ChanServ changed the topic of #freedesktop to: https://www.freedesktop.org infrastructure and online services || for questions about freedesktop.org projects, please see each project's contact || for discussions about specifications, please use https://gitlab.freedesktop.org/xdg or xdg@lists.freedesktop.org
fantom has quit [Ping timeout: 480 seconds]
haaninjo has quit [Quit: Ex-Chat]
scrumplex has joined #freedesktop
scrumplex_ has quit [Ping timeout: 480 seconds]
georgc has joined #freedesktop
gchini has quit [Ping timeout: 480 seconds]
guludo has quit [Ping timeout: 480 seconds]
mrpops2ko has quit [Ping timeout: 480 seconds]
sentry has joined #freedesktop
mrpops2ko has joined #freedesktop
snetry has quit [Ping timeout: 480 seconds]
mrpops2ko has quit []
mrpops2ko has joined #freedesktop
vyivel has quit [Remote host closed the connection]
vyivel has joined #freedesktop
croissant_ has joined #freedesktop
croissant__ has joined #freedesktop
croissant has quit [Ping timeout: 480 seconds]
croissant_ has quit [Ping timeout: 480 seconds]
croissant_ has joined #freedesktop
croissant_ has quit [Remote host closed the connection]
croissant has joined #freedesktop
eluks has quit [Remote host closed the connection]
eluks has joined #freedesktop
croissant__ has quit [Ping timeout: 480 seconds]
mrpops2ko has quit [Ping timeout: 480 seconds]
alarumbe has quit []
swatish2 has joined #freedesktop
ximion has quit [Remote host closed the connection]
swatish21 has joined #freedesktop
swatish2 has quit [Ping timeout: 480 seconds]
mrpops2ko has joined #freedesktop
mrpops2ko has quit []
mrpops2ko has joined #freedesktop
sima has joined #freedesktop
georgc has quit [Quit: Leaving]
gchini has joined #freedesktop
AbleBacon has quit [Read error: Connection reset by peer]
jsa1 has joined #freedesktop
<kode54>
wth
<kode54>
cgit is working again??? wow
haaninjo has joined #freedesktop
mebious_ is now known as mebious
swatish21 has quit [Ping timeout: 480 seconds]
swatish2 has joined #freedesktop
guludo has joined #freedesktop
swatish21 has joined #freedesktop
swatish2 has quit [Ping timeout: 480 seconds]
swatish2 has joined #freedesktop
swatish21 has quit [Ping timeout: 480 seconds]
swatish2 has quit [Ping timeout: 480 seconds]
mrpops2ko has quit []
swatish2 has joined #freedesktop
swatish21 has joined #freedesktop
swatish2 has quit [Ping timeout: 480 seconds]
mrpops2ko has joined #freedesktop
mebious has quit [Remote host closed the connection]
mebious has joined #freedesktop
mebious has quit [Remote host closed the connection]
mebious has joined #freedesktop
ximion has joined #freedesktop
swatish21 has quit [Ping timeout: 480 seconds]
AbleBacon has joined #freedesktop
jsa1 has quit [Ping timeout: 480 seconds]
mripard_ has joined #freedesktop
mripard has quit [Ping timeout: 480 seconds]
mrpops2ko has quit []
mrpops2ko has joined #freedesktop
airlied_ has joined #freedesktop
airlied has quit [Ping timeout: 480 seconds]
jsa1 has joined #freedesktop
swatish2 has joined #freedesktop
olivial has quit [Read error: Connection reset by peer]
olivial has joined #freedesktop
jsa1 has quit [Ping timeout: 480 seconds]
___nick___ has joined #freedesktop
<jrayhawk>
I got annoyed and made some performance improvements to it, but it can still get overwhelmed by overaggressive neural net training scrapers.
<karolherbst>
I suspect the blame page causes most of the load?
<karolherbst>
can we put anything that actually needs a lot of processing power behind anubis?
<jrayhawk>
Blame page?
<karolherbst>
I thoguth cgit does have a git blame frontend?
<karolherbst>
mhhh maybe it doesn't...
<jrayhawk>
I am guessing we disabled that a long time ago
<jrayhawk>
ah, '"enable-blame", which is disabled by default'
<elibrokeit>
tbh putting it behind Anubis isn't wrong either way :)
<elibrokeit>
the whole cgit
<daniels>
yeah, blame is disabled, but really, at this point anyone who wants to use cgit should just git pull to local storage and run their own instance
<elibrokeit>
that would defeat the entire purpose of it though
<elibrokeit>
which is precisely to be able to view stuff from, say, a smartphone
<pinchartl>
elibrokeit: without denying the usefulness, I think the world would be a better place if we didn't view git trees from a phone
<elibrokeit>
generally all I actually need is viewing links to specific commits I'm already interested in :P
<Consolatis>
I have another use-case: being able to browse some code from the browser because there is only 148 MiB space left on the daily driver root partition
<Consolatis>
and cgit is *so* much faster for that compared to gitlab (or github for that matter)
<elibrokeit>
cgit is not faster than gitlab, as that would imply that gitlab would eventually load given enough time, and that's not really guaranteed.
