Interesting. The spam post itself wasn't detected by Smokey, but because it 'bumped' the question, the question and every answer got rescanned and Smokey triggered on one of the other answers.
But this could be a challenging pattern for us. Links to Amazon.com are always 'converted' by Stack Exchange so that they benefit from the referral, but that makes them harder to detect.
We have officially discontinued the Amazon book affiliate remnant ads, since despite our best efforts they don't perform well for our audience and cannot be made to perform well.
We've often wondered, Would it be a problem if all Amazon links were converted to affiliate links?
I'm thinking this...
@Glorfindel That pattern looks like it's already caught by Bad ns for domain in answer and Bad ns for domain in body; append -force if you really want to do that.
@Glorfindel That pattern looks like it's already caught by Bad ns for domain in answer and Bad ns for domain in body; append -force if you really want to do that.
[ SmokeDetector | MS ] Bad NS for domain in body, bad keyword in body, blacklisted website in body, link at end of body, pattern-matching website in body: This consolidate bouncecel be by Hati ere on drupal.SE
[ SmokeDetector | MS ] Bad NS for domain in body, bad keyword in body, blacklisted website in body, link at end of body, pattern-matching website in body: So you can relax confident by Polet1968 on askubuntu.com
Since the page under the link given there is in Chinese, I do not know whether it is just a completely misplaced post, or whether it is indeed spam: About Riemann Hypothesis (MathOverflow).
I'm voting to close this question as off-topic because this is spam. The link given does NOT purport to be a proof or the Riemann hypothesis. — Michael Renardy11 mins ago
@Videonauth You are not a privileged user. Please see the privileges wiki page for information on what privileges are and what is expected of privileged users.
Random-ish question: I assume that findspam.py contains the various spam reasons (URL in title, bad keyword, etc.), but each one appears to return true/false. How does reason weighting work?
@quartata I had a nasty little dream last night that you somehow changed all the file extensions to .py3 and had a lot of obscure names for some of the modules and things in Smokey...
you wouldn't be a person who would do such evils would you? :P
But why for prefix? People misspell thos all the time, and knowing that is better than getting radio silence and thinking maybe something else is going on
We wouldn't have known that I missed an alias on watch without it for instance
I recently reported What are the reasons for using parser combinators? for spam but it was declined. The problem is not immediately obvious for the post, but in fact, it is advertising the author's own library; the answerer simply wrote that he was familiar with the library, but if you were to lo...
if a Helios key isn't available or if Helios isn't working for an extended period of time we don't want to lose our current local persistence behavior. at the sme time, however, managing pickles is a nightmare
what I think we should have are some abstract containers (PersistentDict or some such) that save to either the appropriate Helios end point or a shelf otherwise which does a good job of abstracting the pickle part out
@quartata I'm gonna agree with this, but not for the same reasons
Helios being unavailable is unlikely. It's gonna run on AWS; reliability scores are stupid high.
But - we fairly regularly get people trying to run Smokey for themselves, without all the creds. Expecting those people to set up their own instance of Helios as well is unreasonable; so, the persistence layer works with pickles instead.