« first day (2617 days earlier)      last day (1190 days later) » 

Why is this watch PR showing far more hits ("\>{4,} has been seen in 3145 true positives, 137 false positives, and 8 NAAs.") than the linked MS search returns?
 
@CodyGray Both appear to be wrong.
 
So... the MS search linked in the watch PR is wrong, because it doesn't fill out all 3 fields?
 
I'm not entirely sure what goes wrong here, but it's a bit of an odd regex either way.
I'm not touching that PR.
 
Doesn't seem like that odd of a PR to me.
Just maybe needs to be tweaked.
 
9:14 AM
That's why I'm not throwing it out.
 
If I could figure out a way to exclude matches that are immediately preceded by <code>, I think I'd be happy with it.
 
teward/Osiris: Executing automatic scheduled reboot.
 
A negative lookbehind doesn't consume characters, though, and I don't know what I'm doing when it comes to writing regexes.
 
[ SmokeDetector ] SmokeDetector started at rev 85b936f (SmokeDetector: Auto watch of fashiontiy\.com by Cody Gray) (running on teward/Osiris, Python 3.6.9)
Restart: API quota is 18521.
 
I'm hoping my regex will improve further if I stick around long enough.
But some of the magic I see is way above my head.
As in, I can figure out how it works eventually, but how do you ever come up with it.
Perhaps I'm just too wary of FP.
 
9:17 AM
If I could do what I described, and assuming your proposed search is the correct one, the FPs get down to only 1, which is something I'm happy with having caught anyway, FP or not.
 
The bookmarklet I used:
 
[ SmokeDetector | MS ] Url-only title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +2 more (590): norman-reedus-cbd-oil-97.webself.net/ by peytion larsen on askubuntu.com
 
tpu- feedback received
 
javascript:(function(){
    var text = "";
    const activeEl = document.activeElement;
    const activeElTagName = activeEl ? activeEl.tagName.toLowerCase() : null;
    if ((activeElTagName == "textarea" || activeElTagName == "input") && /^(?:text|textarea|search|password|tel|url)$/i.test(activeEl.type) && (typeof activeEl.selectionStart == "number")) {
        text = activeEl.value.slice(activeEl.selectionStart, activeEl.selectionEnd);
    } else if (window.getSelection) {
        text = window.getSelection().toString();     }
const searchTerm = encodeURIComponent(`\\b(?s:${text})\\b`);
const searchTitle = `&title_is_regex=1&title=${searchTerm}`;
const searchBody = `&body_is_regex=1&body=${searchTerm}`;
const searchUsername = `&username_is_regex=1&username=${searchTerm}`;
const url = `metasmoke.erwaysoftware.com/search?utf8=%E2%9C%93${searchTitle}${searchBody}${searchUsername}&or_search=1`;
window.open(url);
})();
I should attempt to minify that so it fits a single line in chat...
 
I still haven't figured this out.
 
9:25 AM
sdc watch-force norman[\W_]*+reedus[\W_]*+cbd[\W_]*+oil(?:[\W_]*+(?:\d++|[\da-f]{5,}+)s?)*
 
It's making me feel like an idiot.
 
@cigien That pattern looks like it's already caught by Bad keyword in body and Bad keyword in answer; append -force if you really want to do that.
[ SmokeDetector | MS ] Few unique characters in body, no whitespace in body, potentially bad keyword in body, repeating characters in body (261): Расширение sssssss ✏️ by peroni1 on ru.stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Why bullet point font not change in my word press post by Suhail Ahmad on stackoverflow.com
 
fp feedback received
 
@SmokeDetector v
 
9:34 AM
@CodyGray User blacklisted (15017383 on stackoverflow.com).
 
