[ SmokeDetector | MS ] Link at end of answer, pattern-matching website in answer, username similar to website in answer: Mirrored numbers by RFplus on puzzling.SE

tpu- by CalvT븃

10 second backoff on request to questions/340261?site=physics

@Andy again ^

@Undo okay, that should be good to deploy, filter errors fixed

TL;DR: never try to work with string-binary-integer conversions and expect to get it right first try

[ metasmoke ] ci/circleci failure on 5857c54: Your tests failed on CircleCI

:3

[ SmokeDetector | MS ] Blacklisted website in answer: Comparison of actual running time of algorithmic trading software by PracticalTrader on stackoverflow.com

tpu- by J F

@SmokeDetector you had to rub it in, didn't you?

@ArtOfCode That's ... a lot

[ metasmoke ] ci/circleci failure on 65f3b46: Your tests failed on CircleCI

just a few

Whats your opinion about the 198M leak @ArtOfCode?

/me is glad I don't live in the US

@Magisch It’s not actually a leak though – the people who reported it seem trustworthy, and there’s no evidence AFAIK that someone else has a copy of the data.

Well, 16000 primes per one possibility is quite a lot

I'm really curious about how I've been profiled. It'd be interesting to see what the GOP thinks of me as a voter.

okay, breach then

@JF Publicly available on a sub-domain for 6 months... It's been downloaded.

@JF It laid unsecured on a cloud server for 6 months

@JanDvorak Huh? How'd you come up with that?

It was bound to have been downloaded a couple times if only by people running port scans

cyber criminals that do that

[ metasmoke ] ci/circleci success on 06b6d33: Your tests passed on CircleCI!

[ metasmoke ] ci/circleci success on 06b6d33: Your tests passed on CircleCI!

@CalvT븃 I think that's the 3rd time since we re-implemented those. @Undo, it doesn't look like backoffs are a big deal. We've gotten three in the past 4-5 days.

2^1014 out of 2^1000 is 2^14 out of 2^0

@Andy They apparently have predictive modeling and factor in things like facebook and reddit accounts and polling data

@Magisch Well... looks at social media accounts I don't have either of those.

Oh, wait, you interpreted my question to mean Kibits, I guess

Basically, if I ever get arrested, the cops are going to look for me on Facebook, not find me and assume I'm guilty of something much worse

@JanDvorak what's 2^1000

@Andy Correct

unless you mean the xor function rather than exponentiation, in which case you're rather low

@ArtOfCode 1,075 x 10 ^ 301

...

[ metasmoke ] ci/circleci success on dde9ab1: Your tests passed on CircleCI!

what do these numbers mean?

what do they represent?

2^1000 is the number of 1000-bit numbers including those with leading zeroes

1,075 with 301 zeroes

@ArtOfCode So one zero seven five and then 298 more zeroes

that's the number

It's the scientific notation

You display hilariously big numbers in this notation to not have to write them out

Yeah, I can manage that much.

I did do an A-level in maths and physics.

I just don't/didn't know why that number is relevant to the question

Ahh

I misunderstood your question then

>_>

[ SmokeDetector | MS ] Bad keyword in body: If I join AME courses, then which Aircraft Maintenance Engineering Colleges in Delhi will be best? by HDI Dipali on aviation.SE

Possible spam seed?

Aaaand here's the spam

!!/report aviation.stackexchange.com/questions/39086/…

Yeah spam

@CalvT븃 Post 1: Already recently reported

tpu- by Kyll

!!/report aviation.stackexchange.com/a/39087/8300

[ SmokeDetector | MS ] Manually reported answer: If I join AME courses, then which Aircraft Maintenance Engineering Colleges in Delhi will be best? by HDI Dipali on aviation.SE

tpu- by CalvT븃

!!/test igesame.com

> Would not be caught for title, body, and username.

!!/watch igesame\.com

@CalvT븃 You don't have code privileges, but I've created a pull request for you.

CI on ce5520e succeeded.

!!/pull

Restart: API quota is 13005.

