« first day (2615 days earlier)      last day (1204 days later) » 

12:03 AM
[ SmokeDetector | MS ] Link at beginning of answer (34): Change Data Capture - SQL SERVER? by Ryan on stackoverflow.com
 
tpu- feedback received
 
@cigien On the site they moderate, moderators can see the usernames which the user has used on that site. Regular users have no direct way to see prior user names. An SEDE query could get you the username which the user was using in the prior week and you could use archives of the data dump to get even older names.
Three other ways to find old usernames are:
⠀⠀⠀A) search in the comments left on the user's older posts, which can contain an @ mention with an old username;
⠀⠀⠀B) look at the responses to the user's chat messages, which will contain an @ mention; and
⠀⠀⠀C) look at the user's list of "recent" chat messages, which will retain the username which was in use at the time it was posted. A hint with the user's "recent" chat message list is that the data is actually paginated, so you can obtain earlier messages by the user by appending a page parameter to the URL.
There are userscripts which add pagination buttons on chat's user history pages. The two UI looks I'm aware of: spacemonaut's Chat User History Pagination and Sam's Chat Pagination, which is also a part of his Chat Improvements script.
 
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)
 
12:23 AM
fp- feedback received
fp feedback received on [MS] MKV to AVI convertor?
fp feedback received on [MS] My wordpress blog posts page is broken
 
[ SmokeDetector | MS ] Potentially bad ns for domain in body (1): Accidentally printed a chocolate mold instead of the object itself by WhatsYourFunction on 3dprinting.SE
 
12:39 AM
@tripleee Yeah, I, also, have been thinking about something along those lines. Specifically, I've been thing about refactoring findspam.py into a file, findspam.py (or something like "postprocessing.py"), which has all the code which processes posts through the list of detections and something like "detections.py" which contains just the detections. Then we could allow for the existence of something like "detections_custom.py", which, when it exists, could be used instead of "detections.py".
Along with that, I've been planning on moving which SE WebSocket to use into the config file (at least optionally). Doing so would allow people to more easily set up an instance which is just looking at one specific site (or a list of sites), rather than all of SE. At the same time, move the SE API key into config, as such uses aren't really "SmokeDetector" and legitimately should use a different SE API key (and the larger API quota isn't needed to look at one site, even with a queue depth = 1).
 
[ SmokeDetector | MS ] Post is mostly images (14): VBScript script progress notification by Mark Llewellyn on stackoverflow.com
 
fp- feedback received
 
1:06 AM
[ SmokeDetector | MS ] Link at beginning of answer (34): What's the best plugin for a membership based site by Boby Bravo on stackoverflow.com
 
@tripleee As far as I'm aware, chat message throttling is per-user, not per-IP address. However, at the moment, the SD and MS users are being posted from different IP addresses, so even if the throttling is per-IP address, what one does wouldn't affect the other in the current configuration.
@tripleee A post is only considered for autoflagging the first time it is reported to MS. The post is also disqualified from autoflagging if it's been edited prior to the report. OTOH, if a matching spam wave is configured, then an additional report will result in the post being autoflagged. I suspect a spam wave will override the limit on not autoflagging edited posts, but I have neither looked at the code nor had a case which I remember that indicated it was one way or the other.
 
@Makyen Thanks for the information. C works quite nicely. And the "recent" chat messages go back indefinitely, so this lets me at least compile a list of all usernames a profile has ever posted a message with in some chat room.
 
sdc watch billspk\.com
 
@cigien Yes, (C) works well, but only if the user has been active in chat. :;
 
1:31 AM
@Makyen True :) Do those data dump archives go back indefinitely? That would allow one to search for all of them, in theory at least.
 
@cigien I haven't looked.
 
