I was at a loss as to why my attempt to set up spam-stat seemed to
have no effect, only to find (digging in the code) that it was
ignoring most of the sample files due to this undocumented variable.
* doc/misc/gnus.texi (Creating a spam-stat dictionary): Document
the variable 'spam-stat-process-directory-age'. (bug#39780)
file is treated as one non-spam mail.
@end defun
+@defvar spam-stat-process-directory-age
+Maximum age of files to be processed, in days. Without this filter,
+re-training spam-stat with several thousand messages could take a long
+time. The default is 90, but you might want to set this to a bigger
+value during the initial training.
+@end defvar
+
Usually you would call @code{spam-stat-process-spam-directory} on a
directory such as @file{~/Mail/mail/spam} (this usually corresponds to
the group @samp{nnml:mail.spam}), and you would call