On Tue, Aug 03, 2004 at 07:41:20PM +1200, Volker Kuhlmann wrote:
I have a recipe of which I don't understand why it matches:
:0
...
{
:0 B
...
{
...
}
:0 B
* ^Message-id:
{
The email in question is short and doesn't have any line in the body
which starts with "m", but the ^Message-id: is matched some of the
time, depending on who-knows-what. The test email is always the same.
There's a message-id: in the header, but flag B should only search the
body.
Is this a procmail bug? If there's a known bug list, it's not mentioned
on procmail.org.
procmail v3.22 2001/09/10
My bet is that you have induced a known different bug up earlier in
your rc. If you have a recipe with the "H" flag up above anywhere,
then the bug says that cannot be turned off later. So a subsequent
recipe with just "B" is really "HB".
The bug has been discussed many times, including within the last 10
days or so again, so searching for it should not be hard. Nancy McGough
mentions it on her QuickStart pages (linked from www.procmail.org) as
well.
I agree that it would be good if there were a canonical bug list.
In general you should not set the "H" flag at the top of a recipe.
It is, after all, the default behavior. If you need an H limitation
in a recipe, use the alternate syntax:
:0
* H ?? foo
action
--
dman
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail