[Rose]
There's no beanstalkd service.
#
[Rose]
Where would it have come from? I haven't set that up (not in the instructions)
#
sknebel
System Requirements: Beanstalkd
#
[Rose]
As it's including things I presumed that meant it requires the files.
#
Zegnat
These things need to be on the server as baseline: https://ascraeus.org/websub-part-i-watchtower/#system-requirements
#
sknebel
beanstalk is a service
#
[Rose]
Perhaps a line stating it's a service/a link to it
#
aaronpk
so many dependencies, sorry!
#
[Rose]
*wouldn't go amiss
#
Zegnat
There are no official docs ;)
#
[Rose]
Installing now. I suspect that will resolve my issue
#
[Rose]
Even in these unofficial instructions it would be nice 😛
#
Zegnat
sknebel: should the watchtower unit file be updated to depend on / not start before beanstalkd?
#
sknebel
Zegnat: probably?
#
aaronpk
good idea
#
Zegnat
never knows if it is After= or Requires= or something else that is needed for that
#
Zegnat
I have now setup some services, even my own little PHP workers, but I am still guessing the settings 99% of the time
#
aaronpk
ubuntu went through a few iterations of arguing about which startup thingy to use and now I don't know how to use any of them or even know what they are called or which one is the "correct" one to use now
#
[Rose]
Beanstalked disapproves of something, but at least I'm on the right track now
#
sknebel
Zegnat: I'd say After= and Requires=
#
Zegnat
Just use both? Guess that would work too
#
sknebel
Requires= - "this won't work without this other thing"
#
Zegnat
I thought systemd appended one to the other in some magical way. But maybe best not to assume such things
#
sknebel
After= - "this doesn't like it if the other isn't available when it starts, so wait with starting until the other is ready"
#
Zegnat
aaronpk, it is a weird tech landscape. And once you have decided on the whole systemd vs initd, vs whatever, you can still decide to do your own process control with tools like supervisord, daemontools, etc.
#
Zegnat
Confuses the hell out of me
#
Zegnat
I was often being pointed towards http://supervisord.org/ as an alternative for the systemd messing.
#
aaronpk
oh yeah I started using supervisord too
#
sknebel
well, Ubuntu did their own adventure through upstart beforehand
#
aaronpk
so some of my things are run through that now
#
[Rose]
Hmm, apparently invocation ids are required by both beanstalkd and watchtower, and they're not getting them
#
aaronpk
has no idea what an invocation id is
#
[Rose]
Watchtower's throwing that error message at me 😛
#
sknebel
precise error?
#
aaronpk
sound like a systemd thing maybe?
#
[Rose]
It must be a systemd thing, that's all gogole is telling me
#
aaronpk
I wish there was a good way to run stuff in the background that didn't involve this mess
#
[Rose]
"Failed to set invocation ID on control group /system.slice/beanstalkd.service, ignoring: Operation not permitted"
#
[Rose]
And the same with watchtower, but with "watchtower" instead of "beanstalkd"
#
sknebel
that shouldn't affect the start though, I think it's "just" a warning
#
[Rose]
I suspect the second error I get "Failed at step CHROOT spawning /path/to/php" is more important, but I have no idea where it's referencing this.
#
sknebel
you don't have literally "path/to/php" somewhere?
#
[Rose]
No, I checked the .service and it's /usr/bin/php
#
sknebel
is that for watchtower?
#
Zegnat
Hmm... any chance systemd is caching a unit file from before it got changed to /usr/bin/php?
#
[Rose]
I did daemon-reload, and restarted the server when that didn't work for good measure
#
[Rose]
Beanstalkd is not referencing path/to/php either as far as I can find
#
Zegnat
Guess you can always ln /path/to to /usr/bin if nothing else works, hahaha
#
Loqi
hehe
#
[Rose]
I also grepped and couldn't find it inside of the watchtower directory
#
sknebel
just to be sure, you're looking at the same unit file that's listed in the `systemctl status` in the Loaded: line?
#
Zegnat
Feels like the ExecStart line in the unit file, not something that would be in watchtower itself
#
[Rose]
Ahh, but if I grep through the whole server... It's in .viminfo, that doesn't make sense?