@Makyen Ok. No worries, I'll investigate it further myself if I actually end up needing to do this :) Thanks for the tip.
 
np
 
sdc watch blitzz\.io
 
[ SmokeDetector ] Blacklists reloaded at rev 321fa4f (SmokeDetector: Auto watch of blitzz\.io by cigien) (running on teward/Osiris)
 
@thesecretmaster I have some questions: what advantages does managing the blacklists/watchlists on or via MS gain us? What additional dependencies does it introduce (e.g. would we still be able to update the lists when MS doesn't exist, which is a common enough occurrence such that it would be a substantial concern)? If it does introduce additional dependencies, do the expected positives outweigh the potential negatives.
Does having the blacklists/watchlists managed by MS make it such that if someone wants to set up their own instance of SD (which is done by some people, and is definitely done when testing new code for SD), would they now need their own instance of MS in order to have the blacklists/watchlists?
If the blacklists/watchlists are managed by MS, how are we going to interface with those lists? For example, are we still going to have them in GitHub too? Are we going to be able to edit the lists locally in our own editor of choice and push changes (something I do routinely)?
 
@SmokeDetector tp- edit only adds offensive content
 
tp feedback received
 
fp- feedback received
 
[ SmokeDetector | MS ] Offensive answer detected, potentially bad keyword in answer, toxic answer detected (159): How do I relight the pilot light on my Magic Chef stove? by Andrew on diy.SE
 
tpu- feedback received
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): Boys function for Gaussian integrals in ab-initio calculations by R. J. Mathar on chemistry.SE
 
2:28 AM
[ SmokeDetector | MS ] No whitespace in body, no whitespace in title (167): dsjkljlg123kdkfkfkggh ✏️ by jsdifjew on math.SE
 
@SmokeDetector v
 
[ SmokeDetector | MS ] Link at end of body (99): DrupalVM not working by Thakur on drupal.SE
 
fp- feedback received
 
Makyen/EC2-linux: Executing automatic scheduled reboot.
 
2:50 AM
[ SmokeDetector | MS ] Link at beginning of answer (34): Interaction effects between latent variables in SEM in lavaan by Lanny on stackoverflow.com
[ SmokeDetector | MS ] Link at beginning of body (42): Can someone please answer this survey? Use the Link below plz by no one on stackoverflow.com
 
tpu- feedback received
 
iBug/Lily: Executing automatic scheduled reboot.
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Leakage and bias in XGBoost trading strategy by Federico Juvara on quant.SE
 
fp- feedback received
 
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted username, blacklisted website in body, potentially bad ip for hostname in body, +2 more (385): Warum sind Sexpuppen dieses Jahr beliebter? by kobedoll on photo.SE
 
tpu- feedback received
 
3:01 AM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): can someone help me identify this type of camera? by user96977 on photo.SE
[ SmokeDetector | MS ] Link at beginning of answer (34): UEFI Boot options menu at every boot : how to make it work? by Jenny Feng on superuser.com
 
@Makyen Oh, I thought that most of these issues had been discussed back when helios was initially proposed, and the one barrier was dev-time. My first thoughts on advantages would be more automation around blacklist management, and things like treating terms differently if they haven't been seen in years or something. It would also allow fewer git pulls and reboots which sounds like a good thing to me, although I don't really do much with smokey so I'm not sure.
SD would still be able to add and delete *list items, so it wouldn't require MS to be up, although MS would be an alternate way to add/remove items. I believe it would live on an AWS lambda (at least that was the design last time I heard about it), so downtime is very unlikely.
 
[ SmokeDetector ] SmokeDetector started at rev 321fa4f (SmokeDetector: Auto watch of blitzz\.io by cigien) (running on teward/Osiris, Python 3.6.9)
Restart: API quota is 6267.
 
I think we'd want public read-only access and only require a key for writes, so any test SD instance could pull the lists without any authentication
It would break the workflow of manually editing those files, although my hope would be that any of those workflows could be integrated into MS.
 
3:24 AM
iBug/Daisy: Executing automatic scheduled reboot.
 
