#dev 2021-06-01

2021-06-01 UTC
rrix joined the channel
#
superkuh Contents for who can build the indieweb functionality with the most corporate-web dependent parts?
#
superkuh Er, Contest.
[snarfed] joined the channel
#
[snarfed] anyone know of a decent video caching proxy service? we have https://indieweb.org/image_proxy but I haven’t found any similar resources for video
#
[snarfed] notably, i’m looking for one that does pass through fetching to remote origins, eg you can put the URL in the request, something like https://cdn.io/proxy/http://my/video
#
aaronpk oh wow
#
aaronpk would this be for short video clips like you’d see on twitter? or long videos like youtube?
#
[snarfed] arbitrary, my main use case right now is for facebook-atom and instagram-atom. they recently started using tokens in their image and video URLs with extremely short expirations, so they’d always fail when I saw them in my reader hours later
#
[snarfed] I added a caching proxy for the images with a free tier, but haven’t found one for video
#
[snarfed] Cloudinary is almost there, but I haven’t found a way to set the cache TTL or otherwise age them out, so it keeps them forever and eventually starts charging me
#
[snarfed] https://cloudinary.com/documentation/fetch_remote_images#comparing_fetch_to_auto_upload
#
[snarfed] I also need to start using the caching proxy for FB and IG profile pictures in Bridgy, for the same reason 😠
#
aaronpk oof
#
aaronpk doesn’t that imply the caching proxy would need to be aware of which query string parameters are expiring tokens?
samwilson and tomlarkworthy joined the channel
#
tomlarkworthy I think the idea is you cache the content while the token is active, and return the same content after the token expires. Kinda exploits that there is no VARY headers for URL parameters though, its a loophole.
#
tomlarkworthy one would also think the cache TTL should expire with the token URL anyway. You would need a heavily customizable cache to sidestep link expiry
#
tomlarkworthy heavily (mis) configured
[capjamesg], barnaby, ben_thatmustbeme, [jacky], shoesNsocks1, [Murray] and [snarfed] joined the channel
#
[snarfed] right, this is an abuse/misuse of what FB/IG intend. many caching proxies are flexible and let you configure how they obey headers (or not), which this needs
#
aaronpk hacks on some code to move sleep data importing to pull from apple health so he can finally ditch the fitbit
#
aaronpk hmm looks like there is no per-minute data :-/
#
@AndreJaenisch ↩️ What is this namespace thingy? In https://github.com/ryuno-ki/webmention-tools my code was shallow enough to install the whole package. But that wouldn't allow to install a subset from it (your „pip install http://rich.tui“) (twitter.com/_/status/1399775708235419655)
Loqi, stevestreza, deeden, barnaby, [chrisaldrich], [schmarty], [fluffy], jjuran and [KevinMarks] joined the channel