On 02/12/2015 04:14 PM, Zooko O'Whielacronx wrote: > Hm… I'm not sure, but I think what is happening is that spammers have > started submitting complete English texts that they've copied from > somewhere else, and so now the Bayesian classifier is starting to > think that complete English text is likely to be spam… > > Anyway, I've just manually retrained it that the things you submitted > last night that it thought were spam (like the one you quoted in this > email) were actually ham, so maybe now it knows enough to recognize > the difference? If it happens again, please let me know!
Okay Zooko, thank you. I used a non-ASCII character - maybe that is very bad because it could be a smiley or a look-alike character in a fishing URL. Lukas _______________________________________________ tahoe-dev mailing list tahoe-dev@tahoe-lafs.org https://tahoe-lafs.org/cgi-bin/mailman/listinfo/tahoe-dev