@JeffSchaller That pattern looks like it's already caught by Blacklisted website in answer and Blacklisted website in body Append -force to the command word(s) if you really want to add the pattern you provided.
@JeffSchaller 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.
I just flagged/reported a spam post but, on MetaSmoke, I see that I already had reported the post ("from review") as fp. I can't remember doing this, though. Is there a way I can see when I gave that FP?
Yeah - I can find no trace (now) of the earlier FP report I made. I'm concerned that I didn't earlier spot it for spam, though, especially if had been already reported/signalled by Smokey.
I came back to it because Suraj raised the issue in the AI chatroom, where I was asking about an answer it had. Once the "spam" issue was activated in my head, I could see it for what it was.
The really bizarre thing is that the post was "Manually reported by Adrian Mole" and "would not have been caught otherwise". [ MS ] So, how did I get to give FP feedback before it was even registered with M/S???
I guess Advanced Flagging could submit some feedback if you click NAA or etc but the timing would have to be really messed up to provide feedback on a post you reported later
@metasmoke Now I just gave FP feedback without knowing how I did that? My TP has been overridden. Wozappnin?
OK - that one came from (I think) me clicking on the red cross in MS. I thought I would get to see if that was my initial report. But I guess I misunderstood ...
[ SmokeDetector | MS ] Bad keyword in answer, blacklisted website in answer, potentially bad ip for hostname in answer, potentially bad ns for domain in answer, potentially bad keyword in answer (185): Online Job Portal System Use Case Diagrams by Emma Robert on stackoverflow.com
I suggested Movavi Screen Recorder as an answer to this question. My answer was deleted as spam, although I use the software myself and didn't notice any viruses or other adware there. According to the moderator, Movavi marked my reply as spam. Does it mean that all Movavi software is considered ...
^ On posts like that (I flagged it as NAA but didn't send any feedback to SD), should we be calling them as TP? It's sort of abusing the system but basically not really rude in Smokey's sense.
... OK, it was mod-deleted but not as R/A.
But I'm not completely happy sending FP feedback, because it's nice that such posts are detected.
I think Cody tends to more on the "filler characters are an abuse of the system limitations" side
I can definitely see the "we would want this blocked from the SE system-side" argument, but that might be a baby step above "block spam" towards "also block NAA's"
@Spevacus 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.
@CodyGray The checkmarks represent the TP feedback that existed at the time I copied it. I probably should have removed them, but chose not to, as many people use AIM, which would have provided that information if it was an actual SD report. There aren't any ✓, ✗,or 💩 in the others, because there wasn't any feedback on those at the time. If I'd created a script for it, I probably should have made them the same format as normal SD messages and then sent them from the SmokeDetector account.
@CodyGray My main problem from a conceptual POV of testing for "seo" in usernames is that there are quite a number of legitimate users that have "seo" as a word in their username. "Seo" is a real name, so is expected to be seen in usernames. I substantially dislike having detections which will detect users just because they used their real name or semi-common alternate. I expect that having such a test for "seo", even just in usernames, will result in quite a few FP. It would probably be substantially less FP if "SEO" is used (i.e. case sensitive for all caps), but it's unclear to me if that would pick up the usernames that you're wanting to catch. (cc @tripleee) [Note: my removal of the prior watch was because it wasn't doing what it was intended to do, rather than enforcing my concern about not detecting the real names of uninvolved users.]
@tripleee Yeah, that's an extension of the users.yml issue we've had for a while (since I switched the CI testing to be multithreaded). It used to be only a single user missing. I wrote code that added a bunch of locks which should fix that, and some other issues/potential issues (some are CI only issues, while others are real). My concern is that I may have gone a bit overboard on locks and I need to go back though those changes and see which ones are really not needed. I also need to do at least another fairly long run of a fully active SD instance with those changes to see if I munged up any of the locks. I also stalled out on it because it looks like there's a multithreading issue in the regex package, but I didn't actually find it. I worked around it, but I still need to confirm that what I did really resolved "the" multithreading issue there, or if issue isn't really in the regex package and the code I added just moved other processing around enough such that the issue went away. In other words, I've worked on this issue/these issues, but it needs more work.
> Blacklisted website in answer, potentially bad ns for domain in answer ---------- Blacklisted website in answer - Position 13-35: mytravel-insurance.com Potentially bad ns for domain in answer - mytravel-insurance.com NS suspicious ns14.wixdns.net.,ns15.wixdns.net.
@ProcessedMeat User almost certainly refers to metasmoke.erwaysoftware.com/post/407296. The product/website they linked to is blacklisted, so there is some element of "recommendations for this product are always considered spam", and, thus, this may not be a purely innocent question.
@Makyen Yeah, I got that part. To me, that was proof that you copied the messages from your own view of them, rather than getting a script to generate them, because the script would surely not have the feedback displayed inline like AIM does. Which is exactly what you concluded in the last sentence. :-)
@Makyen Hmm, I suppose this is the sticking point. I was not aware that "Seo" is a real name or that it would be expected to be seen in usernames, beyond the rare exception that someone has a promotional username but contributes acceptable content (which does happen, but, to me, is an exceedingly rare edge case and probably a reasonable situation to have FPs).
I think watching for "SEO" in usernames in all caps would be better than nothing, but I'm not sure that it would catch everything. I have not paid enough attention to the things that SD misses (which are the things that I would mostly hope to catch by such an expression) to remember if they are always or generally uppercase.
At any rate, thanks for clarifying why you removed the watch; my impression, based on this message is/was that tripleee still did not fully understand why the watch that was tried was excessively FP-prone.
@CodyGray 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.
