#dev 2018-10-16

2018-10-16 UTC
#
GWG I appear to have violated the Nominatim terms of service with Simple Location
#
[tantek] what is Nominatim?
#
Loqi Nominatim (UI, documentation) is a piece of open source software and a hosted instance based on OpenStreetMap data for geocoding and reverse geocoding, for example deriving coordinates from a physical address or vice versa https://indieweb.org/Nominatim
#
jacky wait how?
#
GWG No idea. I'm sending them an email now. I think it is because I used the default WordPress user agent
#
GWG https://operations.osmfoundation.org//policies//nominatim///
#
GWG I'm sending them a note.
[jgmac1106] joined the channel
#
[jgmac1106] I did install it and try to do a checkin to see if I could map to display
[dave] joined the channel
#
GWG [jgmac1106]: I may need to find a new geocoding service.
#
GWG Does anyone run their own nominatim instance?
KartikPrabhu joined the channel
#
aaronpk oof I've thought about it many many times
#
aaronpk way more work than I want to sign up for
#
jgmac1106 [gwg] I could ask our NYC sponsor, he is my brother-in-law. He showed me a demo where they tracked every single taxi in different parts of NYC at real time
#
jgmac1106 https://www.blursearch.com/ it is an open source search engine he is customizing for large real time map data. Have no idea what service goes for but maybe he could just let us sit on a box in the back office…well its just a one room headquarters now
bradenslen joined the channel
#
GWG aaronpk, I may have to switch to something with an API key
#
GWG I think MapQuest has a Nominatim server
[asuh] and eli_oat joined the channel
#
[tantek] whoa blast from the past
#
[tantek] what is MapQuest
#
Loqi It looks like we don't have a page for "MapQuest" yet. Would you like to create it? (Or just say "MapQuest is ____", a sentence describing the term)
#
GWG tantek, I would have to write new code to use Google reverse geolocation
#
GWG Same with Bing
renem, [tantek] and eli_oat joined the channel
#
@polm23 Reading over the WebMention spec recently seems unfortunate it requires POST requests - if it used GET static sites could just scan HTTP server logs for mentions. https://www.w3.org/TR/webmention/#receiving-webmentions (twitter.com/_/status/1052034615261577216)
[kevinmarks], jjuran, [dave], [Rose] and [tantek] joined the channel
#
[tantek] "just"
[dave], cweiske and [pfefferle] joined the channel
#
Zegnat Interesting feedback on webmentions in logs. Though I wonder if there is a case to be made where you do not want private webmentions in your server logs.
#
Loqi Zegnat: [eddie] left you a message 1 day, 4 hours ago: yeah, Micropub post list query (q=source) is what MarsEdit needs. I’ve flagged the new developments for Daniel so he can provide feedback on anything missing he needs for it
#
petermolnar rewrote my PASTA archiver based on official/semi official libraries (flickr_api, pytumblr, deviantart), now it's 20x slower, than the handcrafted one...
#
petermolnar although now I generate an opml file for the silo follows I have
#
petermolnar found the problem, fast again, yay!
#
Loqi giggles
swentel joined the channel
#
petermolnar fyi: tumblr doesn't show cookie consent on RSS if the user agent is "googlebot"
[Vincent] joined the channel
#
@kevinmarks ↩️ @polm23 You could scan your server logs for referers and send webmentions for them, but that is assuming that no https is involved. (twitter.com/_/status/1052116858785189888)
KevinMarks_ and KevinMarks joined the channel
#
@polm23 ↩️ @kevinmarks I'm talking about receiving webmentions, not sure what HTTPS or referers have to do with that. (twitter.com/_/status/1052132362958467072)
swentel joined the channel
#
petermolnar !tell snarfed I know you don't like these things, because they are dangerous to use and can easily break, but this actually works: https://github.com/ping/instagram_private_api
#
Loqi Ok, I'll tell them that when I see them next
#
Loqi [ping] instagram_private_api: A Python library to access Instagram's private API.
