On 2009-11-10, James <j...@nc.rr.com> wrote: > Any good regex_reply configurations out there?
Here's what I use after several years of tuning. ######################################################################## # # Note: All letters in the reply_regexp must be lower-case, else the # entire pattern will be treated as case-sensitive. # set reply_regexp="^((\ \\[protocol_tc\\]|\ \\[fw_rtp\\]|\ \\[wvware - help\\]|\ \\[[a-z][a-z0-9 :-]+[0-9]\\]|\ out of office autoreply:|\ (\\(fwd\\))|\ (re(\\[[0-9]\\])?|aw|fw|fwd|\\?\\?|ç.å¤.|´ð¸´|\\?\\?\\?\\?):\ )[ \t]*)+" ######################################################################## I subscribe to several mailing lists, mostly company-internal (I deleted some names and altered others in the above), that automatically insert the name of the list in brackets at the start of the Subject. I could use a regular expression for the names, but I prefer enumerating them to avoid unintended matches. The structure I use makes that really easy to do. Including those names in 'reply_regexp' prevents them from building up in successive replies. The pattern immediately above "out of office autoreply:" was written primarily to handle the w3m-dev-en list, whose Subjects begin as this example, "[w3m-dev-en 01113]", where the number is the message number. I added "out of office autoreply:" so that those messages would be threaded with other replies to the original message. The question marks and the two sequences of non-ASCII characters are there to handle replies from coworkers in China who use some sort of Chinese Outlook. Regards, Gary