[ SmokeDetector | MS ] Blacklisted website in body, link at beginning of body, potentially bad keyword in body (142): what chuljang kakaotalk (dia49) mean by 이지영 on stackoverflow.com
 
tp feedback received
 
!!/watch havecamerawilltravel\.com
 
[ SmokeDetector | MS ] Link at beginning of answer (34): I want to move my data from postgresql to ms sql? by VEDAT PALA on stackoverflow.com
 
9:44 AM
"Potentially bad keyword in body" means "I found a keyword that is on the watchlist"?
 
naa feedback received
tp feedback received
 
@CodyGray Yeah, that seems to work nicely; (?<!code)\b(?s:\>{4,})\b keeps all the TP, and drops all FP except the one you mentioned.
This just uses a negative lookbehind though, so I'm not sure what you mean by it not consuming characters, or the issue with that.
 
I have been spinning my wheels for the past... since that message trying to make this work.
I am trying (?<!\<code\>)\>{4,} on the text from here using regex101.com The trailing > from <code> is still getting matched, when it shouldn't be, because the negative lookbehind doesn't consume characters.
 
[ SmokeDetector | MS ] Username similar to website in answer (64): LinkedIn API integration OAuth2.0 by Jeevan Gupta on stackoverflow.com
 
fp feedback received
 
9:51 AM
Also, we can't watch that directly, can we? Aren't watches going to be wrapped in \b(?s: ... \b?
Hmm, I guess that doesn't change the results.
 
@CodyGray Hmm, I'm not sure. If word boundaries are added then we can skip the closing one, but the one after the lookbehind should still be there I think. I'll play around a bit with it to see.
 
Gah.
I find this irritating.
There's so much I don't know, both about how this system works and about regexes in general.
 
Which part? Ah, I see :)
 
[ SmokeDetector | MS ] Potentially bad ns for domain in answer, potentially bad keyword in answer (2): Auto heaphone connecting problem by Sudip Kumar on askubuntu.com
 
I enjoy not knowing how things work. Then I can spend a bunch of time trying to figure it out :)
 
9:55 AM
That feels like wasted time.
 
Wasted? What do you think I do with my time? I'd be getting high and watching youtube otherwise. This way I can get high and learn something about regexes ;)
 
Hmm.
I have lots of things to be doing.
 
Well, I wasn't recommending that you spend time on it. I was just mentioning why I enjoy it.
 
Yes, I get that.
The fact that you have nothing better to do almost frustrates me more.
:-\
No offense intended, of course.
It's great for you; just less so for me.
 
Actually, I have other things I'm supposed to be doing as well. Shhh, don't tell my grad committee.
 
9:59 AM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Reading a video on GPU using C++ and CUDA ✏️ by Billy on stackoverflow.com
 
@SmokeDetector ignore-
 
@CodyGray We could avoid matching the > after code probably, but it wouldn't help with this FP at least, it seems to be caught by arabic>>>>I in the first line.
 
Oops.
I linked you the wrong test-case.
Too many tabs again (for those following along at home)
 
So many tabs, so little time. Did I mention that already?
 
I don't know, there are too many chat messages.
 
10:02 AM
[ SmokeDetector | MS ] Username similar to website in answer (64): How do i setup a api linkedin connection using oAuth and php by Jeevan Gupta on stackoverflow.com
 
tp feedback received
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): MyAdvancedToken Automating Selling and Buying function not working by EMA4869 on ethereum.SE
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): MyAdvancedToken Automating Selling and Buying function not working by Tertius on ethereum.SE
[ SmokeDetector | MS ] Potentially bad ns for domain in answer (1): Spring-Boot-Security: Custom Authenticator by khadar shaik on stackoverflow.com
 
naa feedback received
 
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, pattern-matching product name in body, potentially bad keyword in body, +1 more (388): sites.google.com/view/herbal-grown-cbd-oil-price/ by DeshalWher on apple.SE
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): Arduino Nano programmer not responding by Domkrt on stackoverflow.com
 
fp feedback received on [MS] Arduino Nano programmer not responding
naa feedback received
 
10:10 AM
[ SmokeDetector | MS ] Username similar to website in answer (64): Help with retrieving your profile via LinkedIn's API? by Jeevan Gupta on stackoverflow.com
 
tpu- feedback received
 
@SmokeDetector User's 3rd answer at least linking to their website, no disclosure on two of them, and flimsy disclosure on one.
sdc watch jeevangupta\.com
 
10:29 AM
@cigien When possible, please avoid using look-arounds as the first thing in a regex. Using a look-around as the first thing tends to be computationally expensive. In most cases, the regex can be easily adjusted to have the look-around at least a few characters after the start of the regular expression, which can reduce the computational load substantially (up to more than an order of magnitude, depending on the text being matched and the regular expression).
 
