![]() ![]() |
I think we are through the worst.. there are some scripts who set their updaters to check every 10 ms for updates (probably thinking setTimeout/setInterval took seconds instead of milliseconds). I've added a layer of defense that adds rate limiting and should hopefully protect us. I must sleep. I have a presentation tomorrow morning at work at 8am :( |
![]() ![]() |
Thanks for the great work! |
![]() ![]() |
I'm used to handling 100-200 dynamic req/s -- but 100x as much is much harder to do on a budget. |
![]() ![]() |
Wow... how stupid people can be. Thanks Jesse! |
![]() ![]() |
Thanks Jesse... you beat your estimate. :) Rest well. |
![]() ![]() |
Are you blocking or removing ill-behaving scripts like that? Is there a way we can educate them better? Like, for instance, ask/force them to use the auto-updater that was produced earlier? |
![]() ![]() |
Thanks very much for the time and effort you put into this site. The downtime just highlighted how very useful it is :) |
![]() ![]() |
Wow. Even checking for updates every 10 seconds is extreme. 10 minutes even seems unnecessary. My scripts check just once a day. |
![]() ![]() |
Thanks Jesse... I was about to go crazy. |
![]() ![]() |
edit: adapted the userscripts updater |
![]() ![]() |
i think you need to implement something which checks to see if a new script submission has a reference to userscripts.org/scripts/source/*.user.js and either reject it or flag it for inspection before making it available. |
![]() ![]() |
@Jesse Andrews, I just had to click the install button 3 times in DOMNotify just to get one of my scripts updated and installed... 503 Service Temporarily Unavailable... Between the update notice and the actual click was about 15-20 seconds or so... I am running a few extra scripts that do meta.js calls and one that does a HEAD request to the user.js which might be the catalyst for USO triggering the failure. Anyhow... more incentive (up to several now ;) for adding Just a side note too... it would seem like you are still being attacked from observation on this end as things are quite sluggish here. |
![]() ![]() |
That would give lots of false positives I'm afraid. Just checked one of mys scripts, I have two lines like this:
var UPDATE_CHECK_URL = 'http://userscripts.org/scripts/source/27106.meta.js'; var INSTALL_URL = 'http://userscripts.org/scripts/source/27106.user.js'; This would trigger the code for moderation and would require someone to check that it behaves. But on the other hand, if that someone would be the scriptwriter himself.
So maybe some sort of self-check could be in place? |
![]() ![]() |
Photodeus wrote:Is this an embedded updater or required updater that uses this btw? (send in pm if you want... Thanks) Photodeus wrote:+1... perhaps with a captcha and maybe an exception list for approved externals. |
![]() ![]() |
Photodeus wrote: But that would still leave the door open for miscreants/careless/lazy ppl. |
![]() ![]() |
1nfected wrote:True but they'd be immediately unlisted because they double accepted a TOS if they slip... This issue is affecting even the major updaters now since the caching isn't quite tweaked. Sir Jesse could draw his blade and cast off the demons... sorry... imagination running away today ;) |
![]() ![]() |
Marti wrote:My only concern with self moderation is that an ill-behaving script could still make it to the front page, available for download.1nfected wrote: But that would still leave the door open for miscreants/careless/lazy ppl.True but they'd be immediately unlisted because they double accepted a TOS if they slip... This issue is affecting even the major updaters now since the caching isn't quite tweaked. Sir Jesse could draw his blade and cast off the demons... sorry... imagination running away today ;) Sure it can be unlisted, but what about the ppl who'd have already installed it before it got unlisted? |
![]() ![]() |
1nfected wrote:Well that's what black/white lists exists for... even if USO doesn't want to do this, it's been the way of the internet for several decades now... USO always has the prerogative of actually deleting a script too... and when it tries to update it's going to find nothing. I'm all up for finding a common ground solution if a better one is presented. |
![]() ![]() |
@Jesse Andrews,
|