@thesecretmaster The git pulls and reboots for strictly only blacklists/whitelists (BL/WL) changes are fairly innocuous at this point. For changes which are to only files which contain the BL/WL, SD just pulls the current state of the repository from GitHub and re-reads the files, repopulating the primary BL/WL regular expressions. It's definitely not a full reboot.
Other than the actual pull from GitHub, which is relatively lightweight, there's not a lot which could be saved in there without substantial restructuring. Sure, we could just re-read only the list that actually changed, but we could do that now too.
If we're getting strictly an "add" notification, I guess we could have a separate, short list for each that is updated to save re-processing the whole list, but we'd have to do a full sync-up periodically and for any removal, other than removals from just the short list.
For changes to the files other than the BL/WL files, SD does do as much of a reboot as it ever does w/o someone restarting it at the console.
 
Oh, ok, then I guess there's not much use for a non-github *list manager
 
@thesecretmaster I was largely out of the loop when that was discussed. It was prior to when I was more engaged in the project, so I'm not really familiar with what helios really is, or was intended to be. For me it's always been a fairly nebulous concept as to what it was really intended to accomplish.
 
Honestly it's before my time as well, I just know the 3 MS issues at the very bottom of the issue list that refer to it :P
I thought it would also be a nice-to-have for if other projects wanted to use our lists, but I don't think that's a realistic use-case or one we should worry about at all.
 
@thesecretmaster What would be beneficial is cross-SD instance management of the various lists which are held on a per-instance basis (i.e. that are not on GitHub). That would be things like the user blacklist, user whitelist, chat user notify lists, post ignore list.
 
3:33 AM
Ah, that might be the problem Helios was intended to solve?
Possibly that was the original idea, then black/watchlists got added because there was no reason to keep them in git if they could go in Helios?
 
I was under the impression that it would, but really am not familiar enough with helios to be able to say what it really was going to do.
 
Hmmm I feel like there was a specification somewhere.
I'll go dig
 
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): What would the economy look like on a generation ship? by Tod3 on worldbuilding.SE
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): How to fight browser fingerprinting? ✏️ by cloud em on security.SE
 
It looks like it was notifications and black/watch lists?
 
@thesecretmaster Well, it's something that could be solved with either holding those lists somewhere else, or having the lists held by each instance, but letting the instances communicate in some manner.
For instance, it could be solved by having a WebSocket relay where an SD instance could either send via the WebSocket or POST and have arbitrary JSON data relayed to all the other instances. Then the instances could maintain a local list while getting updates (or any other communication they wanted/needed). There would need to be some data-sync logic in SD, but it shouldn't be too bad.
@thesecretmaster Hmmm... Ok let me take a look.
 
3:56 AM
[ SmokeDetector | MS ] Potentially bad keyword in body (1): can someone help me identify this type of camera? by user96977 on photo.SE
 
@Makyen ...I thought we had that
Will the smoke detector WS channel not do that already?
 
@thesecretmaster For the BL/WL, it sounds like a lot of the reduction in turn-around time which that proposal was intending to gain was not having to wait for CI for BL/WL changes. The full CI is ~2 minutes, or so. However, SD no longer waits for CI to complete in order to make a change to the BL/WL. It just runs through (some of?) the checks which are specific to the BL/WL (e.g. no duplication; no detected by other detections; etc.) and then just makes the change.
Yes, overall, it does take substantial/noticeable time. A portion of that is git/GitHub, but there's also a portion of that which is reloading the lists, rebuilding the regexes (which could be substantially less in many cases with some architecture changes; e.g. breaking the lists), etc.
Overall, it would be interesting to get actual data as to which parts of the process are taking what time, and what those parts are actually costing us (i.e. are we doing other processing while waiting), other than the wall time (e.g. what threads are running concurrently).
Hmmm.... thinking about this, there are some concurrency issues which I'm note sure are appropriately handled. I'd need to look at the code which does all this in more substantially more depth.
 
[ SmokeDetector | MS ] Bad keyword in body, potentially bad keyword in body (96): CBD: Good or Bad? by cyndibrant on medicalsciences.SE
 
@ArtOfCode I've asked if it was possible and, IIRC, you said it wouldn't be hard to do. I'm not aware of it actually existing. I know that MS can send arbitrary JSON data to SD instances, but I'm not aware of a way to have SD pass data to MS to have it directly forwarded to the WebSocket (e.g. received by all instances as {"from_sd": "[Whatever SD sent here]"}). It could exist and I'm just not aware of it.
(or something like that)
 
tpu- feedback received
 
@Makyen I... don't know, to be honest. Need to read the code :)
 
!!/watch theinspirespy\.com
 
@Machavity That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer; append -force if you really want to do that.
 