!!/pull

[ SmokeDetector ] SmokeDetector started at rev ce5520e (A Wegner: Merge pull request #856 from Charcoal-SE/auto-blacklist-1497897554.8933525) (running on Undo's RPi 3)

Restart: API quota is 12996.

!!/test-a igesame\.com

> Would not be caught as an answer.

Huuuh?

!!/test igesame.com

Doesn't like the regex, ok

[ SmokeDetector | MS ] Repeating characters in answer: Is a baptized but not confirmed Catholic really a Catholic? by ddddd on christianity.SE

tpu- by J F

@Andy Yeah, noticed that. Weird.

re-ping @Undo to fix metasmoke when you get home

[ metasmoke ] ci/circleci success on 3d5dee1: Your tests passed on CircleCI!

[ SmokeDetector | MS ] Bad keyword with email in answer: Using data given back with getJSON by Sherry on stackoverflow.com

tpu- by J F

mmm… priestade. Is that a soda or a smoothie?

The wix spam reporter reminds me of this:

s/target's coordinates/the URL you feel is spam/ s/missile launch/spam reporter/

Wix says:

> Thank you. You complaint was submitted.

This makes me less likely to use their product in the future.

Haha :p

[ metasmoke ] ci/circleci success on 8e00abf: Your tests passed on CircleCI!

At least it wasn’t “Thanks in advantage! You complain was submitted.”

[ SmokeDetector | MS ] Mostly non-Latin title: Ubutu update is not working ________________ by Arefin Zamshed on askubuntu.com

fp- by J F

\o

o/

Anything exciting happen while I was gone?

@NobodyNada upguard.com/breaches/the-rnc-files

Yikes

You can say that again.

@JanDvorak We'd just ignore that post

Does that work recursively on answers?

No idea, but we can ignore those too

Ignoring each of them separately doesn't stop them from getting autoflagged

There’s always !!/stopflagging.

Sorry, that was mostly-idle speculation, I don't have any plans at the moment, and would probably give a warning if I did

A full kill isn't a great solution - given the time for which it would have to be active

[ SmokeDetector | MS ] Offensive answer detected: Gasoline is dripping out of a rubber tube outputting on the bottom of my motorcycle - is this normal, or is it a problem? by bradley glass on mechanics.SE

tpu- by NobodyNada

The 'max_rep' in findspam determines how we evaluate the post. Since the vast majority of spam is from low rep users, we have the thresholds set very low.

All this means is that Smokey trusts the more established users. If they are caught spamming, we can blacklist them and they will be dealt with by a local mod.

@Sconibulus We'd appreciate that warning. I'd also be interested in that particular puzzle :)

[ SmokeDetector | MS ] Bad keyword with a link in answer: Disable comments feed, but not the others by Herdin on wordpress.SE

tpu- by J F

[ SmokeDetector | MS ] Link at end of body: Gnome shell launcher icon by Hajthem on askubuntu.com

fp- by NobodyNada

This Safari extension is really awesome

I just went to a bloated webpage and disabled JavaScript just to see what would happen

The download size shrank by several times, the webpage was much faster and more responsive, and I didn't even notice anything wrong with it

So I did a quick Google search and installed that extension so I can whitelist JavaScript only on websites where it's useful

[ SmokeDetector | MS ] Few unique characters in answer, repeating words in answer: How to conference call with FaceTime audio? by Keith on apple.SE

tpu- by J F

I just use Ghostery and AdBlock. They get most of the ads without breaking site functionality.

I already use AdBlock for a while, but I just found out about this

I've configured it so that when I visit a new site, it'll give me a little notification in the corner asking if I'd like to enable JS for that site.

That way, I can see if something is broken, and enable some scripts if so. It'll still block trackers and ads and such

You can also configure it so that it silently blocks trackers without blocking useful scripts

10 second backoff on request to questions/116812;116839?site=diy

Why was that added back?

Huh, looks to me like they just want to see if we're getting hit by backoffs a lot

I wonder if it’s possible to determine which JS minifier is used based on what the shortened variable names are.

[ SmokeDetector | MS ] Pattern-matching website in answer: Installing software using serial numbers available in internet is haram or halal by مصطفی on islam.SE

@Cerbrus That ^ Arabic(?) username seems to break the rep count display. If you put the username in a span with unicode-bidi: isolate in the CSS, the problem seems to be fixed.

