@metasmoke Oh wow, spam for a product I'm familiar with! I've actually played their Enter the Spudnet game. I would not recommend it. It was not very good.
If I recall correctly, the owner threw the game out after we played it.
In case anyone was wondering about the quality of products that get spammed on Stack Exchange :D
Yeah, one that was asked about in the question, that appears to be legitimate, and with no history of excessive promotion. I don't see why it's worth slapping with a tp. See also my comment on the question. Also got no clue how I fat-fingered so hard that I clicked a different link instead of the one in the message xd That's definitely a new one
[ SmokeDetector | MS ] Few unique characters in body, luncheon meat detected, no whitespace in body, repeating characters in body, repeating characters in title (364): xxxxxxxxxxxxxxxxx ✏️ by bluesdrip on stackoverflow.com
@tripleee 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 if you really want to add this pattern.
@CodyGray I had them before, I just temporarily turned them in because I wasn't comfortable doing it without oversight right away after 2 years of break.
@CodyGray The regex contains an unescaped ".", which should be "\." in most cases. Append -force to the command if you really want to add this pattern.
Frankly, the only reason I see for splitting them up is to make editing the file more manageable. I'm not really sure why the blacklists were split up. I assume there's a non-obvious technical reason for that?
Being able to watch usernames would probably be nice. Admittedly, we have a relatively functional hack for that, due to length limits. But it's not perfect.
phone numbers, IP addresses etc are treated differently in the code but user name regexes, domain name regexes, and keyword regexes are all just applied to all the text (albeit sometimes with slightly different anchoring etc)
It trips 3 reasons that have 87%, 85% and 99% accuracy, so together it passes the threshold for autoflagging. The autoflagging threshold was calculated based on seperate reasons for websites/keywords/usernames etc so it would have to be re-done if these were merged.
@Magisch (sep-a-rate is hard for Americans because they don't distinguish unstressed vowels much; but for someone who speaks a less guttural language, this typo should be impossible)
I have a pending PR to refactor the blacklist logic, I really want to get that in first before proceeding with any of this; but I concur with the reasoning
[ SmokeDetector | MS ] Few unique characters in body, mostly punctuation marks in body, repeating characters in title (188): aaaaaaaaaaaaaaa ✏️ by hinanily on ethereum.SE
@Magisch 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 if you really want to add this pattern.
@Magisch those sometimes catch me when the reason has a comma-separated list -- Potentially bad keyword in body - Positions 398-410, 1502-1514: cbd-coin.com, Position 2133-2140: gummies -- "it's in there"
