#social 2019-03-08

2019-03-08 UTC
vitalyster left the channel
#
dansup
aaronpk: are you using guzzle?
#
dansup
or curl
#
dansup
I need to update pixelfed to use concurrent requests, http://docs.guzzlephp.org/en/stable/quickstart.html#concurrent-requests
#
aaronpk
I think I'm using curl directly
#
aaronpk
that's not the problem tho
timbl, ben_thatmustbeme, Guest84_ and xmpp-social joined the channel
#
dansup
aaronpk: I just shipped async delivery with guzzle pools! I think this is the first time I've used yield/Generators, heh https://github.com/pixelfed/pixelfed/blob/dev/app/Jobs/StatusPipeline/StatusActivityPubDeliver.php#L73-L102
#
dansup
It used to take up to 10 minutes to deliver a status to 210 distinct inboxes, now it takes 34 seconds with a concurrency limit of 10 and timeout of 30 seconds
vitalyster, Guest84, dmitriz, dmitriz1, Chocobozzz and timbl joined the channel
#
aaronpk
dansup: nice! Are those running in the fpm process or are you using a background worker?
#
cwebber2
second federation episode of librelounge out
vitalyster joined the channel; vitalyster left the channel
#
jdormit[m]
cwebber2: should I listen to the first federation episode first?
#
fr33domlover
cwebber2 / everyone, I'm wondering, if using ocap authorization tokens / capability invocations, how does a server decide when to send a token and which one? If I have a token giving access to some resource, do I just always send the token when delivering an activity to that actor etc.?
#
dansup
aaronpk: those are running in the horizon job queue, I'm using the redis queue driver.
#
dansup
cwebber2: Nice! Listening now :)
#
dansup
jdormit[m]: can you join #pixelfed-dev on freenode? kaniini and I are trying to debug the pleroma/pixelfed/pterotype following issue
#
dansup
nvm, kaniini found it
timbl_ and timbl joined the channel
#
jdormit[m]
They did, and opened an issue! Was it the same issue in PixelFed?