« first day (2154 days earlier)      last day (3024 days later) » 

04:45
@grochmal so here is the thing, if i am in a shell that does not have a tty but the process for the tty push back has a tty... what should i do in that scenario ? im studying pentesting so im very curious
 
9 hours later…
13:51
HTTPS is not working on META :-(
I can't post on that because of sniffers on the internet
It used to have one problem (it still has), that when I click chat it goes to HTTP URL instead of HTTPS as original URL, since the Cookie is Secure, I have to type "HTTPS" manually on chat or use HTTPS EVERYWHERE
With META, adding HTTPS redirects back to HTTP
 
3 hours later…
17:08
@silverpenguin , well TIOCSTI syscall accepts only a single argument. Therefore the kernel must figure out the TTY from the process calling.
Even if you have a process with the same user using a TTY you cannot do a TIOCSTI because that syscall would need to come from that PID. And you cannot change the PID of a syscall.
in general, people ignored and still ignore TTY pushback because it is a pretty difficult scenario to happen. Someone must make blatant errors in his behaviour (e.g. running an exploitable internet facing daemon from a command line that has been a root shell in the past).
@Aria - Hmm... i thought that *.stackexchange.com in the cert would be enable to allow meat.security.stackexchange.com. But even my FF is complaining, I learned something new!
 
6 hours later…
23:25
@grochmal yeah, they did a nice long blogpost back when, and verified by @ThomasPornin - turns out the spec does not allow for two levels of wildcard domains, or even another subdomain under the wildcard. So .stackexchange.com works, but *..stackexchange.com would not, and neither would meta.*.stackexchange.com. So yeah, metas are screwed (for now).
several meta posts and blogposts discussing this...

« first day (2154 days earlier)      last day (3024 days later) »