@Makyen You’re right, of course, but this is on my list of stuff I don’t want to mess with at all. I only care because my domain is tied to all this stuff, and I’m not interested in dealing with some overzealous law enforcement outfit doing something stupid while they’re following the threads after nuking real CSAM

[ SmokeDetector | MS ] Link at end of answer, potentially bad asn for hostname in answer, potentially bad ns for domain in answer, potentially bad keyword in answer (66): How to trigger the "Refresh" printer entry in Windows Explorer via CLI‭ by Sam‭ on superuser.com

tpu- feedback received on [MS] How to trigger the "Refresh" printer entry in Windows Explorer via CLI‭

@SmokeDetector AIGC w/ spam link

Which should be added to the watchlist. Looks like another keyword to add to the "watch when GH is back up" list

!!/watch- anchorcomputersoftware\.com

watch- anchorcomputersoftware\.com

@Fastnlight Can't do that right now...

I know, thats why I didn't put the !!/ in front of it

Ah, clever. Go back and find them later

Earlier, we were putting watch commands as plain text, so I was following the pattern

!!/isblu history.stackexchange.com/users/69603

@GeneralGrievance User is blacklisted (69603 on history.stackexchange.com).

[ SmokeDetector | MS ] Potentially bad asn for hostname in body (1): server gets a new ip-address every time I reboot on Ubuntue 20.04‭ by user2037551‭ on unix.SE

fp- feedback received on [MS] server gets a new ip-address every time I reboot on Ubuntue 20.04‭

[ SmokeDetector | MS ] Potentially bad ip for hostname in body (1): Embed a dinamic customization into the Storefront using Toolkit‭ by Cesar Mendez‭ on stackoverflow.com

[ SmokeDetector | MS ] Potentially bad ns for domain in body (1): Outline shader/material in unity shader graph ✏️‭ by SeedM‭ on gamedev.SE

[ SmokeDetector | MS ] Potentially bad keyword in answer (1): Why is my whatsapp message not being sent?‭ by Pedro de Castro Tedesco‭ on stackoverflow.com

metasmoke.erwaysoftware.com/review/untagged-domains/381272 what?

is there a way to remove this "app" domain (that's literally the domain, not .app)

@Undo Unfortunately I can completely understand that sentiment.

or if we want to keep app as a domain, give it a new "invalid" tag?

@micsthepick You'll find that's not the only thing picked up erroneously.

do we just skip them, and leave untagged?

There's a few Python constructs that are continuously picked up as false positives as well.

Just to name an example.

@micsthepick They are not a priority.

tp feedback received on [MS] Можно ли создать приложение для Ios и android из сайта‭

naa feedback received on [MS] Можно ли создать приложение для Ios и android из сайта‭

fp feedback received on [MS] Можно ли создать приложение для Ios и android из сайта‭

I don't know why my site specific filter was gone, but I should have at least seen the icon depicting I was looking at a different stack...

ArtOfCode/TRINITY: Executing automatic scheduled reboot.

[ SmokeDetector ] SmokeDetector started in standby mode at rev 839550c (SmokeDetector: Auto watch of ssoidraj\.com by Spevacus) (running on ArtOfCode/TRINITY)

naa feedback received on [MS] How secure is my information on a bootable USB drive?‭

ok, hold onnnn, I SAW my filter there, and after giving that feedback NOW it gives me the relevant site

[ SmokeDetector | MS ] Manually reported answer (93): How to set the picturebox style using SendMessage‭ by Tyson‭ on stackoverflow.com

tp feedback received on [MS] How to set the picturebox style using SendMessage‭

when I go back to the review tab from somewhere else in MS, the first thing it loads is not from puzling, even if it says it is. >:(

tp feedback received on [MS] A Deceitful Rebus‭

[ SmokeDetector | MS ] No whitespace in answer (97): Хеши целых чисел в Python‭ by Эльвина Насардинова‭ on ru.stackoverflow.com

!!/docs

@micsthepick No such command 'docs'.

Yeah, you'll find no luck there at the moment I'm afraid.

The rest of the documentation would be available at charcoal-se.org/about under the Wiki tab.

anyways, I'd like to watch the domain shopee.vn

We can't add watches at the moment, but we can note that for later.

