Hi, Xe here. Admins can confirm via my email address matching what’s on my blog. I’m going to add ASN based filtering (with an example rule for Opera Mini) in the near future. Things are exploding much faster than I thought, so it may take a hot minute. Sorry! Does Opera Mini work on iOS? I don’t have an android phone.
Yes, Opera Mini works on nearly all phones, Android, iOS and even those low-end Java-based ones are still supported as far as I know.
I am still a bit undecided on this topic. While it’s great that Anubis doesn’t block bots that don’t pretend to be browsers, I am not sure how long you can keep that. My personal experience is that while those Browser-like bots are annoying, they are not harmful enough yet for me to care. On the other hand, I have had very bad experience with other bots (e.g. Scrapy) that seem to think that crawling your web site with 16 concurrent connections at full speed is a good idea.
The primary problem currently are AI scrapers that pretend to be browsers and don’t care about any rules (robots.txt) and request the same page over and over again.
This wastes a lot of resources and can bring some server software like cgit down.
Anubis seems to solve that very effectively,at least for now.
Sure a day will come when the scrapers adapt to the methods and try to circumvent Anubis,maybe by sending another user agent that doesn’t contain Mozilla,but then Anubis will surely also receive an update.
Personally,I’d actually prefer blocking all bots except those of smaller and friendly search engines like Mojeek,Stract and Marginalia,and maybe Bing,which isn’t nice by itself but powers nearly all privacy-respecting search engines.
I honestly don’t see a reason not to block random bots that waste resources and do no good,but as long as they’re not a major resource hog (like the AI scrapers currently are),that’s probably a minor issue.
YMMV. That very much depends on what your services are providing and how they are hitting you. I guess for a small static site you’ll only notice some traffic spikes. For other stuff, you can read some examples of the damage they cause in an article linked from anubis’ docs.
Hi Xelaso, welcome. For the record, I did not verify your email, I’m not sure how one would do that.
I wanted to give a small feedback to your blog, which I’ve checked for uodates on anubis: I can’t tell what the dates are. I gather they are posted in chronological order, but reading this format is really confusing to me. Localizing this based on the language may be too difficult, but maybe the ISO date could be used?
I’ll understand if you don’t want to however.
It’s in the en-US middle-endian format. I can switch it over to YYYY-MM-DD though. You’re the first person to ever comment on that!
Workaround: use “Download Linked File” from the context menu instead of direct click of the link.
ok just tested on latest nightly and everything works flawlessly. If you’re on this forum, there’s no good reason not to use the latest Haiku Dev build imho
Now I’m just curious about how Anubis looks like when you’re not deemed worthy to enter the page it protects;-)
yes please switch to a sane format, double +1 from Austria/Europe:=)
It’ll roll out in a few minutes.
It happened now twice that I went to Haiku’s Trac with Web+ (by clicking a link of a notification email) to add a comment. The webpage loaded alright, I entered my comment, but when I clicked the “Submit” button, that infantile Anubis girl popped up, vetted my humanity (I passed), and lo! My comment was gone…
@kallisti5: Web+ failure or is there something to be tuned with Anubis?
It is probably WebPositive bug that cookies and/or HTTP headers are not properly remembered.
Serving out your curiosity …
(it’s cute also ! |) )
I got after I wanted to get back a cgit page from a link from History in Web+
Finally it had turned out … it is because
cc: @kallisti5
WebPositive, you old cookie monster…
As with the HDS webinterface being prone to crashing in Web+, I’ll just have to train myself to copy my texts frequently to the clipboard, esp. before clicking anything…
Same happens for me in Safari mobile. Quite annoying since anubis should have checked first.
Trac loosing comments is something trac does often when loosing connectivity though. Just makes it that much more slow/annoying to work with… maybe we will get a native bugtracking app eventually.
Can you connect to https://halone.within.lgbt and give me the X-Tls-Fingerprint-Ja4
header value? I’ll add WebPositive to my “known good” list.
Thanks!
UPDATE: I got a VM working and confirmed that the latest patch I’m working on for Anubis will let WebPositive through without having to “known good” its TLS signature:
I wonder if anubis would cut down on the number of AI spammers posting to this forum
So far it is still manageable, but it is sometimes a bit annoying, having to read every new flagged post twice to see if they added some AI spam links…
Throw me an email or a DM on OFTC. I’d love to learn more!
If you want you can join our #haiku channel on OFTC