A "Thank You"
David Relson
relson at osagesoftware.com
Tue Feb 25 14:52:53 CET 2003
Nick,
Yesterday I had observed a difference in trace output from the current
lexer trio and the trace output from my unified lexer. Tokens from message
separators were appearing in different places in the two
outputs. Typically they appeared sooner in the unified lexer and later
with the lexer trio. Also, when processing mailboxes, I found that some
tokens were being counted as part of a subsequent message.
Yesterday, I saw this behavior, but didn't understand it.
This morning, as I was reading and replying to your message, the light
dawned and I realized what was wrong. As described earlier, token.c is
changing state in the middle of the message separator which causes tokens
from the message separator to be temporarily forgotten.
Thank you for sparking that understanding.
David
--------------------------------------------------------
David Relson Osage Software Systems, Inc.
relson at osagesoftware.com Ann Arbor, MI 48103
www.osagesoftware.com tel: 734.821.8800
More information about the bogofilter-dev
mailing list