gRegorLove_, bradenslen, [jgmac1106] and [eddie] joined the channel
#
[eddie] tantek Zegnat my thought is if someone wants to build a static site, most of the time it’s for simplicity of maintenance. I could be wrong but I don’t think many static site people would want to fool with scanning their server logs
#
petermolnar I think it's much simpler/easier/even better to rely on things like webmention.io for this
#
aaronpk That tweet sounds like a misunderstanding of webmention
#
@kevinmarks ↩️ Presumably you want to respond to webmentions with some kind of action. You could use a 3rd party webmention handler like http://webmention.herokuapp.com or http://webmention.io or http://mention.tech or code your own. Once you do that, you could do the log parsing I suggested (twitter.com/_/status/1052169045896830977)
[kevinmarks] joined the channel
#
[kevinmarks] lets see if that helps
[Rose], [jgmac1106] and [pfefferle] joined the channel
#
@aaronpk ↩️ Static sites can already scan HTTP server logs for mentions. Webmention enables the sender to notify the receiver immediately instead of waiting for someone to click the link, and also enables static sites to offload receiving to an external service like https://webmention.io (twitter.com/_/status/1052185153160658945)
#
sknebel I think their point was that if the mention Details where in the URL you could read the mentions out of the server log and could do all processing offline, later, without any dynamic code or service
#
aaronpk wow
#
aaronpk did not get that
#
sknebel Welcome to Germany btw :)
#
aaronpk 🎉
[johnjohnston] joined the channel
#
[jgmac1106] Anyone see this error
#
aaronpk argh I gotta fix images in the slack bridge
#
[jgmac1106] Sorry, didn't want to try and retype error message
[Vincent] joined the channel
#
[Vincent] This is how I am (about, just need some time to nix a few bugs and push up the code) to handle sending webmentions on my static site, if it helps!
#
[Vincent] I build a custom JSON file when the website regenerates, which contains the last time the service was processed and an entry for each new article to post (containing source/destination).
#
[Vincent] Basic flow
#
[Vincent] 2. On build, trigger a webhook (netlify)
#
[Vincent] 1. Generate JSON file on build
#
Loqi http://meme.loqi.me/m/4x0ktNf4.jpg
#
[Vincent] 4. Node server loops through JSON file and sends each webmention
#
[Vincent] 3. Pings my Node server to go get the file from Github API.
#
[Vincent] 5. Node server updates the JSON file to current date/time and deletes other entries.
#
[Vincent] 6. Node server pushes the file back in to the Github API.
#
[Vincent] 7. Server will trigger another rebuild now, but do nothing since there are no new webmentions.
#
aaronpk reposting screenshot https://media.aaronpk.com/Screen-Shot-2018-10-16-15-27-55-dakx4hRxeB.jpg
#
[Vincent] lol Loqi
#
petermolnar [Vincent]: interesting approach; I ended up generating a file per outgoing webmention, similar to how incoming is handled, so no double-sends; btw webmention.io can trigger webhooks on incoming webmention with a nice payload of the webmention itself
#
[jgmac1106] I added my Known site to Indigenous as well as my WP. Syndication to Twitter works on Known for the web but not through Indigenous..
#
petermolnar what is Indigenous?
#
Loqi Indigenous is a set of native apps for iOS and Android currently in beta that supports posting your website using Micropub and a built-in reader that supports Microsub https://indieweb.org/Indigenous
#
jeremycherfas I need to rename a bunch of folders from 01.a-few-words to a-few-words. I can find the first part ok with [\d\d\.] but I am not sure how to save all the rest, after the first target, to the end of the line, in order to do the replacement. Help, please.
[Rose] and sebsel joined the channel
#
[Rose] I actually have a script which creates the XX. at the start of the name in the shell
#
jeremycherfas I want to get rid of it, cluttering up my Grav pages.
#
jeremycherfas I think it is something like $p I need, but so far trial and error is not helping me.
#
cweiske $ rename 's/^\d+.//' *.*
#
cweiske (backup first!)
