The issue dugan was running into should be resolved (and nothing related to this would ever result in a perma-ban).

--jeremy

I am seeing the attached whenever I try to edit a blog entry. As I am logged in, I'm not sure why I'm seeing this captcha? (disregard the error, that's not the issue)

Attached Thumbnails

 

Based on a variety of factors this will happen for a very small number of requests (0.00002% over the last 24 hours).

--jeremy

Ok, thanks... let me check it over for any "keywords"...

Managed to edit without triggering the captcha - had to copy and paste and reconstruct in bits. I had to delete the word "like" from three places in double quotes and the word "having" without the double quotes from one particular sentence. If I added back "having" on a line on it's own in the same entry, the captcha is triggered again.

Is it not possible to turn that kind of invasive and annoying crap off for members who have been registered for a few years / made a certain number of posts?

We skip the vast majority of checks for long time members, but a few happen before we have enough information to make that determination.

--jeremy

I was blocked today too (an hour before). Does (can) it depend on the subforum? Or can it depend on the proxy server I have to use?
And I still blocked, but looks like I can post here.

https://www.linuxquestions.org/quest...0&d=1642669384
^ I have seen this today when replying to a post.
Strangely, it happened only on that 1 tab and only with that 1 reply. I could close & reopen the tab all I wanted, whenever I posted the reply I got that again.
While I had ~10 more LQ tabs open where that didn't happen, I could continue posting there.
I had to disable uBlock completely to even get to the captcha - apparently google-analytics needed its pound of flesh, even though the captcha itself isn't google's.
It makes me sad that this is considered normal nowadays - if I were using a service that does that sort of stuff, my trust in it would be lost.


This is what I posted:
Is there something that triggered the captcha in there? Looks normal to me.

This is ridiculous - the "security check" screen can be triggered with nothing more than "ls -l", when placed at the start of a line!

ls -x
ll -l
#ls -l

^ the first two are fine, but remove the hash from the third line and it requires a captcha.

2 members found this post helpful.

It's a cloudflare issue. I don't think their switching to a new captcha provider is related.

It should not be triggering on words or commands like those regardless.

i have to solve CAPTCHA's for 2 times per 10 site reload's.

i posted one site link at Security forum, after that came the CAPTCHA's.

I don't know how can it be a cloudflare issue:
I could post here, but in the same time another subforum was read only, I could not post anything, because I (my IP) was blocked.

Well the "one more step" captcha page in the screenshot I attached is a cloudflare thing. And the captcha is the one cloudflare uses (hcaptcha).

Yet another case:
https://www.linuxquestions.org/quest...1/#post6332790

The "security check" can be bypassed by using the full path, escaping/quoting, or ensuring there's a non-L option first - all of which will maintain the (intended) behaviour:
A few more quick tests show that the trigger is likely the regex pattern "\nls\s+-l" - i.e. a newline, then "ls" then any number of spaces/tabs/newlines (at least one) then "-l".