@ArtOfCode ^

@ArtOfCode The filter string is MLLKIHJMHIHKKFMJLLHGMKIMMGOKFFN

@ArtOfCode Looks like there is an MS API breakage for whatever reason:

@ArtOfCode Well, can we just pipe blazer result to find_all ?

@ArtOfCode I don't think it is for blazer (so if sql search is restricted to core then it is no problem)

@ArtOfCode SELECT id FROM p_posts WHERE body LIKE '%string1%' AND body LIKE '%string2%' AND feedback_count = 1 kind of things

@ArtOfCode Generally, have a search page with a 'sql' input box, and let user select post ids; after that, present those posts the same way as search results (which have fancy UI like spammer name, post creation date and existing feedbacks)

@ArtOfCode Technically, but it requires effort

@ArtOfCode Not with the fancy UI

@ArtOfCode Is it possible for it to get disabled at a server level or kind of things? Because it is very strange for it to be in the scheduler and the job itself not commented out.

@ArtOfCode I think we can fire up a keybase or something because we might want to discuss the previous one as well (which I think we have not yet fully solved)

@ArtOfCode @ThomasWard ^

@ArtOfCode If you have time... Just want to ask if it is okay to render "some #{user_inputted} string", status 200 without sanitizing

@ArtOfCode Ooooops

@ArtOfCode Strange enough, when I later tried to access it, it finished in about 0.3 s and now it's query-times rank dropped to nearly the bottom, so I assume MS is again causing random bug for me due to being annoyed by my tests.

@ArtOfCode Low priority things; you may be interested in this since you are the one implemented it ^

@ArtOfCode If you can still ssh to the server, MS is now down and Osiris is broken

@ArtOfCode Yeah, but I don't think there are anything else which can cause such terrible result and "enumerate ****"

@ArtOfCode But it will tell me no core user did bad things

@ArtOfCode But there is no recent code changes remotely likely to cause the issue...

@ArtOfCode what does htop say?

@ArtOfCode I don't think so

@ArtOfCode I would assume it is a block on the router

@ArtOfCode Mine?

@ArtOfCode Maybe look through p_blazer_audit?

@SmokeDetector @ArtOfCode ...

@ArtOfCode I guess some kind of policy DROP because my connection just times out

@ArtOfCode Including iptables?

@ArtOfCode But it kinda stops FIRE from working

@ArtOfCode I got blocked!!!!!

@ArtOfCode Oh please do fix it and check nginx log

@Daniil Or maybe @ArtOfCode

@ArtOfCode really?

Looks like my feedback won't persist on metasmoke.erwaysoftware.com/post/259561 . (@ArtOfCode @thesecretmaster)

@ArtOfCode Basically, I am thinking about just have the empty tree created as migration, and after MS is online, let someone write a script that sends insert range request to dev controller

@ArtOfCode I have spinning disks, which severely degraded the performance

@ArtOfCode Likely; my current implementation is to just have MS online and migrate via the insert range functionality in dev controller

@ArtOfCode which operation?

@ArtOfCode Yes.

@ArtOfCode How can I tell Rails to "execute something for once when this commit get deployed/tested" ?

@ArtOfCode Which file in MS repo contains AppConfig data?

@ArtOfCode Nope, "potentially bad"

@ArtOfCode I believe utf-8 is variable width, so one need to iterate through the string to find an index (this is also where the awkward u32str_seek() comes from - it was u8str_seek() which actually increased performance significantly)

@ArtOfCode Because utf-8 is not O(1) indexable

@ArtOfCode Well, if you are interested in checking for backdoors random issues, you may want to look at the prototype here while I try to understand the terrible Ruby ffi

@ArtOfCode This is when the ugly GIL stuff kicks in :-/

Maybe we should tell @ArtOfCode that the index is not helpful and causes people to use regex, which has more overhead, not less :-/

@ArtOfCode The idea is that you cannot really trust me and my code even if I say I am doing a not-so-formal-at-all verification for now, and that is a security risk to run such code without actually reading it very carefully (pointer magic warning)

@ArtOfCode What if I hide some buffer overflows in the code that is just as innocent as programming bugs? :-P

@ArtOfCode I bet you won't actually read thousands lines of C with a bunch of pointer ops :-P