[ SmokeDetector | MS ] Linked punctuation in answer, pattern-matching website in answer (140): What's The Difference Between Paper and Long-Lasting Filters by Andreo Steven on coffee.SE
 
tpu- feedback received
 
[ SmokeDetector | MS ] Potentially bad ns for domain in body, potentially bad ns for domain in title (2): Curve.fi deposit - Where is my money? by Richard Nies on ethereum.SE
 
sdc watch bestdehumidifier2021\.com
 
[ SmokeDetector | MS ] Url-only title, bad keyword in body, bad keyword in title, potentially bad keyword in body, potentially bad keyword in title (294): normanreeduscbdoil.wixsite.com/norman-reedus-cbd by EthekHenriksen on askubuntu.com
 
tpu- feedback received
 
@SurajRao Looks like it is relevant to the question
 
the contact info here suggests undisclosed affiliation
 
10:50 AM
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer, potentially bad ns for domain in answer (2): How to toggle sidebar menu with pure CSS? by Neeraj Kumawat on stackoverflow.com
 
tpu- feedback received
 
sdc watch baexamresult\.in
 
sdc bisect norman-reedus-cbd-oil-97.webself.net
 
Matched by the following regexes:
norman[\W_]*+reedus[\W_]*+cbd[\W_]*+oil(?:[\W_]*+(?:\d++|[\da-f]{5,}+)s?)* on line 26393 of watched_keywords.txt
webself\.net on line 2835 of blacklisted_websites.txt
 
10:52 AM
Ah
 
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, pattern-matching product name in body, potentially bad keyword in body, +1 more (388): sites.google.com/view/herbal-grown-cbd-oil-order/ by MerryRosz on apple.SE
 
tpu- feedback received
 
[ SmokeDetector | MS ] Link at beginning of answer (34): Water simulation doesn't work properly by Karol on blender.SE
 
sdc blacklist-keyword-force herbal[\W_]*+(?:grown|native)(?:[\W_]*+(?:cbd|oil|order|price|\d++|[\da-f]{5,}+)s?)*
 
@cigien That pattern looks like it's already caught by Bad keyword in body and Bad keyword in answer; append -force if you really want to do that.
 
10:55 AM
IIRC, a watch that's not identical to a pattern that's blacklisted should be unwatched, correct?
sdc bisect herbal grown
 
Matched by the following regexes:
herbal[\W_]*+(?:grown|native)(?:[\W_]*+(?:cbd|oil|order|price|\d++|[\da-f]{5,}+)s?)* on line 2424 of bad_keywords.txt
herbal[\W_]*+(?:grown|native)(?:[\W_]*+(?:cbd|oil|\d++|[\da-f]{5,}+)s?)* on line 26367 of watched_keywords.txt
 
@cigien If it effectively duplicates the other watch, then yes. Things shouldn't be double detected for the same thing.
 
[ SmokeDetector | MS ] Link at end of answer, potentially problematic ns configuration in answer (62): Docker based Web Hosting by gHosting LLC on stackoverflow.com
 
sdc unwatch herbal[\W_]*+(?:grown|native)(?:[\W_]*+(?:cbd|oil|\d++|[\da-f]{5,}+)s?)*
 
10:58 AM
@Makyen Got it, thanks.
 
tpu- feedback received
 
np
 
@Makyen I'm still reading this, and working through it, but minor point, teh --> the in the second sentence.
 
