[FFML] [ADMIN] Archiver issues, and minor thing to deal with this

[Mark A. Biggar]

Dennis Carr wrote:
> Some of you have been noticing the way the archiver here on Chez Vrolet
> has been handling inbound messages.  One of the most common ones I've
> seen thus far is what appears to be a spillover at 10.2 kilobytes of
> message, then continuing to a new message.
> 
> Digging around, I found this is a known issue - with the file format.
> What happens is that the archiver sees a line with the word "From" in
> the very front (no spaces, no nothing), and assumes it's a new message,
> so it's more how it's parsing the archive datasets.  It's actually due
> to the format of the archive file itself not coinciding with what is
> based in reality. =O.o=
> 
> I'm currently researching a way to fix this - it's the first time I've
> run into this issue, to be perfectly honest.  What I do know is that the
> data is intact, just parsing funny.  Hopefully I figure out a fix to
> this soon. =^^=

This is a standard problem that most mail and news processors have. It's 
actually working correctly according to the RFC-822 Internet email 
standard, which says that in a file containing multiple messages that 
each message starts with a line that starts with the string 'from' 
ignoring case.

Most email processing software (for example sendmail) has an option to 
filter incoming email and add a '>' (or some other non-alpha character) 
on the front of any line (except the very first line of a message) that 
starts with 'from'.  This is why most of our list subscribers don't see 
this problem with the messages they receive from the list as their local 
email processor is most likely set up to fix this problem.

Fixing that will take care of future messages, but the fixing the 
archive is harder. The last time I had to fix this problem in an email 
archive I had to hand edit it and fix each occurrence manually.

-- 
mark at biggar.org
mark.a.biggar at comcast.net