@ArtOfCode :) We should probably also decide if the data should be sent via the WebSocket or if it should be sent just as part of the response to status pings. It may be sufficient for such messages to be sent just in response to the status ping. Doing so wouldn't add the requirement that the standby instances keep the MS WebSocket open. Although it shouldn't be too bad to add that requirement, given that it's low traffic.
 
Easier technically to send it via WS, so we don't have to store it and keep track of what's been sent where
 
4:23 AM
If it's easier for SD to get it via status pings, it could go in redis with a short TTL
 
API quota rolled over with 5851 requests remaining. Current quota: 19999.
stackoverflow: 1894
math: 1177
ru.stackoverflow: 920
superuser: 550
askubuntu: 493
physics: 471
es.stackoverflow: 422
electronics: 370
unix: 362
stats: 320
gis: 304
pt.stackoverflow: 259
serverfault: 252
tex: 243
mathoverflow.net: 222
blender: 208
salesforce: 205
ell: 187
dba: 178
worldbuilding: 165
mathematica: 163
diy: 157
english: 140
wordpress: 128
apple: 122
codegolf: 120
magento: 119
codereview: 109
politics: 96
scifi: 94
gaming: 88
ethereum: 88
sharepoint: 87
rpg: 87
drupal: 84
security: 83
 
tpu- feedback received
 
4:56 AM
[ SmokeDetector | MS ] Potentially problematic ns configuration in body (1): 404 request to "?standalone=true" with SSL by Rooholamin Honarvar on stackoverflow.com
 
fp- feedback received
 
tpu- feedback received
 
sdc watch- aspiresoftwareconsultancy\.com
 
@ppwater The regex contains an unescaped "."; in most cases, it should be "\."; append -force if you really want to do that.
 
5:01 AM
PR#5657 ("ppwater: Watch aspiresoftwareconsultancy\.com") opened by SmokeDetector
 
[ SmokeDetector | MS ] Link at beginning of answer (34): Looking for a desktop app to sort my photos library by face by a deleted user on softwarerecs.SE
 
!!/approve 5657
 
sdc approve 5657
 
Hmm, are 2 merges like that an issue?
 
5:03 AM
OK. Only merged once.
There were two reloads of the blacklist (I assume), but there's only one merge.
 
I see. If 2 merges do happen, I take it they should be unwatched to remove them both, and then watched again?
 
That would be a reasonable way to deal with it.
 
Ok, thanks.
 
np. Thank you.
 
fp- feedback received
 
5:09 AM
[ SmokeDetector | MS ] Potentially problematic ns configuration in answer (1): Does copyright law apply if I download films in Iran? by Morteza on law.SE
 
tpu- feedback received
 
@Makyen no but there is a (human) race condition in that "fp- feedback received" from metasmoke which corresponds to the latest SD report metasmoke had seen at the time might not be the latest SD report in chat ... maybe those too should be linked now
 
sdc watch vakiltop\.com
 
re: chat search I have used the RSS search button which comes up on the lower right in the chat search results, though (weirdly) that doesn't restrict the search to a single chat room
 
5:13 AM
@tripleee Yeah, I've also thought that just having a single format for the MS feedback messages would be a good idea. There's always going to be the possibility of a race condition there, even once the MS messages are back to being in the main SD instance (as coming from the metasmoke user).
 
[ SmokeDetector | MS ] Potentially bad ns for domain in body (1): Why are world maps cut off? by a deleted user on gis.SE
 
fp- feedback received
 
@SmokeDetector that is an image of a map. Nothing to do with homoeopathy
 
@tripleee IIRC, that RSS feed ignores both the room and user restriction. OTOH, if you have custom code which is receiving it, rather than just a viewer, you could filter it upon receipt.
 