[ SmokeDetector | MS ] Bad keyword in title, link at beginning of body, potentially bad ip for hostname in body, potentially bad keyword in body (142): Which is the best QuickBooks Institute in Kochi, Kerala? by user15018086 on meta.SE
 
tpu- feedback received
 
11:01 AM
@cigien Thanks.
 
np
 
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer (1): typing a backtick on Mac OS X Snow Leopard by Rik Wouters on apple.SE
 
fp- feedback received
 
I'm still trying to figure out how either one of these is "obvious".
 
fp- feedback received on [MS] Water simulation doesn't work properly
naa feedback received on [MS] Water simulation doesn't work properly
 
11:05 AM
!!/bisect ghosting.ge
 
@CodyGray 'ghosting.ge' is not caught by a blacklist or watchlist item.
 
!!/watch ghosting\.ge
 
sdc bisect logicindia.net
 
@ppwater Matched by logicindia\.net on line 26196 of watched_keywords.txt
[ SmokeDetector | MS ] Potentially bad keyword in body (1): What’s the best security for own server law firm or beauty company by Stany Capler on serverfault.com
 
11:10 AM
fp- feedback received
 
@CodyGray I guess it depends on your point of view. The first one seems more obvious to me that it's looking for a static string of characters which must match the case supplied, then looking backwards to something, which will, of necessity, mean that it's looking backwards for something that is prior to the already found string. Given that it was looking for a static string, it wouldn't make sense for it to be only re-verifying that the static string was what it found.
It does have to include the static string in the look-behind, but it's a given that that portion of the look-behind matches.
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): adding a cuda file to an existing c project in visual studio by DataCrusade1999 on stackoverflow.com
 
@CodyGray Yes. "Bad keyword" means it's on the blacklist.
 
@Mast but why Potentially?
 
Because the watchlist can and will produce false positives.
 
11:20 AM
Wait. You're mixing watchlist and blacklist.
 
That's also why the weight is forced at 1.
@CodyGray No, I'm pointing out 2 situations with bad phrasing.
I'm confirming to you that Potentially means it's on the watchlist.
While adding the same reason without the Potentially indicates it's on the blacklist.
 
Oh, I see.
Why can't the messages be more explicit? Is that an attempt to avoid spammers figuring them out too easily?
 
lol
 
@CodyGray Probably in part due to legacy. Once you know it, it doesn't bother you as much thus none of the regulars mind it much.
 
@cigien Merging two entries like this is going to depend on if you feel that it's the same product by the same people. Effectively merging this is grafting the prior 6 TP onto the new item, which may, technically, allow blacklisting the combined regex prior to the new product getting all that many TP or testing/exposure on the watchlist. It's probably OK in this instance, but it's not really something we should be doing on a regular basis.
It would be OK to try to find a pattern that more generally matched multiple occurrences, but this instance feels like it's trying to merge them in order to blacklist the newer product name sooner.
It also concerns me a bit because the pattern that's being used is not all that restrictive. If the "CDB oil" portion was required instead of being optional, I'd feel more comfortable about it. Basically, "herbal native" doesn't fit into as many "normal" sentence structures as "herbal grown" may, which could mean that "herbal grown" might end up seeing some FP.
I see that I should have gone through the transcript earlier, or at least gotten back to to this ping sooner, as you've already made the adjustment. It's reasonable for you to have assumed I would have replied prior to this (i.e. my fault). We're probably OK with it as it is now, particularly given the additional TP posts with "herbal grown" which have shown up since the message I'm replying to.
 
11:29 AM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): How appear sticky ads adsense? by Yogi Aprilio on stackoverflow.com
 
@Mast Aaaand, we can't change them now, because then that would confuse the regulars...
 
fp- feedback received
 
@CodyGray Unfortunately, yes. It would also require a substantial adjustment to multiple fields in the database on MS, with ~90k posts affected.
 
@Makyen You can't be everywhere at once.
 
