On Thu, Feb 07, 2013 at 10:47:57AM +0100, Andre Klärner <kan...@ak-online.be>
put forth the proposition:
On Thu, Feb 07, 2013 at 08:44:34AM +0000, David Woodfall wrote:
On Thu, Feb 07, 2013 at 08:01:21AM +0000, James Griffin <jmz.grif...@kode5.net>
put forth the proposition:
>--> David Woodfall <d...@dawoodfall.net> [2013-02-06 21:59:48 +0000]:
No it works fine now. Just trying out a couple of shell scripts
similar to Andre's, but much simpler:
mailboxes.sh:
##########################
#!/bin/sh
set -e
MAILDIR="/home/david/mail"
HOSTNAME="blackswan"
for f in $(find $MAILDIR -name "*," -o -name "*$HOSTNAME" | cut -d'/' -f5 | \
sort -u | sed 's/^\.//')
do
if [ "$f" == "new" ] || [ "$f" == "cur" ]; then
f="INBOX"
fi
echo -n " +"$f
done
echo ########################
It seems to be working fine but for some reason I'm unable to bind it to a
macro in browser:
macro browser z "|~/.mutt/mailboxes.sh\n"
This just tells me that key is unbound, and I've tried it elsewhere like
index, which works, but I'd rather have it in browser where I'm actually
looking at the mailboxes.
Well, not quite obvious at first, but easy to understand: Your script
generates just a list of mailbox names, e.g. "+foo +bar +baz". That output
is usable by feeding it into a mailboxes-command. e.g. "mailboxes
`mailboxes.sh`".
And that command in turn you can bind to a macro.: "macro browser z "mailboxes
`mailboxes.sh`"
But: that macro will be generated while parsing the config, so it will execute
"mailboxes +foo +bar +baz" when it is called. So you need to source that
out into a seperate file, that than in turn is sourced by the macron on
demand. So you essentially end up with a state that I already had: commit
17cb8a4ff0a455073fba19fdb95871b73ddf480e in my mutt-repo. At that time I
was still using the 1st version of my mailboxes-script, but it got
unmanageable as some of my folders grew too big. So the runtime of the
script ended up being 30-60s, depending on the speed of my NFS-server
serving the Maildir.
So I reimplemented it using perl (because I wanted to do perl anyway some
day) and build it to be completely optimized. The issue I diagnosed with
the old version was that it ran through all directories and file to find
unread mails, but that takes too long. In the perl-version I stop at the
first occurance of an unread mail or don't even check the cur-directory if
already the new-directory has contents. And I also avoid going into cur,
new or tmp directories if it should print all anyway.
Yes, my first effort was much too slow, but piping from find takes < a
second to output the list.
I've also taken out the subscribed part and just subscribe to all when
mutt loads.
Thanks for the info on macros. I didn't know they were generated like
that.
