#dev 2018-02-12

2018-02-12 UTC
dgy_ and dgy__ joined the channel
#
snarfed shipped: bridgy publish starring github repos with u-like-of. feel free to try! https://brid.gy/
#
snarfed it also renders full github-flavored markdown in previews now
[miklb] joined the channel
#
KartikPrabhu snarfed++ for bridgying with github
#
Loqi snarfed has 34 karma in this channel (342 overall)
#
snarfed ugh, broke it. back in a sec :P
#
KartikPrabhu snarfed -- for breaking the bridgy
#
snarfed lol
#
KartikPrabhu ;)
dgy_ joined the channel
#
snarfed fixed
leg, snarfed and mblaney joined the channel
#
mblaney unlikes bridgy to like bridgy using bridgy...
#
mblaney woot it works!
#
Loqi 😊
#
KartikPrabhu lol!
#
KartikPrabhu bridygception
#
@miklb @beep very cool. Also, while snooping in your code, I see you’ve added webmentions, which is awesome. How long have you had them enabled? (twitter.com/_/status/962852434208067584)
AngeloGladding and [kevinmarks] joined the channel
#
[kevinmarks] Remember me going on about hashes? https://scotthelme.co.uk/protect-site-from-cyrptojacking-csp-sri/
AngeloGladding and tantek joined the channel
#
loqi.me created /Favrd (+256) "prompted by tantek and dfn added by tantek" (view diff)
#
kaja.sknebel.net edited /Favrd (+1) "linkify ('x is y. <url>.' pattern)" (view diff)
#
tantek close enough
[eddie] joined the channel
#
[eddie] !tell aaronpk: I got a Micropub post to Aperture to work except the content. I did a content attribute inside of properties, with an array of objects. It contained an html and a text attribute which both contained a string. Aperture is returning all my properties except no content attribute.
#
Loqi Ok, I'll tell them that when I see them next
#
[eddie] !tell aaronpk: I’m assuming I did something wrong? I sent it via JSON
#
Loqi Ok, I'll tell them that when I see them next
tantek joined the channel
#
GWG Once again getting sample data for future location display
#
aaronpk Oh yeah I did that too
#
Loqi aaronpk: [eddie] left you a message 9 minutes ago: I got a Micropub post to Aperture to work except the content. I did a content attribute inside of properties, with an array of objects. It contained an html and a text attribute which both contained a string. Aperture is returning all my properties except no content attribute.
#
Loqi aaronpk: [eddie] left you a message 9 minutes ago: I’m assuming I did something wrong? I sent it via JSON
#
aaronpk check the Microformats docs again ;-)
#
[eddie] Ohhh value and HTML?
#
[eddie] Bingo!
#
[eddie] Lol tricky mf2
#
[eddie] gives Loqi mf2
#
tantek gives Loqi a blank stare
#
Loqi grabs the blank stare
#
[eddie] Darn you Loqi!
#
[eddie] Lol ever elusive
#
KartikPrabhu gives Loqi tricky mf2
#
KartikPrabhu gives Loqi a tricky mf2
#
Loqi jumps on the tricky mf2
#
KartikPrabhu good job Loqi
#
KartikPrabhu it is weird that Loqi only accepts one object at a time
tantek joined the channel
#
@TeamWembassy @Dries does your blog support pingback / webmentions yet? (twitter.com/_/status/962919320236867584)
barpthewire and [kevinmarks] joined the channel
#
[kevinmarks] gives loqi a media endpoint
#
Loqi eyes the media endpoint
jjuran, jeremycherfas, [gerwitz] and [kevinmarks] joined the channel
#
@TeamWembassy @stevepurkiss yea i get it i think it serves a good purpose. my goal is to have something super simple and specific use case, kind of a drop-in and go sort of deal. I probably will lean on them for things and give them some of what im doing with Webmentions as well. (twitter.com/_/status/962995099675209729)
[kevinmarks], [tantek] and tantek joined the channel
#
tantek good morning #indieweb-dev
#
GWG Good morning, tantek
#
tantek I am starting to implement POSSE to GitHub via Bridgy
Kyle-K and leg joined the channel
#
Zegnat https << “Beginning in July 2018 with the release of Chrome 68, Chrome will mark all HTTP sites as “not secure”.” - https://security.googleblog.com/2018/02/a-secure-web-is-here-to-stay.html
#
Loqi ok, I added "“Beginning in July 2018 with the release of Chrome 68, Chrome will mark all HTTP sites as “not secure”.” - https://security.googleblog.com/2018/02/a-secure-web-is-here-to-stay.html" to the "See Also" section of /HTTPS
#
tantek I'll believe it when I see it
#
tantek still likely to trigger the "alarm fatigue" problem. a red light that is always (or frequently) on tends to get ignored
#
tantek I'm considering hiding all replies from my atom feed
#
tantek I don't think they make sense out of context
#
tantek and old style feed readers (the ones that only support sidefiles) rarely (if ever?) show any context
#
tantek or have any knowledge of replies
#
Zegnat It’ll be interesting to see the reaction of the public. The reaction of commercial websites might be to get rid of the warning ASAP, which is probably exactly what Google wants
#
tantek Does anyone else here omit replies from the Atom/RSS feeds?