!!/watch shopee\.vn

ah ok, didn't quite remember if there was another command for a domain

can I find a listing for a word?

scrabble2019.txt

tp feedback received on [MS] What's this component found in a Wecon HMI? Need replace‭

@metasmoke tpu-

sdc addblu- puzzling.stackexchange.com/users/80517

tp feedback received on [MS] zeta:know $PROOF Puzzle‭

[ SmokeDetector | MS ] Offensive answer detected (78): Failing To Setup Dynamic Enum Popout Menu‭ by Vona‭ on blender.SE

fp feedback received on [MS] Failing To Setup Dynamic Enum Popout Menu‭

naa feedback received on [MS] I am a non zero number‭

tp feedback received on [MS] ARG: youtube.com/@donde_esta_andrea‭

tp feedback received on [MS] How can one solve a Rubik's cube without relying on guides/algorithms?‭

[ SmokeDetector | MS ] Potentially bad keyword in username (1): how to run android on qemu‭ by jhhhhhhhhhh‭ on stackoverflow.com

can someone explain blog.com#deleted? is it all as it seems?

(the tag)

@SmokeDetector uptime

!!/blame micsthepick

@micsthepick It's Johannes Kuhn's fault.

@micsthepick I think it might be quite literal: Blog links that no longer resolve. I do not think that tag is all that great or helpful, though...

!!/ms-status

@micsthepick metasmoke is up. Current failure count: 0 (teward/Cerberus)

watch djobzy.com

!!/test site=puzzling djobzy

!!/test smartkeda

> Would not be caught as a post, title or username.

!!/test smarkeda.com

> Would not be caught as a post, title or username.

watch smartkeda

tp feedback received on [MS] Data Sufficiency Quiz : Is the average age of the students of a school less than 17 years?‭

!!/test -u propertybro

> Would not be caught as a post, title or username.

meh

tp feedback received on [MS] A website I made a year or two ago with a series of puzzles‭

tp feedback received on [MS] A website I made a year or two ago with a series of puzzles‭

@micsthepick I'm not sure what you're asking for, but I think you're wanting what the entry would be for "scrabble2019.txt". If so, it would be scrabble2019\.txt.

!!/test scrabble2019.txt

this should not be so

x.com/adguard/status/1831040130641252839

another repo containing big lists of spammy domains that GH unexpectedly nuked today

okay, something's happening

!!/bisect scrabble2019.txt

@CodyGray Matched by (?:[a-z0-9-]{1,40}?scrabble[a-z0-9-]{0,40}?|[a-z0-9-]{0,40}?scrabble[a-z0-9-]{2,40}?)\b(?!\.vip(?<=scrabblewordfinder\.vip))(?!\.weebly\.com(?<=officialscrabbledictionary\.weebly\.com))(?!\.blogspot(?<=scrabblewordfinders\.blogspot))(?!\.yolasite\.com(?<=officialscrabblewordfinder\.yolasite\.com))(?:\.[a-z0-9-]{1,40}?)+ on line 13038 of watched_keywords.txt

teward/Cerberus: Executing automatic scheduled reboot.

[ SmokeDetector ] SmokeDetector started at rev 21de582757 (SmokeDetector: Merge pull request #13070 from Charcoal-SE/auto-blacklist-1725364721) (running on teward/Cerberus, Python 3.10.12)

Restart: API quota is 89457.

@SmokeDetector That's the expression that it is matching, @micsthepick. If you're seeing a significant number of false positives for "scrabble2019.txt", then that can be added as an explicit exception, like several others have been. (But it can't be done right now, because the repo is offline.)

I don't know the exact flavor. I believe it's a specific Python package that it uses. If you're just interested in testing an expression (e.g., on regex101), the PCRE or PCRE2 flavors are good choices.

@NobodyNada That situation appears to have already been resolved, which is good. However, frustratingly, "We received the official response explaining that this was a mistake."

There is nothing that I could find in the ToS that would suggest there is remotely any problem with storing this type of information on GitHub.

i would not be surprised if GitHub does their own SmokeDetector-like "nuke repos with this spammy domain" and managed to mess it up in this case

But... doing that based on textual matches is very, very stupid.

There are lots of legitimate reasons why spammy domains might exist in a GH repo. There are even legitimate reasons why actual malware might be stored in a GH repo (namely, research).

So, frankly, I find this kind of difficult to believe, but maybe.

yeah it's really stupid, and unambiguously a mistake (maybe they meant to filter it to brand new repos from brand new users but messed up their filters)

but hey, at least it's not as bad as that one time Twitter banned everyone who tweeted the word "Memphis"

I mean, I can't see how a human could have disabled adguard of all things, especially without leaving any information...automation seems marginally more likely than the alternatives.

Does there actually exist a situation where one cannot say, "hey, at least it's not as bad as that one time Twitter/X ..."?