@Mast True. As it worked out, the last ~36 hours have been substantially below average wrt. my time participating on SE. :;
 
11:48 AM
@RyanM The deletion tracking isn't all that robust. It's not uncommon for it to miss that a post has been deleted. A very quick check using the test in this question didn't show any problems with the SE WebSocket. However, that test is really intended to detect only a specific small subset of the possible problems on SE's side. All things considered, I wouldn't read much wrt. issues on SE's side into SD/MS missing detections of posts being deleted.
 
fp- feedback received
 
@Makyen I hadn't looked at whether the posts were the same product by the same people. I hadn't even realized that that was particularly relevant, especially for the cbd-oil and keto spam. I'll definitely keep that in mind when deciding whether to merge entries in the future.
Looking at the posts more closely, I don't see any evidence that they are the same spammers. In fact, the native series being grouped tightly a year ago, and the grown series in the last 2 days suggests that they are not the same.
It seems that you're comfortable with how my adjustments ended up, but if it's ok with you I think I'd like to split those entries up. The native series can be blacklisted given the TP on them. Additionally, the "cbd oil" portion can be made mandatory, as it catches all TP, even though you made the original watch with that being optional, and as you point out, "herbal native" is unlikely to generate FP on its own.
For the grown series, I agree that making "cbd oil" mandatory would be much better, so I'd like to do that. Also, it seems it might not technically qualify for blacklisting, but making the adjustment for "cbd oil" to be mandatory makes it quite unlikely to generate FP, so I think I'll blacklist it anyway instead of waiting for additional TP. Is this reasonable? I won't make any of these adjustments till you respond. Take as long as you need, of course.
And please don't apologize for not responding within some time frame; I went ahead with the adjustments because I felt they were reasonable, not because I assumed you'd have responded by then. If I do need you specifically to respond, I'd ping you at at least a second time as a reminder.
 
fp- feedback received
 
11:58 AM
@cigien Sounds fine. Feel free to go for it. :)
 
@Makyen Ok. I'll double check that I got all that right, and then make the adjustments. Thanks.
 
np
 
[ SmokeDetector | MS ] Bad ns for domain in answer, bad keyword in link text in answer, potentially bad keyword in answer, username similar to website in answer (222): Task Management and Reminder Tool by Utter Coupons on softwarerecs.SE
 
tpu- feedback received
 
!!/blacklist-website uttercoupons\.com
 
12:05 PM
 
Am I misunderstanding something here? What's the logic for needing to keep watches separated based on the product/person who is posting?
 
So we know who to notify if anything goes wrong with it.
Or if we merge something.
 
I think it's more what Makyen said here:
44 mins ago, by Makyen
@cigien Merging two entries like this is going to depend on if you feel that it's the same product by the same people. Effectively merging this is grafting the prior 6 TP onto the new item, which may, technically, allow blacklisting the combined regex prior to the new product getting all that many TP or testing/exposure on the watchlist. It's probably OK in this instance, but it's not really something we should be doing on a regular basis.
 
Oh, that.
 
For example, imagine two entirely made-up hypothetical unrelated Java help websites. java-help.com and java-examples.com. We could catch both with a watch on java-(?:help|examples)\.com, but we wouldn't want to do that, because one getting itself added to the blacklist shouldn't result in the other getting added to the blacklist as well.
 
12:14 PM
Makyen/EC2-num02: Executing automatic scheduled reboot.
[ SmokeDetector ] SmokeDetector started at rev 63b113a (Makyen: This is the MS instance, so needs to post feedback) (running on Makyen/EC2-num02, Python 3.6.12)
 
@RyanM Doesn't something being added to the blacklist have to be manual action?
 
Yes, but it's often done at the granularity of an entire watch (and is far, far easier to do if done that way, because the blacklist commands also remove the duplicate entry from the watchlist)
 
Yes, often the entire watch is upgraded at once.
So it's important to only put in a watch what we think is similar to each other.
 
Otherwise you'd have to manually notice that they're unaffiliated (which you might not, since you'd probably assume they were watched together for a reason) and then unwatch the combined one, manually rewrite the regex into two regexes, watch one and blacklist the other. This would be extremely annoying.
 
Otherwise we might as well rewrite all existing watches into one big watch.
I'd recommend doing that automated though...
 
12:18 PM
@Mast Amusingly, this is exactly how SD functions internally
 
So that's how it keeps up the performance.
 