5:28 AM
@thesecretmaster Aye, but then you have to keep track of what instance has had it or not... it's just easier to be stateless and dump it straight to the socket
 
Yeah, I mean stateless is always easier
 
tpu- feedback received
 
sdc bisect aakash.ac.in
 
@ppwater Matched by aakash\.ac\.in on line 24075 of watched_keywords.txt
 
5:32 AM
As a student who studied in Allen, I have solemnly sworn to hate Aakash Institute :P
sdc bisect NCERT Solutions
 
@Yatin 'NCERT Solutions' is not caught by a blacklist or watchlist item.
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer (1): CPQ - Disable Proration by salesforceninja on salesforce.SE
[ SmokeDetector | MS ] Potentially bad keyword in answer (1): How to run Android ICS on a PC (win7) inside a VM? by William Acker on android.SE (@AndrewT.)
 
tpu- feedback received
 
sdc watch- fingerskilz\.tv
 
PR#5658 ("ppwater: Watch fingerskilz\.tv") opened by SmokeDetector
 
sdc approve 5658
 
@SmokeDetector tp undisclosed affiliation
 
@cigien Registered answer as true positive. If you want to blacklist the poster, use trueu or tpu.
 
fp- feedback received on [MS] CPQ - Disable Proration
 
@metasmoke Why is that FP? @tripleee
 
@Yatin one more and we can blacklist them
@cigien by bad, thanks for the ping
 
[ SmokeDetector | MS ] Link at beginning of body (42): What kind of token does 'Addlogin' method take? by pswffox on stackoverflow.com
 
fp- feedback received
 
@tripleee Oh my old classmates would love that news xD
 
sdc watch salesforceninjacom\.wordpress
 
@ArtOfCode Yep, stateless is substantially easier on MS. It doesn't increase the load too much on a standby instance, assuming the standby instance is on a machine which is otherwise always on. OTOH, it does change the requirement away from the standby instance being able to completely power down between pings to needing to be on and available to receive the WS 100% of the time. I guess wake-on-LAN sleep states would still be possible for a "standby" instance.
 
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, potentially bad keyword in body (194): daisy may cooper keto by OrphKoc on apple.SE
 
@SmokeDetector that needs to be migrated
 
@Yatin your comment is very brief, people who end up posting on meta often have no idea about the surrounding network etc
 
6:11 AM
@SmokeDetector fp- mod flagging for migration
@tripleee Oh
Which SE site was that? Where people asked for the song name?
 
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)
 
@Yatin It's a Beta site, musicfans.
 
@Yatin the ones where it would be topical I think would require a bit more background research though
 
@cigien Yep that is the one, thanks :)
@tripleee meaning?
 
@Yatin I suspect triplee is referring to "Lacks research effort". That question might be downvoted for that, but it's definitely on-topic on musicfans.se
 
6:18 AM
Ohk
 
6:39 AM
[ SmokeDetector | MS ] Url-only title, bad keyword in body, bad keyword in title, body starts with title and ends in url, link at beginning of body, +5 more (631): mfdoomcbdoil.wixsite.com/mf-doom-cbd-oil/ by Emiliaboncak018 on askubuntu.com
 
tpu- feedback received
 
sdc bisect mfdoomcbdoil.wixsite.com
 
@ppwater Matched by wixsite\.com on line 195 of watched_keywords.txt
[ SmokeDetector | MS ] Potentially bad ip for hostname in answer, username similar to website in answer (65): How much should I feed my puppy? by home petsusa on pets.SE
 
tpu- feedback received
 
!!/watch homepetusa\.com
 
6:42 AM
sdc bisect homepetusa.com/
 
[ SmokeDetector ] Blacklists reloaded at rev 0b9adf2 (SmokeDetector: Auto watch of homepetusa\.com by tripleee) (running on teward/Osiris)
@ppwater Matched by homepetusa\.com on line 26329 of watched_keywords.txt
 
lol
 
tpu- feedback received
 
6:43 AM
!!/watch (?-i:69fU72GxTuA)(?# youtube)
 
Ah, dang, ninja'd by @tripleee... was just looking up the right regex :-)
 
@tripleee Is it okay to do a space after the #?
@CodyGray same :)
 
@ppwater yeah, it's just a free-form comment
 
Ah. thanks
 
6:45 AM
[ SmokeDetector | MS ] Potentially bad ns for domain in answer (1): TypeORM: Query Many-to-Many with Custom Properties by scott_long on stackoverflow.com
 
And do I have to leave a comment when I am doing a youtube watch?
 