Not a very high bar.

...good point.

I have no idea about that particular incident. I'm saved a lot of drama by not actually having or using Twitter.

Memphis is, however, a fantastic song by Kitten.

@CodyGray except this one doesn't match...?

oh, silly me, forgot that fake domain

@micsthepick What do you mean? Yes it does. Test on regex101.

Ah, the ".txt". Are there other FP matches with that suffix?

literally .t is the matched part

also, does this part of the regex make sense? (?!\.yolasite\.com(?<=officialscrabblewordfinder\.yolasite\.com))(?:\.[a-z0-9-]{1,40}?)+

What, specifically, are you asking about? What is your concern with that portion?

I guess it does but personally I would just match officialscrabblefilnder in an optional group unless performance reasons

it just seems nonstandard and unwieldy, like it's kind of (not) DRY

Isn't that excluding "officialscrabblewordfinder.yolasite.com"?

(?!(?:officialscrabblewordfinder)?\.yolasite\.com))

Regexes aren't really DRY...

that should be euqivalent

unless it's meant to be a NEGATIVE lookbehind?

I'm actually not sure the details. As far as I know, this is the standard pattern that is used when we want to exclude something from matching, where a simple negative lookahead is not sufficient because the "anchor" portion is variable.

[ SmokeDetector | MS ] Potentially bad asn for hostname in body, potentially bad keyword in body (2): MacOS shell script (video contact sheet) creates files with no permissions on SMB server‭ by J Wolf‭ on superuser.com

actually, that's correct as is, but we could move the negative lookbehind to the left

It is no different than the other exclusion patterns in that very regex.

and save characters and time spent matching those characters in a limited domain

fp feedback received on [MS] MacOS shell script (video contact sheet) creates files with no permissions on SMB server‭

@CodyGray "a foolish consistency is the hobgoblin of little minds adored"

Sorry, I'm confused. Which thing are you moving? There's no negative lookbehind in that expression that I can see.

@micsthepick I mean, yes, it absolutely is. I'd say that most of us here do not have big enough brains to fully understand regex. I am definitely in that camp. If there are standard patterns that work, I can learn and use those. Outside of that, if it gets too complicated, I struggle.

[ SmokeDetector | MS ] Potentially bad keyword in body (1): Problem with adb root on Samsung Galaxy A04e on Android 12‭ by Garby‭ on stackoverflow.com

I'm ok at regexes, when I have the tools handy or the documentation and some time to wrap my brain around the regex

I did get a bit confused earlier though, that regex (?!(?:officialscrabblewordfinder)?\.yolasite\.com)) is not the same

With sufficient time and documentation, I can usually figure out what an existing expression does, but creating patterns from scratch is still mostly a no-go.

fp feedback received on [MS] Problem with adb root on Samsung Galaxy A04e on Android 12‭

technically (?!(?<=officialscrabblewordfinder)\.yolasite\.com)) is the better way of doing it

(without the ? after the lookbehind)

is that more readable, IMO yes.

I'd wager that it wouldn't ever hurt performance my way

plus an already long regex is shortened, that's gotta be beneficial

hmm

Hmm, that's interesting. So, not sure if you're aware of this, but the most common place that pattern is used is if we want to watch, say, the name of a company, but exclude their domain (because that's matched by another watch). For example, if you have a company named "Foo Bar", with a URL of "foobar.com", we'd typically create two watches: foobar\.com and foo[\W_]*+bar(?!\.com(?<=foobar\.com)).

Your permutation of foo[\W_]*+bar(?!(?<=foobar)\.com) seems to work equally well, and be more optimal in the sense that it requires fewer matches.

I'm not sure what other limitations it might have, but it's an interesting suggestion anyway.

[ SmokeDetector | MS ] Potentially bad keyword in answer (1): How to represent a limit of an infinite sum as an integral‭ by EMar‭ on math.SE

fp feedback received on [MS] How to represent a limit of an infinite sum as an integral‭

Makyen/MS-Relay: Executing automatic scheduled reboot.

[ SmokeDetector ] SmokeDetector started at rev ee0f84fecd (Makyen: Log as debug MS WebSocket messages) (running on Makyen/MS-Relay, Python 3.11.3)

Makyen/EC2-linux: Executing automatic scheduled reboot.

[ SmokeDetector ] SmokeDetector started in standby mode at rev 0bf1657028 (Makyen: Revert "Don't verify metasmoke SSL certificate") (running on Makyen/EC2-linux)