In an older episode (Friday 15 April 2005 02:02), [EMAIL PROTECTED] wrote:
> \b means "word-ish character on one side, non-wordish character on the other" good explanation to help me understand my misconception, thanks. > | is not a word-ish character i was aware of that, but tried to "consider | to be wordish", exactly (which was even unnecessary in case of mai\|). > A better way to match the "end of a word, considering | to be word-ish" might be > \bmai\|(?:\W|$) > > and to match beginning of a word > (?:\W|^)\|etter\b thanks! In an older episode (Thursday 14 April 2005 02:31), Robert Menschel wrote: > Use the command > > ./spamassassin -D -t <testfile >outfile 2>msgfile > and peruse the msgfile output ... you will at least have a full list > of all the subtests that fired and thanks for that one, too. that subtests list was very useful when trying to debug why arithmetic metarules didn't do what i expected, and the subtests were the wrongly used \|\b\| ones. regards, wolfgang