Hal DeVore <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] wrote:
>> I don't have anything against Content-Length, I just don't see why
>> anything but inc should even care. What am I missing?
>
>I'm with you, Andy!
>
>I don't want any tool messing with the content of the mail. It's a bad
>idea. It breaks signatures (I'm talking PGP, GNUPG, that sort). And
>it's ONLY necessary if you're foolish enough to allow your mail to be
>stored into a mailbox file. Us MH/nmh users don't have to put up with the
>stupidity of that file format.
>
>Use procmail to feed the mail to rcvstore and you'll never have the problem
>of a body being mistaken for a header line to begin with.
>
>Content-Length is evil. "From munging" is only slightly less evil.
Hmmm...I guess I'm a bit behind on state-of-the-art mail delivery
techniques. I'll have to go play with procmail on my Linux system at
home before messing with the local mail server here at work, though.
Improved mail delivery techniques aside, if one WERE foolish enough to
use sendmail with mbox-formatted files, isn't reading the Content-Length
field (as it is set by sendmail at delivery time) the proper technique
to use to determine the length of each message?
Note that my patch doesn't CHANGE the value of content-length, it simply
uses it to determine whether the '\n\nFrom ' pattern found is truly at
the end of the message. Of course, if you can't assume that the
Content-Length field is correct (either a bug in sendmail or a security
hole in permissions on /var/mail/ or its mail files), then you'll
incorporate two messages as a single message in the worst case. The
worst case currently is that the misidentified 'From ' line is
completely trashed and you end up with a split message.
BTW, why is Content-Length so evil in the context of sendmail and mbox
mail files? Or is using sendmail and mbox mail files the evil part?
Dave