#
tantek (but keep them on their h-feed obviously)
#
tantek (this would basically drop all "respones", likes, reposts, etc. from side files)
#
tantek because AFAIK, no "old" feed readers do anything intelligent with those anyway
#
tantek so presumably they show up as noise for the folks using such software
#
sknebel I know many hide them from the homepage feed by default, not sure if anyone treats atom differently (well, many people use granary, so those don't likely)
#
sknebel and if your post has even a minimal context in textual format ("I like...", "Reply to") I don't think they are useless in a traditional reader
#
tantek the "Reply to" context is only on the permalink display
#
tantek I do put "likes ... " in the fallback summary
#
tantek until someone shows me a feed reader that *does not* support h-feed, AND yet somehow supports doing something useful with like/reply posts, I'm not going to worry about it
#
tantek going to try deploying it and see if anyone notices
#
tantek deployed
#
tantek this is part of my work toward implementing issues / replies to issues
#
tantek I've decided issues (like github issues) are too out of context to bother showing in feed files
#
tantek and since issues have names and content, I'm building them out of article posts that are in reply to a particular issues list
#
tantek I'm also omitting them from my "articles box" on my home page
eli_oat, kapowaz, tantek and [gerwitz] joined the channel
#
[gerwitz] +1 FWIW … replies and reactions require the context of a content resource. I don’t like elevating them to be a peer to original posts. (Ev Williams is the smartest-on-this person I know who disagrees.)
jeremych_, jeremycherfas, [kevinmarks], AngeloGladding and barpthewire joined the channel
#
aaronpk [eddie] I just pushed up a bunch of changes for read state tracking! I'm excited to try it out! let me know what you think of the api
snarfed joined the channel
#
aaronparecki.com edited /Microsub-spec (+340) "/* Channels */ add unread parameter to channel list. for https://github.com/indieweb/microsub/issues/4" (view diff)
[eddie] joined the channel
#
[eddie] aaronpk: Oh wow! I’ll take a look!
#
[eddie] aaronpk: Seems pretty good! I’ll let you know more when I actually get around to playing with it
#
aaronpk kewl
#
[eddie] Question: read/unread is per entry, per channel? So a post with the same url in two different feeds has two read states?
#
aaronpk correct
#
[eddie] Gotcha :thumbsup:
#
aaronparecki.com edited /Microsub-spec (+1094) "/* Channels */ channel ordering" (view diff)
snarfed, tantek, KartikPrabhu and snarfed1 joined the channel
#
tantek does anyone do /autoembed of Soundcloud track links?
#
aaronpk I used to
#
kaja.sknebel.net edited /User:Kaja.sknebel.net/upcoming-hwcs (+201) "update from /Events" (view diff)
#
tantek aaronpk: do you post Soundcloud track links in general?