@CodyGray this is one of the things that bookmarklet of mine does, although it takes text, not a regex, as input (unlike the PRs)
it copies the watch command to the clipboard and also searches SE + MS
 
sdc unblacklist herbal[\W_]*+(?:grown|native)(?:[\W_]*+(?:cbd|oil|order|price|\d++|[\da-f]{5,}+)s?)*
 
sdc blacklist-keyword-force herbal[\W_]*+native[\W_]*+cbd[\W_]*+oil(?:[\W_]*+(?:\d++|[\da-f]{5,}+)s?)*
 
12:22 PM
@cigien That pattern looks like it's already caught by Bad keyword in body and Bad keyword in answer; append -force if you really want to do that.
 
We really need to put those bookmarklets on the GH repo.
 
sdc blacklist-keyword-force herbal[\W_]*+grown[\W_]*+cbd[\W_]*+oil(?:[\W_]*+(?:order|price|\d++|[\da-f]{5,}+)s?)*
 
@cigien That pattern looks like it's already caught by Bad keyword in body and Bad keyword in answer; append -force if you really want to do that.
 
@Mast that's probably a good idea...or at least link to 'em somewhere
 
I'd prefer if you just submitted them to SD directly as a !!/search command :-)
That way, when I don't have the bookmarklet handy, I am less frustrated.
 
12:27 PM
I could also do that...
one of these days I'll get set up for SD dev
 
@RyanM We could always put them under Utilities, but I think it deserves its own sub-repo like we have for Userscripts.
 
A userscript would also be cool. I don't know if you already know how to actually program JS.
 
I have that baseline level of "can sorta write JavaScript" that many programmers have, but I've done very little work in JS
 
"can write enough JS to be dangerous"
Yes, understood.
 
But I'd have to make it work on at least chat and MS, given my usage...it's convenient to have in the bookmarks bar as an omnipresent button
rather than modifying the markup of two sites
 
12:32 PM
I don't have a bookmarks bar taking up precious vertical screen real estate
 
I have them in a separate folder under my bookmarks.
Select text, ctrl+b, hit button, done.
 
Also, I really don't like the idea of something clobbering my clipboard..
 
It doesn't enter your clipboard.
 
13 mins ago, by Ryan M
it copies the watch command to the clipboard and also searches SE + MS
 
Ah, that one possibly does.
The one I posted earlier doesn't.
So it's possible to cut that part out and still make it work.
 
12:34 PM
[ SmokeDetector | MS ] Potentially bad keyword in answer, blacklisted user (72): User Profile Synchronization Service Stuck on starting by azharalibuttar on sharepoint.SE
 
For me, I'm often using it from MS pages, so the clipboard is an extremely convenient way to get it over to chat
 
naa- feedback received
 
@SmokeDetector Very old answer on a crap question.
 
@SmokeDetector probably spam, would need a custom flag though: metasmoke.erwaysoftware.com/domains/63744
 
@SmokeDetector tpu-
 
12:37 PM
> Aug 31 '20 at 8:58
Declined spam flag on that.
 
naa- feedback received
 
Yeah, it's not remotely obvious without the context on MS
did you custom flag it?
 
No, NAA.
If the mods won't handle it, perhaps the community does.
Related:
 
[ SmokeDetector | MS ] Potentially bad keyword in answer, blacklisted user (72): User Profile Synchronization Service won't start by user45135 on sharepoint.SE
 
naa- feedback received
 
12:39 PM
Which I also have a rejected spam flag on by the way.
 
@SmokeDetector naa-
 
Apparently they do things differently over there...
 
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer (1): Check if a package exists on Android Market by hasanadiguzel on stackoverflow.com
 
fp feedback received
 
@SmokeDetector Ah one of those uncloseable recommendation requests :/
 
12:41 PM
@SmokeDetector tpu- undisclosed affiliation, username identical to developer.
 
@Mast That spam flag would be very likely to be declined on many sites; I would always flag something subtle like that with a custom flag, because context (from a link) is needed..
@Yatin No, it's a how-to question where people have provided wrong answers recommending apps for a question that asks how to do something programmatically...
...plus an NAA ("Does this work ...") with an upvote. Amazing.
 
@SmokeDetector tp- This website has been developed by Hasan Adıgüzel to be offered to the service of the end user.
 
