@RyanM That pattern is going to be relatively expensive. While I can understand the desire to include the three words prior to "trading platform" in the why data, unless you have a specific reason, it's not worth the CPU cost.
FYI: I've locked the regex package version in requirements.txt to be <2020.10.22. Today's version prevents any explicitly compiled pattern from being placed in the compiled regex cache (i.e. only implicitly compiled patterns are cached). Using 2020.10.22 will have some performance impact, but what that impact will be is unclear. A quick grep through SD's code indicates we tend to explicitly compile patterns, even in cases where an implicit compile could be easily used.
For a bit more information, please see: Issue #387 in the regex repository on bitbucket.org. [Repository moved to GitHub; this issue on GitHub]
^ lot of spam on that domain still on the network if anyone wants a rabbit hole full of flags. Every post on MS is spam, and so are at least a few unreported ones.
!!/report stackoverflow.com/a/56481672 "Poster has the same display name (Bhupesh) as other accounts spamming this domain. This one was not posted the same day as the linked article"
!!/report stackoverflow.com/a/56778902 "Poster has the same display name (Bhupesh) as other accounts spamming this domain and has posted another link to this site. This one was not posted the same day as the linked article."
!!/report codereview.stackexchange.com/a/223010 "Poster has a typo'd version of the same display name (Bhupesh) as other accounts spamming this domain. This one was not posted the same day as the linked article."
!!/report stackoverflow.com/a/40918150stackoverflow.com/a/40895547 "Undisclosed affiliation: Poster has the same display name (Bhupesh) as other accounts spamming this domain and has posted 2 other links to this site, domain has a history of spam, posted same day as the article"
!!/report unix.stackexchange.com/a/360226 "Poster has the same display name (Bhupesh) as other accounts spamming this domain. This one was not posted the same day as the linked article"
@AndrewT. It does that if the post would be caught if it were scanned today. In this case, it's since acquired a potentially bad NS configuration, and I watched the domain after seeing the first three.
Or is the confusion over what exactly ?: does? It's actually unrelated to it being optional, it's just a performance optimization since we don't use backreferences.
Yeah, these are nonsense advertisements that I keep seeing on youtube too... Indian scammers aren't the smartest of people and they believe beautiful girls get clicks (which I am afraid is true) so without context, any spammy ad will have a cute girl in it
So the equation is: 1. put a list of products/services. 2. grab a cute Indian girl pic from online 3. mix them together and ask the user to click on the link that you want them to click on
the only exception here is there was no such link... which i will attribute to them not being that good at their game
Also, I wanted to ask a question, is smokey good at detecting spam in other languages? I found 2-3 answers that were in different languages and translating them revealed that they were spam... I reported them on SOCVR
@Yatin It's not as good, since we all speak English but only a scattering of other languages...so our ability to write complex regexes is somewhat impaired and there's more guesswork.
if you need access to Blazer that's probably more convenient than the rather crude query facilities in metasmoke itself, it's unfortunate that it's behind rather steep privileges but if it helps you help us I don't think that will be a problem
speaking of Blazer, "make a search diffing function" is on my vague to-do list...it's awkward, though, because Blazer doesn't seem to be capable of escaping input...like regex.
The only time I've used it to write regexes was a particularly ...vexing group of spammers that seemed to have quite a lot of phone numbers, and I needed to extract the phone numbers to find out what they had in common.
@Yatin but yes, feel free to ask questions. hopefully they won't be buried in too much spam. Charcoal's data suggest that spammers seem to take the weekend off, so you should be good there :-)
> Bad keyword in answer, potentially bad keyword in answer ---------- Bad keyword in answer - Position 9-13: keto Potentially bad keyword in answer - Position 0-13: one shot keto