Looks like it's because of an RTL character in that name

I wonder why it doesn't break Smokey's report though

RTL usernames break MS @Art/@Undo.

RTL breaks everything

@NobodyNada Arabic characters are naturally RTL.

[ SmokeDetector | MS ] Blacklisted website in answer: Is there an electric field around neutral atoms? by mohsen lutephy on physics.SE

4 tps, 9 fps

@SmokeDetector f

You’d never guess that’s 5 characters: ‭م‭ ص‭ ط‭ ف‭ ی => ‭مصطفی

[ SmokeDetector | MS ] Link at end of answer: PGP command-line program t be called from script or batch file? by Todd Mitchell on superuser.com

fp- by J F

I had a hypothesis that the poor performance we're seeing there was due to Art's API queue size changes causing backoffs, but that didn't end up being it

Still, the queue size changes are suspiciously well-timed.

Huh

What else happened early April?

MS or Smokey

CI on b7716b0 succeeded. Message contains 'autopull', pulling...

[ SmokeDetector ] SmokeDetector started at rev b7716b0 (Undo1: Re-remove backoff messages --autopull) (running on Undo's RPi 3)

@ArtOfCode I looked a little bit, but not as thoroughly as I probably should have.

Restart: API quota is 10736.

one theory: I started my Real job on Apr. 10th :P

Could it be some psychological effect related to the timings, like with the review icon on SO?

@NobodyNada That's a pretty marked shift, though

@ArtOfCode MS ready to go?

aye

My theory is that it was caused by this

actually... we can test that, by un-trusting the key for a week or so

Done

@ArtOfCode Apr 5 is too early, though. Unless it took me a long time to deploy it

@Undo coulda done. I don't suppose you get deploy logs?

Could dig them up

Because I think how FIRE works is it only checks for deletion state when a user clicks the FIRE button, which could be multiple tens of minutes after the real deletion. If it sends a deletion log then, that skews stats heavily towards too long.

> Branch master (at b80bd0e2d0b5973f4fae66fa744fb6479edb06a7) deployed as release 20170405220936 by me

Deployed Apr 5

waffles

But yeah, that'd make a lot of sense

let's test this. I'm gonna un-trust the FIRE key for a while and hope that doesn't break it too badly.

If it does... well, it's only a checkbox.

Might be easier to just have @Andy try regenerating without deletion logs from that key?

We have that info

we can try both. Is deletion log API key ID in the API?

yep

uhhhh maybe

It's in the data

FIRE is key 15

It's not available in filters, so unless we jbuilder'd it in, it's not in the API

Maybe it's not there, then

doesn't seem to be

more to the point, I don't actually know how Andy generates those graphs

I was just thinking that :P

Hang on, I have a relatively recent dump. I could try manually generating a less good version of that graph, without key 15.

Wait, isn't that graph the reason the ApiController#detailed_ttd method exists?

Maybe?

No, those are hour-of-day

shrug I can generate one

^ with key 15

Huh, creeping up even before Apr 9?

Looks like it, then relatively stable after ~2017-04-14

What if you filter out all where api_key_id is set?

@ArtOfCode Still need to do this?

@Undo your config/config.yml needs to have sensitive_fields in it.

kk

Okay, with key 15 excluded, this is interesting:

lemme put both on one graph

o.O

[ SmokeDetector | MS ] Offensive answer detected: Why is there an iPhone in "All Eyez On Me"? by Pilgrim on movies.SE

Identical up to the 25th

So that's not the issue

Nor is the queue thing

@ArtOfCode How about a plot of what percentage of tp posts have deletion logs?

Not the original issue, but FIRE is causing deletion times to look longer than they actually are

@Undo over time?

aye

Would be interesting to know if Smokey's websocket deletion monitor thing deteriorated.

Um

Not sure my SQL skills stand up to that

oh

... I'm not sure how to do that either

Might be possible in combined Ruby/SQL, hang on

Raw SQL query for that involves subqueries, which don't play nice with groupdate