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, pattern-matching product name in body, pattern-matching product name in title, +4 more (684): supplementsonlinestore.com/life-cbd-oil/ by Wells Norman on apple.SE
 
tpu- feedback received
 
ah, much better...
 
12:45 PM
Some stuff evaporated.
 
Haha
Nice.
 
1:00 PM
[ SmokeDetector | MS ] Toxic answer detected (80): O terminal do Visual studio code não está funcionando corretamente by Nycolazs on pt.stackoverflow.com
 
tp feedback received
fp- feedback received
 
@SmokeDetector I'm interpreting that as "this may be a stupid question, but did you try reinstalling?", but I don't speak Portuguese
 
Google says: "The question seems stupid, but have you tried to reinstall?" But I'm willing to AGF on this as well.
 
Yeah, that's me assuming "the question" refers to "have you tried to reinstall?"
 
[ SmokeDetector | MS ] Link at end of answer, username similar to website in answer (125): USB-C power delivery safe? by Pivoi on apple.SE
 
1:11 PM
Oh, as in referring to their own question. I see.
 
That's their own question?
Ok, that's different.
 
Wait, no, it's not.
 
tp feedback received
 
!!/watch pivoi\.com
 
1:15 PM
[ SmokeDetector ] Blacklists reloaded at rev 0500f60 (SmokeDetector: Auto watch of pivoi\.com by Cody Gray) (running on teward/Osiris)
[ SmokeDetector | MS ] Mostly punctuation marks in answer (36): (a-b)%m Modulo arithmetic by Mangipudi Prashanth on stackoverflow.com
 
fp feedback received
naa feedback received
 
1:29 PM
[ SmokeDetector | MS ] Offensive answer detected, toxic answer detected (158): Why does "position: absolute; left: 0; right: 0; width: XYpx; margin: 0 auto" actually center? by ManiloxGT on stackoverflow.com
 
tp feedback received
naa- feedback received
 
1:42 PM
@SmokeDetector n
 
sdc watch packagenamechecker\.com
 
[ SmokeDetector | MS ] Manually reported answer (94): Read OTP from SMS flutter by user15004183 on stackoverflow.com
 
[ SmokeDetector | MS ] Manually reported answer (94): Read OTP from SMS flutter by cool dude69 on stackoverflow.com
 
2:05 PM
@SmokeDetector k
@SmokeDetector k
 
!!/watch how-to-auto-verify-otp-in-flutter-with\.html
 
@SmokeDetector k
 
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, potentially bad keyword in body (293): Is BioWellnessX CBD Gummies Developed? by biowellsaif on askubuntu.com
 
tp feedback received
 
2:12 PM
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer (1): CNN autoencoder for feature extraction for a chess position by flyer2403 on chess.SE
[ SmokeDetector | MS ] Manually reported answer (batch report: post 1 out of 3) (94): How to check if a linked list is a palindrome or not in Java? by Abhi Patel on stackoverflow.com
[ SmokeDetector | MS ] Manually reported answer (batch report: post 2 out of 3) (94): Remove Duplicates from Linked List by Abhi Patel on stackoverflow.com
 
tpu- feedback received on [MS] Remove Duplicates from Linked List
 
[ SmokeDetector | MS ] Manually reported answer (batch report: post 3 out of 3) (94): How to add preloader loadeing gif to my website by Abhi Patel on stackoverflow.com
 
tpu- feedback received
 
sdc watch technovichar\.com
 
2:16 PM
[ SmokeDetector ] Blacklists reloaded at rev 456b070 (SmokeDetector: Auto watch of technovichar\.com by cigien) (running on teward/Osiris)
 
fp- feedback received
 
sdc watch-force biowellnessx[\W_]*+cbd[\W_]*+gummies(?:[\W_]*+(?:review|\d++|[\da-f]{5,}+)s?)*
 
@cigien That pattern looks like it's already caught by Bad keyword in body and Bad keyword in answer; append -force if you really want to do that.
 
2:35 PM
Makyen/EC2-linux: Executing automatic scheduled reboot.
 