#
[Rose] I prefer the XX. naming, it makes it easier to find things which aren't date named - that's me though.
#
[Vincent] @petermolnar I didn’t realise webmention.io could do that, I’ll take a look thanks!
#
jeremycherfas I have `([\d\d\.])(\S+$)` but the second group matches the 2nd and 3rd characters too.
#
sebsel jeremycherfas that's because of the []
#
sebsel that means 'one of these'
#
jeremycherfas Yay! Thank you sebsel.
#
Loqi woot
#
petermolnar [Vincent]: log in to webmention.io then go to https://webmention.io/settings - you'll see a Web Hook heading
#
Zegnat I like to throw regex into Regex 101
#
Zegnat https://regex101.co
#
Zegnat m
#
Zegnat which tries to give you an explanation of what will be parsed in English rather than computer speak, jeremycherfa
#
Zegnat Wow... Something got pasted wrong there with the URL.
#
[Vincent] Thanks 🙂
jgmac1106_, [chegalabonga] and snarfed joined the channel
#
jeremycherfas Thanks zegnat; I love that site, but couldn't find the problem with what I had.
jgmac1106_, KevinMarks, KartikPrabhu, sebsel and swentel joined the channel
#
swentel mm interesting, I see my posts now in aperture
snarfed, [tantek], jackjamieson, [Rose], [jgmac1106] and [pfefferle] joined the channel
#
snarfed swentel: your patch in https://github.com/snarfed/bridgy-fed/issues/34#issuecomment-430331815 looks good! thank you!
#
Loqi snarfed: petermolnar left you a message 7 hours, 56 minutes ago: I know you don't like these things, because they are dangerous to use and can easily break, but this actually works: https://github.com/ping/instagram_private_api
#
Loqi [swentel] So, I gave it a shot. I'm entirely not sure whether mention['href'] = target.get('attributedTo') is right, but I think the rest should be fine. Didn't want to create a pull request, because I couldn't even run it (at least, all the code in webmention...
#
snarfed i'd love to get you back up and running. remind me where you're stuck right now?
#
snarfed ...at least, if you want to. we could also just run with this split, where you do the hard work to figure out what to send, and i blindly put that in the code :P
#
swentel well, figuring out is nice, I like doing that :)
#
snarfed petermolnar: hah. i'm glad they exist! i expect that one's "web api" is similar to granary's current scraping. the "app api" is scarier, since it needs the user's password. i hate to train users to give out passwords or cookies more broadly
#
snarfed swentel++ heh ok!
#
Loqi swentel has 13 karma in this channel over the last year (25 in all channels)
#
swentel snarfed, for some reason, authentication/signature fails on my instance
#
snarfed swentel: against real mastodon instances? or your own that you run locally?
#
swentel against real ones
#
snarfed ok. so you don't have any way right now to test against actual mastodon? ok if so, just checking.
#
swentel no, not from my local bridgy fed
#
swentel I'm able to send AP requests from PHP though
#
swentel that's why I'm able to figure out at least what requests should contain
#
snarfed oh! right. that's good enough then.
#
swentel that's a big thing for me
#
snarfed yes! you don't really need to test from your local bridgy fed then.
#
snarfed if you do want to just use the unit tests to test bridgy fed code, i'd still happily accept PRs
#
swentel I can't get those working either :/
#
snarfed but i'm also happy to write the code myself. it's pretty quick. eitherw ay
#
swentel I forget the error messages
#
swentel but I know we tried debugging them at some point
#
snarfed no problem. i'm happy to help if you want. or we can keep doing this. either way!
#
snarfed i'll put in the mention now
#
swentel oh cool :)
#
swentel I'm going to debug it at some point again
#
swentel even just seeing requests
#
swentel and adding logging
#
swentel so I see the objects
#
swentel makes it probably a bit easier to actually write code that should work :)
#
snarfed definitely!
#
swentel I'm going to try that with the accept functionality (when a follow request comes in)
#
swentel snarfed, do you use a specific IDE to write python in ?