We can watch that username too. I don't think they are up to any good: On MW tech channel you will find more interested and earning videos and able to earn yourself.
 
Yeah, you wouldn't want to match the literal string "youtube", since they have other alternate URLs, like "youtu.be" or whatever.
 
fp- feedback received
 
@Yatin The name of what user? :-p
 
6:46 AM
!!/watch mw[\W_]*+tech
 
@CodyGray Heh
 
the name that used to be where the smoking crater is now emitting radiation
 
Do we see a lot of benefit from doing that, anecdotally?
If the spammer creates a new account with the same PII, it should already be caught by the system and suspended.
So they'd have to use different PII so as to have the system not connect it to the original account, but the same display name.
 
everything in watched_keywords.txt which has been there for more than a few weeks is scattershot anyway; some of them surprisingly deliver value after a long time
 
[ SmokeDetector | MS ] Manually reported question (93): How can Agriculture crops can be prevent from lots of damage? by Ipl biological on stackoverflow.com
 
tp feedback received
 
@CodyGray You don't use AF?
 
No
 
You can just select Spam in it and it will report it here for you :)
 
6:53 AM
I haven't even managed to get FIRE authorized on a site where I'm a mod.
 
!!/watch international[\W_]*+panaacea
 
@CodyGray Oh
 
I don't know if AF works well for mods. Haven't tried it.
 
[ SmokeDetector | MS ] Potentially bad asn for hostname in answer, potentially bad ip for hostname in answer (2): Intellij settings set to CNY currency but I want $ instead by Aparna Dey on stackoverflow.com
 
6:54 AM
Oh, I didn't know that
 
Since mod spam flags are binding and immediately delete the post, is the script aware of that? Does it delay for a sufficient amount of time for SD to scan it?
 
fp- feedback received
 
@CodyGray Ah makes sense
 
sdc watch-force daisy[\W_]*+may[\W_]*+cooper[\W_]*+keto(?:[\W_]*+(?:buy|order|\d++|[\da-f]{5,}+)s?)*
 
I don't know
 
6:57 AM
@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.
 
sdc watch-force mf[\W_]*+doom[\W_]*+cbd[\W_]*+oil(?:[\W_]*+(?:\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.
 
@ppwater generally the point of the comments on these is so that we can tell the random identifier being watched from the various other sites that use random identifier. They're at least strongly encouraged.
 
So, the same reason you write comments everywhere ;-)
 
Yup!
Anything that's not obvious from looking at it
 
7:03 AM
Good thing all of my assembly code is obvious!
 
[ SmokeDetector | MS ] Bad keyword in answer, link at end of answer, pattern-matching website in answer, potentially bad ns for domain in answer (226): Does bcrypt have a maximum password length? by Putin on security.SE
 
@RyanM Ah, thanks
 
tpu- feedback received
 
sdc bisect cbdoilslist.com
 
@ppwater 'cbdoilslist.com' is not caught by a blacklist or watchlist item.
 
7:09 AM
sdc watch- cbdoilslist\.com
 
@ppwater That pattern looks like it's already caught by Bad keyword in answer and Bad keyword in body; append -force if you really want to do that.
 