2:50 PM
In an attempt to start collecting bookmarklets, I've created the following gist. Anyone, feel free to run with it, none of them are mine anyway, and let me know if someone has a bigger collection stored anywhere.
@RyanM ^
 
In which I suddenly realize I misspelled "based" as "beased"...
 
iBug/Lily: Executing automatic scheduled reboot.
[ SmokeDetector | MS ] Potentially bad asn for hostname in answer, potentially bad ns for domain in answer (2): "Да сам большой" by Некто on rus.SE
 
3:15 PM
teward/Osiris: Executing automatic scheduled reboot.
Restart: API quota is 15616.
[ SmokeDetector | MS ] Repeating characters in answer (80): Roblox In-Game Ban System ✏️ by Mister_Mistico on stackoverflow.com
 
tpu- feedback received
 
sdc bisect pizzahutcouponcode.com
 
@RyanM Matched by CouponCode\.com on line 1136 of blacklisted_websites.txt
 
okay that's...not exactly the same thing...
but...it is spam...
I feel like blacklisted_websites not requiring word boundaries is gonna bite us one of these days
also I love the edit summary:
> You are recharging from there is fair, and you get that thing cheap, if you use promo code, you will take the same thing, if you do not get the expensive something, then the company does it because you are with the company. Stay connected.
 
3:42 PM
[ SmokeDetector | MS ] Link at beginning of answer (34): Language in chrome changed to Thai from English by Dennis Starkey on superuser.com
 
3:54 PM
[ SmokeDetector | MS ] Few unique characters in answer, luncheon meat detected, no whitespace in answer (225): How to create a custom elementor template library? by QuestionsBoy on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad asn for hostname in answer, potentially bad ns for domain in answer (2): "Да сам большой" ✏️ by Некто on rus.SE
 
[ SmokeDetector | MS ] Few unique characters in answer, luncheon meat detected, no whitespace in answer (225): Why/How is `value="javascript:alert(1)"` considered as a XSS vulnerability in OWASP's ZAP tool? by user15019488 on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Wix Java Search How To Use Multiple Inputs by JC Bluntch on stackoverflow.com
 
4:11 PM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): GetSpectrumData() doesn`t work in OnTriggerStay() Unity 2D by Артур Васильков on stackoverflow.com
 
fp feedback received
 
[ SmokeDetector | MS ] Link at end of answer, potentially bad ns for domain in answer (62): self reset password with MFA by Alex J. on security.SE
[ SmokeDetector | MS ] Potentially bad keyword in answer, username similar to website in answer (65): Review all my changes in git using kdiff3 by Apks Club on superuser.com
[ SmokeDetector | MS ] Blacklisted user (71): Create post button disabled in reddit by Jimmy A. Schmidt on webapps.SE
 
@SmokeDetector odd.. the SU link seems to be to the same question. the second link is not added properly. failed spam attempt?
 
4:26 PM
[ SmokeDetector | MS ] Link following arrow in body (97): Problem with array modifier, incorrect orientation and deformation by Neail on blender.SE
 
fp feedback received
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): How to Make PHP Pagination Start on Page 1 Rather Than Page 0? by N. Scribner on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): What happens to cash used to buy bitcoins? ✏️ by BleepBloop on bitcoin.SE
 
4:47 PM
fp- feedback received on [MS] "Да сам большой"
 
[ SmokeDetector | MS ] No whitespace in title, repeating characters in title (135): sdfsdggggggggggggggggggggggggggggggggggggggggggggggggg ✏️ by Isabella C. LAURSEN on stackoverflow.com
 
5:40 PM
@RyanM Not having bookending on the blacklisted websites does bite us from time to time. When it does, we've added a couple/few different types of bookending to the specific entry. It would be reasonable for us to add it to the baseline, but being thorough in changing it would require searching MS for each entry with and without the new bookending and seeing where there's a difference to determine which are relying on there not being bookending in order to be effective.
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Fiverr I'm new, my given gig is ok? Would you please let me know? by careteam1 on stackoverflow.com
 

« first day (2617 days earlier)      last day (1190 days later) »