#
aaronpk not much anymore
#
tantek it looks like a pretty straightforward iframe
#
aaronpk here is how I used to do it https://aaronparecki.com/2013/05/09/1/experimenting-with-auto-embedding-content##SoundCloud
#
tantek hmm except the src needs a trackID whereas most soundcloud links have a human friendly path
#
Loqi [Aaron Parecki] Experimenting with auto-embedding content
snarfed joined the channel
#
schmarty tantek: i used to autoembed soundcloud links. generating an embed iframe from a track's browser-readable URL required an extra processing step to get the trackID, as you noted.
#
KartikPrabhu tantek: I think I do something similar for youtube and vimeo and Flickr
#
KartikPrabhu extract some sort of ID from the URL and then embed it
#
schmarty oops. by processing step i meant fetching the page to find the trackID embedded somewhere in it.
#
KartikPrabhu schmarty: oh! the URL itself does not give you that?
#
aaronpk their oembed API will give you the info you need
#
KartikPrabhu <sigh> if something is in the API it can on the page no?
#
schmarty KartikPrabhu, aaronpk: i hadn't thought to try oembed, so i fetched the page and looked for a proprietary meta tag containing a soundcloud: url
#
KartikPrabhu oh boy!
#
schmarty something like soundcloud://sounds:TRACKID
#
schmarty not the best! i have since dropped the hack and decided to simply not show soundcloud embeds automatically for now. :}
#
KartikPrabhu yeah that whole thing sounds annoying
jackjamieson joined the channel
#
tantek yeah too annoying to bother with
[cleverdevil] joined the channel
#
snarfed [cleverdevil]: re AWS and Athena (moving here...)
#
snarfed AWS definitely has an equivalent competitor in GCP. migrating either direction is totally doable, esp if you're using basic services like EC2, S3, etc
#
[cleverdevil] Ah, cool, I thought here might be more appropriate 😉
#
[cleverdevil] Indeed.
#
snarfed and Azure is fairly close too (but not quite as close)
#
[cleverdevil] The base object store side is totally commoditized.
#
snarfed athena is an odd choice though. do you maybe mean aurora/
#
[cleverdevil] Nope.
#
snarfed oh nm i see
#
aaronpk athena looks really interesting
#
[cleverdevil] Athena allows you to query S3 objects via SQL.
#
snarfed i evaluated athena a bit ago. definitely interesting, but pretty limited, at least as a data warehouse
#
[cleverdevil] Including JSON objects.
#
[cleverdevil] I think for the purposes of an IndieWeb site, it could be totally adequate.
#
snarfed maybe better for OLTP (like webapps) ...but i dunno
#
[cleverdevil] This all came from my desire to move off of MySQL and store all of my data in JSON files on disk.
leg joined the channel
#
[cleverdevil] But, then I started thinking about search/query, and how I'd have to do all of this work to build indexes, etc.
#
snarfed what are your search/indexing use cases?
#
[cleverdevil] The appeal of having all of my data on S3, which is super redundant, reliable, and distributed is pretty enormous.
#
snarfed honestly for a personal web site, you can probably get by with just filesystem dates, simple linear search, etc, and maybe google site search or aws's hosted elastic for site search if you really need it
#
aaronpk I wonder if I could use that for indexing the data in compass
#
snarfed yup, S3 is nice
#
[cleverdevil] So, I use Known for my website, and I'd likely want to implement this for Known as a storage backend.
#
[cleverdevil] It only supports a very, very limited set of queries out of the box.
#
aaronpk for my site I implemented indexes in mysql because I wanted things like tag pages
#
snarfed S3 will be totally straightforward. athena will be a snowflake. maybe start with just S3 and see
#
[cleverdevil] Essentially, it lets you fetch posts by "type," with date and count limits, and basic sorting.
#
[cleverdevil] I added "<" and ">" for dates, to implement my "On this Day" page.
#
[cleverdevil] That's basically it.
#
[cleverdevil] Other nice thing about this is that I could "publish" new things directly by copying things to S3 from anywhere.
#
[cleverdevil] Rather than having to worry about Known's Micropub endpoint, and constantly fighting with extending it to support more things.
#
snarfed known already has a flat file storage backend right? easy quick MVP would be that + S3 FUSE filesystem, no new code needed, at least to start
#
[cleverdevil] Nope, it just has MySQL and MongoDB.