Huh
sdc bisect cbdoilslist.com
 
@ppwater 'cbdoilslist.com' is not caught by a blacklist or watchlist item.
 
sdc test cbdoilslist.com
 
> Bad keyword in body, bad keyword in title, bad keyword in username, pattern-matching website in body, pattern-matching website in title, potentially bad ns for domain in body, potentially bad ns for domain in title
----------
Bad keyword in body - Position 0-6: cbdoil
Bad keyword in title - Position 0-6: cbdoil
Bad keyword in username - Position 0-6: cbdoil
Pattern-matching website in body - Position 0-15: cbdoilslist.com
Pattern-matching website in title - Position 0-15: cbdoilslist.com
Potentially bad ns for domain in body - cbdoilslist.com NS suspicious sam.ns.cloudflare.com.,elinore.n
[ SmokeDetector | MS ] Potentially bad keyword in body (1): Binance Future Quantity Number by S M on stackoverflow.com
 
7:10 AM
fp- feedback received
 
@SmokeDetector might be off-topic but idk
 
looks like general computing to me
or finance, or $pick_your_favorite_cryptocurrency
 
[ SmokeDetector | MS ] Bad keyword in answer, bad keyword with email in answer, pattern-matching email in answer, potentially bad keyword in answer (270): Simple bank-atm system design OOP by Miss Jessica55 on codereview.SE
 
> Binance is a cryptocurrency exchange website. This tag is for questions related to the Binance API and usage of Binance API wrappers. This tag is not specific to any programming language
Tag info^
 
tpu- feedback received
 
7:14 AM
[ SmokeDetector | MS ] Potentially bad keyword in username (1): After power hit, wallet won't load by Antier Solutions on bitcoin.SE
 
They haven't added any programming language so I think so too @RyanM
 
fp- feedback received
 
@ppwater It appears to be a check in findspam.py.
 
@cigien Ah, thanks
 
np
 
[ userscripts ] continuous-integration/travis-ci/pr success on ba2ddfb: The Travis CI build passed
fp feedback received
 
[ SmokeDetector | MS ] Blacklisted website in body, potentially bad ip for hostname in body (100): Does anybody have any knowledge of a successful ERP school management system module? by priti on stackoverflow.com
 
sdc watch- dienmayhoanglien\.vn
 
PR#5659 ("Yatin: Watch dienmayhoanglien\.vn") opened by SmokeDetector
 
7:29 AM
!!/approve 5659
 
fp feedback received
 
[ SmokeDetector | MS ] Potentially bad keyword in body (1): AWS Lightsail 'Exception occurred processing WSGI script' by Kyle McReynolds on stackoverflow.com
 
7:34 AM
@Yatin Yeah, someone pointed out in SOCVR that the poster had that link in their profile.
Otherwise, it looked like a legitimate, if off-topic, question.
Regardless of the fact that previous spam has been posted with that same link.
 
Yep
 
[ SmokeDetector | MS ] Manually reported question (93): How to translate the feeling into sound by coding? ✏️ by yegane on stackoverflow.com
 
tpu- feedback received
 
!!/watch- (?-i:q5yxIzs5Wug)(?#youtube)
 
Wait
 
@CodyGray Why are you watching a TedX talk?
 
Why would someone post a TedX video
@cigien Yea
 
Why does anyone watch TedX talks?
 
Good point :) Doesn't make it spam though.
 
7:37 AM
Looks like spam to me.
 
Yea
Off-topic, not spam
 
What would it need to have to make it spam?
 
@CodyGray Promoting some service or product.
 
Be in a tin can?
 
I mean.. TedX isn't gonna spam us
 
7:38 AM
Uh, it promotes a YouTube video.
@Yatin You say that, but then we have ongoing waves of spam from Microsoft. So....
 
Um... I don't really think so... I feel like they just wanted to have a discussion on that video... The video was the reference material... Microsoft spam was different. It was on mass. This is one post
Again, off-topic for sure
But not spam
 
We just did an identical example of someone promoting their YouTube video.
 
fp feedback received
 
How do you know that that was their video?
 
7:40 AM
@SmokeDetector fp- OP has linked to TedX talk about coding and music, and wants to know if coding will replace songwriting. Way off-topic on SO, but not spam.
 
^
 
This is materially identical to this previous case, where the person was promoting a YouTube video.
Or, maybe that person just wanted to start a discussion about the best way to do a search on Fiverr using keywords?
Maybe they were writing a Python program to do it and needing programming help?
It's fun to make up scenarios, I guess, but that doesn't make the unsolicited promotion of unwanted, blatantly off-topic content any less spam.
 