#
sknebel
viminfo just has a cache of stuff you've done in vim, that's harmless
#
Zegnat
Any way to have systemctl echo whatever it thinks the ExecStart line of a unit is?
#
[Rose]
So, in conclusion, it's nowhere on my server according to grep, but it is according to systemd
#
[Rose]
I'm trying turning it off and turning it back on again
#
sknebel
systemctl show watchtower.service?
#
[Rose]
ExecStart shows /usr/bin/php
#
[Rose]
ExecStart={ path=/usr/bin/php ; argv[]=/usr/bin/php watchtower.php ; ignore_errors=no ;
#
[Rose]
I'll poke at some of the other errors, see if I get anywhere
#
[Rose]
Now it runs, crazy
#
[Rose]
(I hadn't done anything more than open a folder via SFTP, I'm convinced there's a gremlin in my server)
#
sknebel
did the errors have timestamps? maybe it was showing you old logs?
#
[Rose]
16:26 was the timestamp.
#
[Rose]
Maybe yesterday? I'll ssh back in later and check
#
ben_thatmustbeme
is anyone using 2fa on their indieweb sites?
#
[smerrill]
not me.
#
Zegnat
Once upon a time I had 2FA on my IndieAuth endpoint. Hoping to bring that back
snarfed joined the channel
#
Zegnat
I actually got a yubikey at IWC Berlin and have wanted to sit down and try a passwordless IndieAuth ever since, ben_thatmustbeme
#
aaronpk
ben_thatmustbeme: have you been able to get webauthn stuff working?
#
ben_thatmustbeme
Zegnat: aaronpk: i got a sample working at one point, it wasn't too bad, someone had a nice sample
#
aaronpk
do you remember the link?
#
ben_thatmustbeme
looking it up
#
aaronpk
last time I tried was when everyone announced the browsers were supporting it
#
aaronpk
but I couldn't get the sample code to work
#
aaronpk
the demo worked for me in chrome but I couldn't figure out how to do it myself from scratch
#
Loqi
[davidearl] webauthn: An implementation of webauthn in PHP on the server side (e.g Yubico 2 and Google Titan keys)
#
ben_thatmustbeme
that works for me in FF
#
ben_thatmustbeme
not sure if thats because i turned on u2f flag, but it shouldn't be, looked at the code and it should be using the webauthn code
#
ben_thatmustbeme
yubikey works perfectly with it
#
ben_thatmustbeme
the CBOR stuff is really.... weird. would love to know who had that idea so i could yell at them, haha, but yeah, i was writing it all myself until i found that code and it was basically exactly what i was writing
#
Zegnat
Hmm, I had never ran into CBOR before.
#
Zegnat
reads RFC
#
[smerrill]
can anyone identify what I might be doing wrong here? https://github.com/aaronpk/Quill/issues/107
#
[smerrill]
So far as I can see, my server is responding “201 Created” with the header; but Quill doesnt seem to agree.
#
aaronpk
[smerrill]: in the quill note interface, when you upload an image, does it show you the URL?
#
[smerrill]
I do, if I wait a bit…
#
aaronpk
ok that's good, that means quill is recognizing the upload to the media endpoint succeeded
#
[smerrill]
soooo! I guess I was being impatient.
#
aaronpk
ah lol
#
[smerrill]
I just tried again, waiting to get that URL back. Previously I had NOT been waiting long enough to get that.
#
aaronpk
so I need some sort of progress indicator in quill
#
[smerrill]
i’ll close my ticket. silly me.