#
[cleverdevil] But, since MongoDB came first in Known's history, the MySQL database is basically a bunch of tables with indexed columns that also have the JSON document embedded.
#
snarfed huh ok. maybe start with a simple flat file backend then, so it's usable both on and off S3
#
[cleverdevil] Yeah, probably wise.
#
[cleverdevil] Mostly just for saving and retrieving by ID.
#
aaronpk [eddie] if you add a syndication link from your indie post to your github post, i'll be able to reply to your indie post and have my reply also syndicated to github :)
#
aaronpk whoa, my tags got added as github labels!
#
aaronpk silopub++
#
Loqi silopub has 1 karma in this channel (3 overall)
#
aaronpk although I don't really want them as github labels
#
snarfed heh
[kevinmarks] and tantek joined the channel
#
@miklb @beep understood. Are you using @AaronGustafson’s Jekyll plugin for webmentions/webmention.io? If not, may I suggest a look? (twitter.com/_/status/963119687902158849)
tantek and ben_thatmustbeme joined the channel
#
aaronpk *whew* k that's all the microsub/aperture work for me for the day
[cleverdevil] joined the channel
#
[cleverdevil] You've been making a ton of progress.
#
[cleverdevil] 🙂
snarfed joined the channel
#
[cleverdevil] https://aws.amazon.com/blogs/aws/s3-glacier-select/
#
[cleverdevil] But, no querying afaict across objects.
[eddie] joined the channel
#
[eddie] aaronpk: I definitely want to get syndication links added, but that will probably happen once I update how my site syndication works.
#
[eddie] Right now I syndicate by posting to my site, updating jekyll and then manually sending a Webmention to bridgy publish through Telegraph. Which means until I get automated webmention/syndication logic in my site, I would have to copy the url returned on telegraph, open the post on my server and manually paste the syndication link in.
#
aaronpk aha yeah
#
[eddie] But fixing up my syndication story is definitely on my list
#
[eddie] So one day that’ll be better 🙂
#
aaronpk also: https://aaronparecki.com/2018/02/12/18/ 😂
#
Loqi [Aaron Parecki] Channel icons?
#
[eddie] Haha, that’s awesome!
#
Loqi awesome
#
schmarty everyone is having so much fun with aperture / indigenous
#
aaronpk it's pretty great. want to join the party?
#
dgold its pretty excellent fun, schmarty. move fast and break things
#
schmarty is really tempted to get an aperture deployment going
#
aaronpk if you want to work on a client i'm happy to let you use my aperture
#
loqi.me edited /silo-quits (+111) "tantek added "~2018-02-05 @rafaorr quit [[Facebook]] per 2018-02-12 https://twitter.com/rafaorrr/status/963118776618188803" to "See Also"" (view diff)
#
tantek aaropnk - that's a misuse of the Warning emoji for Notifications
snarfed joined the channel
#
tantek unless you really mean it to be warning alerts like something might be wrong
#
aaronpk hey I can do whatever I want :P
#
aaronpk it's just the display name of the channel, it's not like it's in the spec
#
tantek of course you can, just thought you'd care about the misuse
#
tantek but now it's on your blog, which people take seriously :)
#
aaronpk is there a better emoji? I couldn't think of one
#
tantek yes one sec
#
aaronpk bell?
#
tantek ok so this is my fault, apologies
#
schmarty i've been using tt-rss + woodwind for a bit and am ready to move past them. this flurry of new microsub stuff makes me feel like it'd work out for me okay. :}
#
tantek bell could work (that's what FB uses now instead of the globe!)
#
tantek I think ℹ is nice and neutral
#
aaronpk I have the globe still on facebook
#
tantek 🔔
#
tantek orly
#
tantek even on mobile web?
#
[eddie] “backdate entries when adding a new feed to a channel” Blast! Right after I added several new feeds today. Haha
#
aaronpk whoa, no on mobile it's the bell
#
tantek.com created /🔔 (+26) "r" (view diff)
#
aaronpk [eddie]: lol! well you can remove them and re-add them!
#
tantek.com edited /notification (+1) "bell is better than warning" (view diff)
#
aaronpk I made the scary red button work
#
[eddie] Lol probably not that important. But that’s true!
