https://issues.apache.org/SpamAssassin/show_bug.cgi?id=7127
--- Comment #34 from RW <[email protected]> --- (In reply to Reindl Harald from comment #33) > >> read the initial report again > >> > >> * i saw *repeatly* whole message parts in bayes_seen > >> * i saw that over months independent how often rebuild > > > > Neither of these statements is in the original report - I can't find > > anything > > that says this in the entire thread > > WHAT? > "bayes_seen contains whole message parts" is the subject of the bugreport There's no mention of "*repeatly*" and "over months" in that original report. There's nothing there to suggest you saw corruption in more than one file. > and https://issues.apache.org/SpamAssassin/show_bug.cgi?id=7127#c6 is also > pretty clear - so *what* exactly are you missing here? That just says that you "don't see that issue no longer". It doesn't imply that you reproduced the bayes_seen corruption and established repeatable behavior. If you have, you might have mentioned it before. > no idea what in "i don't see that issue no longer as well es no training > messages are ignored" is unclear "that issue" logically belong to the > subject of the burgreport "bayes_seen contains whole message parts" I've no idea what that was intended to mean, but my point is that in some places you seemed to be inferring corruption from the output of sa-learn, rather than from observing actual corruption. -- You are receiving this mail because: You are the assignee for the bug.
