@SmokeDetector This looks suspicious: website splinxtechnologies (see address at bottom of page), and (from username) cITsecure are both based in "Eden Prairie, Minnesota 55344", especially given the website has only no other hits on SE, and one likely spam post a few days ago.
The quota_remaining property was not in the API response. Error: no method found with this name at 02:33:00 UTC. Backing off on requests for the next 12 seconds. Previous URL: https://api.stackexchange.com/2.2/questions
The quota_remaining property was not in the API response. Error: no method found with this name at 02:33:07 UTC. Backing off on requests for the next 12 seconds. Previous URL: https://api.stackexchange.com/2.2/questions
@SmokeDetector fp- So it seems this isn't actually spam, despite looking very spammy. From the declined message on my spam flag "This passage appeared in newspapers in Dec 2005 and now appears in a Vietnamese online test.". The question has also been edited on the site to include attribution to the original source.
@Makyen I don't see how it's plagiarism. The user isn't taking credit for the text, or in any way claiming to pass it off as their own work. They're asking for the meaning of "them" in the text, that's all. It's missing attribution, but that wouldn't count as plagiarism, would it?
@Makyen Ehhh... the original title was "what are them in this passage?". So while they should have made it more clear by adding the source, I don't think they had bad intent given they implied it was a passage from something else...
Stack Exchange questions and answers are written by you, the members of the Stack Exchange community. Your work is valuable to the community, and you get credit in the form of reputation for the value that you create. Like many institutions, such as schools, journals, and news organizations, Stac...
@SmokeDetector fp- likewise, from the discussion above, this appears to be a cross-site duplicate, asking how a particular word is used in a passage, and not intended to spam anything. Has been deleted on the site, with (at least) my spam flag marked as "disputed".
@cigien I agree my response stating plagiarism was a bit flippant.
I did, at the time I first saw the post and gave TP feedback, recognize that it could be an actual question, but, IMO, it wasn't appropriate for SE in that form, even if it was intended as an actual question.
I took into account the circumstances surrounding post, including the combination of subject matter (which some people will find offensive); lack of indication that the OP was quoting something (although there is the indication of emphasis placed on "them" in the body and the question in the title, which tends to lean a bit towards a real question, but is also consistent with things many spammers/trolls do);
that the question was in that form, effectively, spam; that the user is both unregistered and created their account simultaneously with asking the question; and that the user is using a generic username.
@cocomac Regardless of my response being a bit flippant, it, technically, is plagiarism, because it did not meet the minimums of the referencing requirements for SE. If that in and of itself justifies TP feedback in this case is, at best, debatable. I also recognize that the users subsequent actions tend to indicate there wasn't the intent to claim it was their own work.
I want to test some text (either a title + body content, or just answer content) against the SD tests. I'm aware of !!/test, but I'd like to do this offline (no, I'm not trying to evade the tests). Is this possible...?
@cocomac Of course it's possible. How easy it is to do is another story. Probably the easiest way to accomplish that is to leverage the existing tests in test_findspam.py. You could do that by adding it as one of the parameterized tests for test_findspam() and add some additional log() or print() output to more easily see what you're interested in, along with forcing the test to fail, so the output from the test is displayed.
@cigien That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body Append -force to the command word(s) if you really want to add the pattern you provided.
@Makyen Hmmm... I'll look more tomorrow, but if I were to add a function to do that and get the numeric value of spam likeliness (like 1 for potentially bad keyword), which file should I start looking in? (I'm looking to try this on an SE data dump, so looking by hand would be rather slow. Yes, I'm aware this will be slow. I intent to start on one of the smallest sites for testing. I have a machine with a large amount of RAM + CPU to try)
I'm trying to see if I can find any random ancient spam posts that fell through the cracks like this one
the people who are active in this room have probably already clicked on it, but I just brought it up in the Tavern (which seems pretty quiet right now too, though)
@Makyen I forgot to ask it explicitly yesterday, but do we want to enable the "username similar to website in {}" reason for questions as well? If so, and assuming the fix I mentioned is correct/appropriate, should I go ahead and make an issue/PR?
@cigien That appeared to be disabled almost immediately after it was created because it showed a substantial number of FP. Unless we have reason to believe that's no longer the case, it seems like it would be a good idea to try to get some data prior to turning it on.
@Makyen Oh, I didn't realize it was intentionally disabled; I couldn't find any discussion of this in the transcript. Ok, I'll hold off on the PR then. By testing, you mean try out the rule in a separate SD instance, and see how it performs?
or you could make a slightly more involved PR to make it an experimental rule "potentially XX"
but even then separate testing before merging would probably be a good idea
gist.github.com/tripleee/f586ea491298e401d860b17483bcf99e has an old snippet for testing findspam.py locally on search results from metasmoke (haven't reviewed it; it probably has some old cruft I no longer use, and/or needs changes to work correctly with current finspam.py)
@cocomac you mean have the tag added? sure I can do that; we might also consider marking it as whitelisted in findspam.py if the problem is substantial
actually the post was reported because request.host in the Rails code was evaluated as a host name, and resolved to the watched IP address 45.13.252.116
... not even watched IP, it's just in an AS which is watched
and also the name server for that domain is watched
but a common bulk registrar with many FPs
in fact our whitelist should probably include common SQL terms etc which look like domain names like user.id, host.name, and so on (and Perl script names like scan.pl)
[ SmokeDetector | MS ] Bad ip for hostname in body, bad ns for domain in body, blacklisted website in body, link at beginning of body, potentially bad keyword in body, +1 more (336): Eaze Promo Code by jasonholder on worldbuilding.SE
[ SmokeDetector | MS ] Link at beginning of body, potentially bad asn for hostname in body, potentially bad ip for hostname in body, potentially bad keyword in body, potentially problematic ns configuration in body (44): Bulk SMS Provider Delhi by Staticking5 on superuser.com (@Ollie)
[ SmokeDetector | MS ] Potentially bad asn for hostname in body, potentially bad ns for domain in body, potentially problematic ns configuration in body (3): Is lens a type of mirror? by Adarsh P S on physics.SE
[ SmokeDetector | MS ] Bad ip for hostname in answer, blacklisted website in answer, potentially bad ns for domain in answer, potentially bad keyword in username (193): vcf file import error by paylorsmith on stackoverflow.com
[ SmokeDetector | MS ] Potentially bad asn for hostname in body, potentially bad ip for hostname in body, potentially problematic ns configuration in body (3): Tplink router setup by praveen kumar on superuser.com (@Ollie)
Maybe there aren't enough people around here with an account on MathOverflow to actually help...? (Of course, I understand that removing spam on that site should be primarily the concern of MathOverflow users.)
@JeffSchaller That pattern may be a North American number. If it is, please use a format which starts with an optional 1 followed by possible separator text and has the main number in the format 829-580-0749 where - could be a single alpha character or any [\W_]*+. Alternately, you can add the comment (?#IS NorAm) to the end of the pattern to force also using the alternate normalized form, or (?#NO NorAm) if it's not a North American phone number and it's incorrectly recognized as one. Perhaps try !!/watch-number 829-580-0749. Append -force to the command word(s) if you really want to add the pattern you provided.
@JeffSchaller In order for a "number" to make an exact match, the pattern must end with a digit. Append -force to the command word(s) if you really want to add the pattern you provided.
@Makyen That pattern looks like it's already caught by Potentially bad keyword in body and Potentially bad keyword in answer Append -force to the command word(s) if you really want to add the pattern you provided.
sdc report stackoverflow.com/a/72659468 "Undisclosed affiliation with the linked repo (usernames match exactly), post is VLQ (link-only). Custom flagged."