#
tantek.com edited /⚠ (-5) "until it's there" (view diff)
#
tantek aaronpk: sorry about that: https://indieweb.org/wiki/index.php?title=notification&type=revision&diff=16300&oldid=16266
#
aaronpk aha that's probably why it was in the back of my mind
#
[eddie] Hopefully this code signing stuff is an easy fix with Apple so I can push out my bug fix release on Indigenous and start on support for some of the new Microsub read/unread stuff! Not knowing the state of stuff in my channels has been one of the biggest things messing with my Microsub use
#
aaronpk yeah I am looking forward to that for sure!
#
aaronpk I think next on my list is getting source info into entries in case there is no author info
#
[eddie] Nice!
#
aaronpk then I think i'm gonna start working on my UI
#
tantek.com edited /notification (+641) "/* Silo Examples */ note icon changes recently" (view diff)
#
tantek ok added more specifics about notification iconography in FB and IG for now
#
schmarty aaronpk: thanks! i don't think i'm ready to start work on a client yet. i'm interested in getting my hands dirty setting this stuff up to learn the state of things.
#
aaronpk kewl
[colinwalker] joined the channel
#
tantek Twitter also uses a bell
#
tantek once Twitter and FB agree on a UI convention it's pretty much a done deal
#
[eddie] schmarty: Feel free to reach out if you run into any issues. The two articles by dgold should help give you a general direction: https://ascraeus.org/websub-part-i-watchtower/ https://ascraeus.org/websub-part-ii-aperture/
#
Loqi [Daniel Goldsmith] WebSub Part I: Watchtower
#
Zegnat I think I'll wait for the spec to stabilise a bit more.
#
tantek.com edited /notification (+230) "/* Twitter */ uses a bell also" (view diff)
#
tantek whoa Twitter now has venue pages?!? (powered by Foursquare) e.g. https://twitter.com/search?q=place%3A07d9e38a8ac88001
#
aaronpk wat whoa
[jjdelc] joined the channel
#
[jjdelc] I got a question about bridgy and my identity. My identity auth site is www.domain but my blog and posts live under jj.domain so it looks like bridgy is never finding my new posts when scanning www. It is assuming that my posts will be on my identity page. or am I missing something? Only when I tell bridgy to crawl a specific post it can then post the mentions to it.
#
[jjdelc] both my identity page and blog index point to each other with rel=me, but there's nothing special about the link in my identity page to tell "here be the posts" and not any other outbound rel=me post
leg, [kevinmarks] and snarfed joined the channel
#
KartikPrabhu [jjdelc]: I think you can use rel-feed to point to the page that has a feed of posts
#
KartikPrabhu [jjdelc]: I do that on my site https://kartikprabhu.com/ see the navigation at the bottom
snarfed joined the channel
#
[jjdelc] ah, rel="feed"? I didn't think of that
#
KartikPrabhu yeah, I think bridgy uses that on my site to find older posts
#
gRegorLove Yep, Bridgy does that with my site
#
[jjdelc] it has to point to an rss/atom feed? or it can be a mf2 html doc?
#
gRegorLove https://brid.gy/about#link
#
gRegorLove HTML doc with h-entrys
#
gRegorLove My site is gregorlove.com. The "notes" link in navigation has rel="feed"
#
[jjdelc] perfect, that's what I needed
#
[jjdelc] I recall reading through that in the doc, but I didn't remember the rel=feed link
#
[jjdelc] Thanks a lot @KartikPrabhu and [gregorlove]
#
snarfed hi [jjdelc]! an easier way is to just add a link to jj.domain to your silo profile, then sign up again
#
snarfed that's where bridgy finds your web sites to crawl
#
snarfed https://brid.gy/about#profile-links
#
KartikPrabhu right ^. rel-feed is more useful if you have multiple pages with posts like separate Articles and Notes pages
#
[jjdelc] Ah, good that it searches in the desc fields so I can have it unlinked there but my main link in the url page for twitter/ig
#
[jjdelc] I'll try rel=feed first , it feels more protocolary
#
[jjdelc] Thanks everyone
#
snarfed i recommend adding it to your profile (and re-authing bridgy) anyway. it behaves subtly differently when it believes a web site is actually yours vs not
#
[jjdelc] 👍
#
[jjdelc] I have another question about the inclusion of webmentions in an entry. Should the list of webmentions be inside the h-entry? because when I parse the links in an h-entry I will find all the webmention links again, but my entry isn't really pointing to them. I've seen some implementations and the list of metsions is inside the h-entry.