The username matched there. The video was clearly spam. That channel was meant to earn money. TedX is a non-profit organisation that makes videos on YouTube. The people who speak there are normal people with stories to tell. TedX won't spam us. :/
 
I'm fascinated that you know the motives of all spammers and can predict the future.
If you make a list of all the people who won't ever spam us, maybe we could make that into a whitelist.
 
I can't but I like to give the benefit of the doubt
 
7:44 AM
[ SmokeDetector | MS ] Offensive answer detected, potentially bad keyword in answer, toxic answer detected (159): What do you call someone who doesn't know how to swim? by user411432 on english.SE
 
tpu- feedback received
 
Again, I agree that that question had to go. But I don't want OP to face any repercussions for being labelled as a spammer.
 
Well, that ship has long sailed.
 
Hm
Np
 
@CodyGray That video is about making money using tricks that the OP has come up with. It's also very similar to other spam. It's also promoting their own service. I agree that it's not always cut and dried, but these 2 cases are quite different beasts, and I'm comfortable with the FP feedback on the TedX post vs the TP feedback on this one.
 
7:46 AM
I think it's horribly wrong to try and assess the nature of the spam.
"Oh, this seems harmless, so it must not be spam."
It's just like what we get all the time: "Oh, I'm not spamming. You see, this project is open source!"
You legitimately think I need to watch the video in order to decide whether something is spam?
This is another one of those discussions we had before: do I need to go to the site and download the file to see if it contains malware?
 
[ SmokeDetector | MS ] Manually reported answer (94): No main manifest attribute, in tempCodeRunnerFile.jar by Ihte Sham on stackoverflow.com
 
fp- feedback received
 
@SmokeDetector Ah sorry, mis click
Ah
 
OK, I was wondering...
 
@Yatin why do you report naa
 
7:53 AM
@SmokeDetector naa- Accidently reported by me while reaching out for NAA in AF
 
Ah
 
@CodyGray I haven't been participating in this project long enough to feel comfortable making broad generalizations about what is or isn't spam. At the moment I'm learning how to do that by assessing individual posts (and using additional context where I can of course). As I said, in this case, one is undoubtedly TP and the other FP, for the reasons I stated.
It's not for me to tell you how much to dig into a post's contents; whether to watch the videos, or download potential malware, or whatever. I will point out that you have at least on one ocassion given TP feedback on a post linking to a completely innocent youtube channel that was relevant to the question being asked.
 
@Yatin Could you delete the misclick comment? (in MS)
 
In MS?
 
@ppwater Why? It was a misclick.
 
7:54 AM
Yes
 
would make sense to leave it actually
 
But yatin left another comment
 
Removed the first one :)
 
Thanks :)
 
!!/watch h[oó]a[\W_]*+ch[aấ]t[\W_]*+r[uử]a[\W_]*+xe
 
"car wash chemicals" in Vietnamese
somewhat to my surprise, the character classes there actually seem to work
>>> regex.findall(r'h[oó]a[\W_]*+ch[aấ]t[\W_]*+r[uử]a[\W_]*+xe', 'Vì sao bạn nên sử dụng hóa chất rửa xe thay vì xà bông thông thường dienmayhoanglien.vn/hoa-chat-rua-xe-chuyen-dung.html')
['hóa chất rửa xe', 'hoa-chat-rua-xe']
 
[ SmokeDetector | MS ] Bad keyword with email in body, potentially bad keyword in body (57): The best Taxi Services in Gloucestershire? by Sabscars on meta.SE
 
tpu- feedback received
 
[ SmokeDetector | MS ] Bad keyword with email in answer, potentially bad keyword in answer, username similar to website in answer, blacklisted user (220): The best Taxi Services in Gloucestershire? by Sabscars on meta.SE
 
tpu- feedback received
 
7:59 AM
sdc watch-number +44 (0) 1452 933 689
 
!!/watch sabscars\.co\.uk
 
!!/watch sabscars(?!\.co\.uk)
 

« first day (2615 days earlier)      last day (1204 days later) »