<elibrokeit>
and GitHub is a barren wasteland since freedesktop has deleted all GitHub mirrors :)
<karolherbst>
yeah.. with anubis now, the time it takes to browse to a file isn't different between gitlab and cgit
<jrayhawk>
also being supported by hundreds of browser engines instead of 2.5 engines is helpful, especially when debugging on systems with broken EXAs
<elibrokeit>
Anubis is much lighter on my phone than gitlab is. and it also only runs once, not every page load...
<jrayhawk>
Anubis moves back into 2.5 engine territory, which eliminates my interest in cgit
<karolherbst>
gitlab even has a handy "t" shortcut you can search file with very quickly :)
<Consolatis>
gitlab is way slower for me (slow hardware) due to it loading everything and the kitchen sink via XHR
<elibrokeit>
that's nice, how do I use that shortcut on my phone :)
<karolherbst>
we already established you shouldn't browse git repos on your phone :P
<elibrokeit>
even on a laptop the t shortcut is even more of a memory hog than just regularly using gitlab
<elibrokeit>
jrayhawk: is that just because the JavaScript requirement?
<elibrokeit>
Anubis is working on a non-javascript approach
<karolherbst>
newer versions do
<karolherbst>
not sure if the fdo one is new enough
<jrayhawk>
Oh? That sounds great, but I can't even imagine what the design would be.
<elibrokeit>
there is also "go-away" which purports to be Anubis except with non-javascript challenges support
<karolherbst>
also why would the t shortcut be slow on laptops?
<karolherbst>
though maybe my hardware is just overspeced, heh
<elibrokeit>
one simple obvious form of check is to pass a cookie to a login-protected page and require that your browser can successfully reach that page with a 200 return code
<immibis>
jrayhawk: AFAIK there's still no real evidence those have anything to do with neural nets. Most of the evidence offered is that people started getting openly scraped by OpenAI etc. around the same time they started getting DDoSed by unknown actors - therefore those unknown actors must be OpenAI, right?
alanc has quit [Remote host closed the connection]
<Consolatis>
honestly, does it really matter what the traffic is caused by? lots of requests to software which apparently was not designed from the start to be public on the internet (or intended to be deployed at a massive scale and with caching layers in front) will degrade that software's performance
alanc has joined #freedesktop
<jrayhawk>
Yes, it matters. The neural net scraping communities are smart people who are not deterred by cookie requirements.
<jrayhawk>
And yes, they routinely functionally DDoS websites.
<Consolatis>
you can pick a random group of people likely behind lots of requests but that is not the core issue at play here
<daniels>
as someone who has to keep a service running for ~50k people who all have browsers who can complete JS challenges in milliseconds, it genuinely doesn't matter whether it's AI scrapers or DDoS
<daniels>
if placing that JS challenge makes things work for everyone, then great, that's what we'll do - same reason we implemented reCAPTCHA when our outbound mail delivery rate started diving below 25%
<daniels>
I'm sorry that it takes 22 seconds for you, but that is a deeply unusual situation and I cannot imagine which of your technology choices have caused it, but it's presumably within your control to do otherwise
<Consolatis>
if you refer to the 24 seconds above then thats just gitlab for me, anubis is actually fine now, takes at most 3 seconds or something
<karolherbst>
is that on an old smartphone or something?
<karolherbst>
just curious
<elibrokeit>
jrayhawk: (for clarity, the cookie thing is a login cookie, the idea being that you can bypass JavaScript checks if you can successfully login with a username+password)
<Consolatis>
karolherbst: its on a aarch64 single board computer, similar to a raspberry
JanC is now known as Guest21174
JanC has joined #freedesktop
<karolherbst>
ohh.. actually
<karolherbst>
file search functions just as well on the smartphone
<karolherbst>
open the left drawer menu and search for `~$file_name`, the ~ is important
<karolherbst>
ahh..
<karolherbst>
can also click the filter button and select files
<daniels>
Consolatis: I regularly work on AArch64 SBCs and that is not even close to expected behaviour
<daniels>
so I would assume you are lacking JIT or something
<Consolatis>
thanks, that could be a potential bottleneck indeed
<Consolatis>
things are indeed lots of faster on the raspberry 5 for me compared to my daily driver Odroid C4
<karolherbst>
on my oneplus 7 it's also blazingly fast, but that one has an actual competent SoC on it
<karolherbst>
while yes, I do think that it's nice to be able to browse it on low end hardware, there is just so much we can do. But anyway.. it sounds like putting cgit behind anubis could be fin in general..
<daniels>
Consolatis: ah yeah, A55 is not super fun, but indeed I would guess at lacking JIT
<karolherbst>
worst case the challenge difficulty could be lower for cgit
<daniels>
karolherbst: my opinion on cgit is that all its users are on IRC, and setting up cgit is no more difficult than setting up an IRC bouncer
<jrayhawk>
I think I am the only person dogfooding it, and I will stop doing so if it requires javascript.
Guest21174 has quit [Ping timeout: 480 seconds]
<Consolatis>
accessibility with things like elinks is indeed another good point