[davidmead] joined the channel
#
aaronpk
heh ok
#
aaronpk
ben_thatmustbeme: wow this code looks like stuff I never want to write https://github.com/davidearl/webauthn/blob/master/src/webauthn.php
#
@mauricehh
Sorry, ich muss ein bisschen spammen, versuche gerade #webmentions zum Laufen zu bewegen, es läuft suboptimal… https://maurice-renck.de/de/blog/holen-wir-uns-das-netz-zurueck
(twitter.com/_/status/1081224517073190913)
#
Zegnat
I am a little upset about the PHP code style used there, but it looks understandable enough
[sebsel] and krychu_ joined the channel
#
ben_thatmustbeme
aaronpk: exactly, its pretty rough to do the underlying spec bits
#
ben_thatmustbeme
need to make a really nice lib out of it
#
ben_thatmustbeme
Zegnat: to fix it :D
#
Zegnat
You know what, ben_thatmustbeme, if nothing else comes up this weekend maybe I will! Hahaha
#
Loqi
awesome
#
ben_thatmustbeme
i agree Loqi
[kevinmarks] joined the channel
#
ben_thatmustbeme
gives Loqi a pat on the back
#
Loqi
investigates the pat on the back
[Ana_R] joined the channel
#
[Ana_R]
Hi everyone - I could use bounce back some ideas because i think i’ve messed up in my blog 😞 I have a jekyll blog and I’m using a micropub that is a fork of https://github.com/voxpelli/webpage-micropub-to-github
#
[Ana_R]
Out of box they work fine. They have specific categories and I can filter them in jekyll but it means that my `_posts_ folder has my normal posts, instagram posts and bookmarks. I think I want to organise them better, especially the bookmarks.
#
[Ana_R]
I use that micropub for OwnYourGram and most recently for Quill (I tried to implement bookmarks on my blog).
#
[Ana_R]
Does this make sense? I don’t even know what I exactly asking.
#
[Ana_R]
Currently, my bookmarks live here (https://www.ohhelloana.blog/bookmarks) but if, for any reason, I want to share the link of a bookmark, I think I would prefer if it looked like: `https://www.ohhelloana.blog/link/me-in-mastodon-social` instead of `https://www.ohhelloana.blog/me-in-mastodon-social` (as it is currently).
#
[Ana_R]
Maybe my question should be, does it make sense not to have all types of `posts` in the same `_post`folder? I don’t know if this is something i may regret in the long run especially after I add lots of bookmarks.
#
[Ana_R]
Thank you!
#
[Ana_R]
Also, how in jekyll with my current implementation, could I change my bookmark’s urls to have `\link\`?
#
jacky
I think that's a Jekyll thing
#
jacky
(but never fear!)
#
jacky
Like you might able to make a custom path for collections
#
jacky
I use that to make a custom path for my FAQ, portfolio and other pages
[metbril] joined the channel
#
jacky
but that'd require your posts to be identifiable to Jekyll as a part of a collection
#
[metbril]
[Rose] could you share your Micropub code for Grav once your server issues have been resolved?
#
jacky
does that help [Ana_R]?
#
[Ana_R]
it does help! i’m going to give it a go and write here how it goes. Thanks for sharing your example!
[tantek] joined the channel
#
jacky
Totally! :)
#
jacky
also I like that border effect you have of your Instagram posts to your site
[kevinmarks] and [cleverdevil] joined the channel
#
[Ana_R]
Thank you!! 😄
[Rose] joined the channel
#
[Rose]
As I said, it's not really for Grav
chrisaldrich joined the channel
#
[Rose]
It's just the micropub endpoint Aaron has built a few times at IWCs, with the save locations set to config parameters
[davidmead], snarfed, [schmarty] and swentel joined the channel
#
[Ana_R]
yeah, turns out that i can’t add normal posts to collections 😞
#
[Ana_R]
My “bookmarks” are generated with Quill + the micropub endpoint i’m using so the `slug` is built there, otherwise i would add `\link\` to it.
jackjamieson and snarfed joined the channel
#
[Ana_R]
I think i can survive this 😂 I wanted to have fancy urls but i guess it won’t be possible as it is at the moment 😞
#
[Ana_R]
I want to confirm as well, am I doing the right thing here? Bookmarks and instagram posts are `posts` from Jekyll’s perspective?
[smerrill] joined the channel
#
[smerrill]
\o/ yeah! I think I finally have my micropub endpoint doing all the things I want, in a way that works, and site build times are down overall. That’s enough tinkering for now!
#
[schmarty]
[Ana_R]: that sounds right. everything in the _posts folder is part of the posts collection.