#
[jjdelc] I that the links I parse in my posts should be inside my e-content to get around it
#
snarfed [jjdelc]: https://indieweb.org/comments#How_to_markup
#
KartikPrabhu [jjdelc]: I have mine inside the h-entry markedup as p-comment. But when I send my webmentions I only use e-content
#
[jjdelc] Ah so many pages to read, I didn't stop through that one yet
#
KartikPrabhu [jjdelc]: it is still ok to ask here since others might know which page to read :P
#
[jjdelc] @KartikPrabhu thanks, do you include your "u-bookmark-of", "u-like-of" likns inside your e-content as well?
#
KartikPrabhu oh! good point. I don't so I wonder what I'm doing for webmention sending :P
#
[jjdelc] Right now I have those outside my e-content because it is not content I've written (not my body), but a different attribute of the post, still inside h-entry but outside e-content
#
KartikPrabhu has forgotten his own code
#
[jjdelc] https://kartikprabhu.com/notes/like-find-your-own-way <- yes you keep it out of e-content too
#
Loqi [Kartik Prabhu] “If those were the only two stories about writers that you know, you’d never start. Find your own way.”
#
KartikPrabhu [jjdelc]: yeah I do. so I am confused how my webmention sender works! :P
#
aaronpk wow my aws bandwidth spiked again this month
#
snarfed aaronpk: i feel your pain! (w/huffduff-video)
#
aaronpk cloudflare has been doing an excellent job of serving about 90% cached requests, so i'm not sure what happened
#
aaronpk I always have way too much of my linode bandwidth quota available every month so I might move off of AWS back to that
#
[jjdelc] @KartikPrabhu I am in the same situation right now, I'm adding the special bookmark, like links as extra on top of what e-content has inside.
#
aaronpk from AWS: 1.681 GB in December, 72.575 GB in January. maybe cloudflare did a cache purge?
#
KartikPrabhu what is ronkyuu?
#
Loqi ronkyuu is a Python open source project that is being designed to help with Webmention requirements for static sites https://indieweb.org/ronkyuu
#
KartikPrabhu [jjdelc]: I use (possibly an older version) of ronkyuu so I will have to see if I am handling that at all!
#
aaronpk 497,478 Requests in dec, 4,662,996 Requests in jan
#
KartikPrabhu wow! aaronpk
#
aaronpk all this just to serve tiny copies of ppls profile photos on my posts :P
#
aaronpk you're welcome everyone else for not hotlinking those :P
#
KartikPrabhu ha!
#
KartikPrabhu but then those requests would be distributed over the network so....
#
[jjdelc] ha, well you don't want to get pwned like what the oatmeal made when he got hotlinked 😛
#
snarfed aaronpk: so that's $7ish of bandwidth for the month? noticeable, but maybe not enough to warrant a migration?
#
aaronpk went from $2 to $10 this month
#
snarfed hmm. aws outbound data is $.09/G in us west, so $7 bandwidth + $3 something else? ec2?
#
aaronpk bandwidth plus requests
#
snarfed right
#
aaronpk oh yeah, this is not just my site, it's also the webmention.io avatars. maybe someone with a lot of traffic recently started showing webmentions from there
#
sknebel always surprised how expensive AWS bandwidth is
#
snarfed huffduff-video has been serving multiple TB per month recently. :( i fixed robots.txt and blocked one bot recently, helped so far, but still.
#
snarfed https://github.com/snarfed/huffduff-video/blob/master/README.md#understanding-bandwidth-usage
#
aaronpk wow
#
aaronpk have you asked adactio about huffduffer?
#
sknebel aaronpk: curious, how much are you paying for storage for all those/how much data is it?