#
swentel that would help me speed up as well, I sometimes need to lookup stupid basic things haha
#
snarfed great! i just emacs. editor and tools are very personal. if you have an editor you like, stick with it
#
sknebel PyCharm by jetbrains is the best "full IDE" for Python
#
snarfed i like dash for reading docs if you're on mac
jgmac1106_ joined the channel
#
swentel on linux
#
swentel vim has some basic highlighting, will check whether there's some plugins
[cleverdevil] joined the channel
#
snarfed swentel: https://github.com/snarfed/bridgy-fed/pull/36
#
Loqi [snarfed] #36 add Mention tag for Mastodon replies
#
snarfed oops, fixing a bug
#
snarfed done
#
swentel cool !
#
snarfed feel free to review if you want, otherwise i'll merge
#
swentel look fine to me
#
swentel good to lear about .setdefault and such
#
swentel learn
[asuh] joined the channel
#
snarfed great! i've merged and deployed, feel free to test
#
swentel ok, checking
#
swentel yes!
#
swentel awesome, thanks :)
#
swentel snarfed++
#
Loqi snarfed has 70 karma in this channel over the last year (122 in all channels)
#
swentel it's funny how other users are testing on that thread heh
#
snarfed hah yes
#
swentel hmm getting No module named future_builtins again, it's really weird, sometimes it runs fine, sometimes 500 :/
#
snarfed bleh
bradenslen, dougbeal|mb1 and KevinMarks_ joined the channel
#
swentel a PR (but a very easy one heh)
#
snarfed thanks! i'll take it
#
snarfed deployed!
#
swentel ah crap, did a wrong merge in my fork on github
#
swentel I sometimes really hate that workflow
#
swentel oh well
#
snarfed git is the worst
#
AngeloGladding aaronpk HEAD requests to my homepage were returning 204 No Content and Aperture wanted 200 OK -- that was painful
#
dougbeal snarfed: I want post-git posthaste
#
snarfed googles
jaggendaggen and KevinMarks joined the channel
#
aaronpk AngeloGladding: 204 sounds like the wrong response to return for a HEAD request if it returns 200 for a GET request
#
AngeloGladding aaahhh you're right
[pfefferle], [eddie] and eli_oat joined the channel
#
swentel aha!
#
swentel snarfed, so I looked at the commit history
#
swentel and you changed something in app.yml
#
swentel +#includes:
#
swentel +# - local/lib/python2.7/site-packages/oauth_dropins/webutil/app.common.yaml
#
swentel when uncommented, I can test again on my local fed instance
#
swentel (don't ask my why though)
[Rose] joined the channel
#
snarfed swentel: wait what? that looks uncommented in master to me. i don't actually see in history that it was ever commented out
#
snarfed https://github.com/snarfed/bridgy-fed/blob/master/app.yaml#L10-L11
#
snarfed (nice find though, regardless!)
#
swentel yeah the paste is already what I did
#
swentel hmm, it's an old commit
#
swentel so I wonder why it matters now
#
swentel oh well, I can paste
#
swentel euh post
#
snarfed ah i see. huh. very weird. that included file doesn't do much. regardless, glad you found it!
#
snarfed hey swentel, just curious, if you run `gcloud info`, which version does it say you're on? the very first output line
#
swentel swentel@swenteldoos:/home/other/bridgy-fed$ gcloud info
#
swentel Google Cloud SDK [217.0.0]
#
snarfed hmm ok i think that's recent enough for 'includes'. shrug
[eddie] and KartikPrabhu joined the channel
#
swentel ok, bed time
#
swentel snarfed, posted a diff which should work for accept, but ignore for today, it's time to sleep - at least for me :)
#
snarfed swentel++ will do!
#
Loqi swentel has 14 karma in this channel over the last year (26 in all channels)
#
@jgmac1106 How to Add to the Zombie Roll in Webmentions vs Zombies: http://youtu.be/4E9W_znT8rc?a via @YouTube (twitter.com/_/status/1052315626088386560)
[schmarty], [tantek], [eddie] and KevinMarks joined the channel