Justin Gombos wrote:
I know that my version of sed
supports case insenstive matching ...
But when I try to use the 'I' flag on the following experimental
version of the script above (it doesn't):
:0 fbw
* ^X-BeenThere:.*procmail
| sed -e '/^[ \t]*$/N \
/^[ \t]*\n__*/N \
/^[ \t]*\n__*[ \t]*\nProcMail.*org\//,/http.*ProcMail$/Id'
it fails to delete the address space, and the footer remains in the
message. Can anyone see what my problem is?
I don't know how well the '\
' syntax in a procmail action line becomes a hard newline to sed, for
one thing; "\\
" does pass backslash-newline to sed, I've found. You might try it
without the backslashes, since procmail will honor the apostrophes and
take it as one long action line in my experience. Personally, I prefer
evading it by using multiple -e options when I call sed inside a
procmail rcfile.
But that would create a major sed syntax error in your procmail log, and
you didn't say you have one.
There's also the possibility that the sed you invoke during an
interactive shell login is not the one procmail finds. Your shell login
and your procmail invocations could run on different machines or have
different paths. But then, too, if procmail is finding a sed that
doesn't grok I, you'd get an error in your log.
So, OK, if you turn on verbose logging, what does the error message say?
Somehow or other, sed is reporting a bad exit and, because of the w
flag, procmail is recovering the unfiltered text.
____________________________________________________________
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