#
aaronpk looks like about 3.5gb of storage
#
aaronpk most of the cost is the bandwidth and number of requests. the 4.6 million requests last month cost $1.87, and 72gb of bandwidth cost $6.53
#
snarfed aaronpk: asked adactio what? he knows about hdv, naturally
#
aaronpk about why it downloads a copy
#
snarfed ahhhhhhh
#
sknebel so if that keeps up it would be cheaper to get a small linode instance to serve them instead :D
#
snarfed no, but good point!
#
aaronpk sknebel: yeah i'm gonna wait til the end of this billing cycle and see if january was just an outlier. but if this keeps up then i'll probably move them to a linode
#
snarfed aaronpk: do you serve the files? or does S3?
#
aaronpk cloudflare does, with s3 as the backend
#
aaronpk er no, cloudflare serves mine. webmention.io's are direclty served from s3
#
snarfed right, and the s3 serving is what you're noticing here. no need to wait for another bill, you can turn on (almost) realtime logs and calculate the rate pretty quickly. details and shell snippets in https://github.com/snarfed/huffduff-video#understanding-bandwidth-usage
#
Loqi [snarfed] huffduff-video: Extract the audio from videos on YouTube, Vimeo, and other sites and send it to Huffduffer.
#
aaronpk ah true
#
aaronpk I don't think I care that much to go through the trouble :P
AngeloGladding and [miklb] joined the channel
#
[miklb] lol
#
snarfed yeah, at $7, good call
tantek joined the channel
#
aaronpk speaking of things that are barely worth the trouble, I want to delete everything from aws glacier that I was experimenting with for a while. but you can only delete the vault if it's empty, and there is no UI to empty it.
#
aaronpk this is costing me $0.30/month but it's a little irritating
#
tantek that is annoying
#
KartikPrabhu that sounds like a bad buisness practice
#
tantek literal monthly nickle & diming (2 of each)
#
aaronpk 😂
#
tantek is it called a glacier because it slowly encroaches on your bank account?
#
tantek or (global warming version) because it causes your bank account to slowly recede?
#
KartikPrabhu lol
#
KartikPrabhu maybe because it is creepy
#
aaronpk wow, glacier/vault really is an appropriate name... "Inventory not ready, sleep for 10 mins..."
#
aaronpk I found a script to deal with it. it has to first retrieve the vault before it can delete it
#
aaronpk I wouldn't be surprised if someone is literally fetching a hard drive right now to install it somewhere for me
#
KartikPrabhu aaronpk: might want to document this process in a post
#
tantek while listening to the menamena song
#
KartikPrabhu read that as numanuma
#
sknebel it's probably a robot doing it, but something like that (is it public what glacier is based on? I remember there being tons of speculation at some point, but not sure if it ever was clarified)
#
tantek sknebel, it could be a robot literally fetching a hard drive
#
sknebel I know
#
sknebel or a tape
#
snarfed afaik it is. tape, but otherwise ys
#
snarfed heh
#
sknebel I think the most common assumption is tape
#
tantek KartikPrabhu: I was close, it's Mahna Mahna
#
tantek https://www.youtube.com/watch?v=X4jbq9SVxAc
#
tantek you're welcome
#
KartikPrabhu tantek: I was thinking of this meme from way back https://www.youtube.com/watch?v=KmtzQCSh6xk
[kevinmarks], snarfed and [cleverdevil] joined the channel
#
[cleverdevil] [aaron_pk] the secret to deleting things from s3/glacier is using bucket lifecycles.
#
[cleverdevil] Just tell the bucket to expire the data yesterday.
#
[cleverdevil] Then, it'll automatically delete all the data for you.
#
[cleverdevil] You'll then be able to delete the bucket.
#
loqi.me edited /Facebook (+374) "tantek added "2018-02-12 CNN: [http://money.cnn.com/2018/02/12/media/unilever-advertising-facebook-google-swamp/index.html Unilever to Facebook and Google: Clean up 'swamp' or we'll pull ads] <blockquote>"We cannot continue to prop up a digital supply..." (view diff)
snarfed joined the channel
#
aaronpk [cleverdevil] that sounds promising, but I cant find any expiration settings in Glacier