#
[schmarty]
if i remember correctly - if you want to separate out links, they would go in a _links folder and you'd need to add some stuff to _config.yml to declare the collection.
#
[Ana_R]
Yeah that’s correct but my bookmarks are being added as `posts` and straight to the main `_posts` folder… which is why I can’t add it to a jekyll collection.
#
[Ana_R]
This is because I wanted to be fancy and add them via a client instead of manually.
#
[Ana_R]
My understanding is that when I use a client (e.g. Quill) to add bookmarks, things are added by default to the default `_posts` folder. But because the micropub endpoint i’m using wasn’t build by me, there is a big chance that I am wrong in my assumptions 😕 I think I understand things separately but putting all the pieces together is a bit more tricky.
[asuh] joined the channel
#
[Ana_R]
wait lol i’m wrong
#
[schmarty]
[Ana_R] the micropub endpoint may have a way of changing where the file ends up based on the contents of the post. hopefully!
#
[Ana_R]
Ok, I think I know where I am wrong in everything. Yeah, I think that’s my problem [schmarty] - thank you
snarfed joined the channel
#
[Ana_R]
i guess i have to build my own endpoint and my own custom way to deploy to my blog depending on the type of content 🙃
#
[Ana_R]
ohh interesting! thank you for sharing!
snarfed joined the channel
#
[schmarty]
[smerrill] good idea! i've been meaning to review your recent changes as well.
snarfed and krychu joined the channel
#
@janboddez
↩️ Also, mark up 'tweet comments'—or, rather, 'webmentions'—as h-cite rather than h-entry.
(twitter.com/_/status/1081282493742047233)
#
swentel
snarfed, woah that's fast :)
#
swentel
snarfed++
#
Loqi
snarfed has 69 karma in this channel over the last year (114 in all channels)
callMeBaby, snarfed and kisik21 joined the channel
#
@mauricehh
↩️ @MacBoeck Ich baue webmentions ein, erwähnst du hier einen Beitrag, soll der tweet als Kommentar im Blog auftauchen. Theoretisch...
(twitter.com/_/status/1081314389762949120)
#
swentel
snarfed, no luck so far, I still need to hack to much into the controllers and helpers to get completely through
#
swentel
for instance it assumes 'attributedTo' is a string and not an array in the AP request
#
swentel
that's not good
#
swentel
but not fed.bridgy's fault (it works for mastodon, so both should work imo)
snarfed joined the channel
#
snarfed
!tell swentel got it. thanks for trying! fortunately dansup is indieweb friendly and likely to merge PRs if you send him one or two to fix those issues
#
Loqi
Ok, I'll tell them that when I see them next
#
snarfed
!tell swentel eg feel free to send him https://www.w3.org/TR/activitystreams-vocabulary/#dfn-attributedto which shows that attributedTo can be an array
#
Loqi
Ok, I'll tell them that when I see them next
[eddie] joined the channel
#
[eddie]
That seems like something that would be pretty cool
#
[eddie]
All this to say I'm jealous because your monthly recap pages look awesome
[Rose] joined the channel
#
[Rose]
I need to figure out how to do a monthly recap page.
[cleverdevil] joined the channel
#
[cleverdevil]
Ah, good call.
#
[cleverdevil]
Fixed. Thanks for the tip, [eddie]
#
[cleverdevil]
I also have a Lambda function now running on the first of each month which publishes a monthly summary to my Aperture.
snarfed joined the channel
#
[eddie]
That’s awesome! How does it get the data? API to your site? Parsing your site’s mf2?
#
[cleverdevil]
Its... gross.
#
[cleverdevil]
Basically, I fetch the page's HTML, then I extract the content and manually shove some of it into an MF2 JSON blob and publish it.
#
[cleverdevil]
Currently, it only publishes the very high level bit at the top ("Jonathan has posted 124 times in December, 2018. Recipes: 1, Posts: 4, Locations: 39, ...")
#
[cleverdevil]
With a link to the full summary.
#
[cleverdevil]
Its not really nice to consume the content itself in my reader.