Re: [mutt] generating mailboxes command (was: split display?)

[lee]

On Tue, Jul 21, 2009 at 02:52:57PM +0100, Adam Wellings wrote:

 If the set of maildirs rarely changes, would it be faster to
 generate a file and just source that in your .muttrc?

It depends on what information from the file system is already
cached. Assuming that nothing is cached, I would think that a file is
faster.

 I can't do that as I use procmail to 
 sometimes create new maildirs in some situations.
 
 Also, I've looked up why my script is no longer used, it took ~8-9s to 
 produce the list. I'm sure it could be improved upon generally, but as find 
 does the job, it'd purely be an academic exercise.

Then mutt-mb is for you, it's pretty fast :)

Re: [mutt] generating mailboxes command (was: split display?)

[Adam Wellings]

On Tue, 21 Jul 2009, Christian Ebert wrote:

 * Adam Wellings on Tuesday, July 21, 2009 at 11:17:09 +0100
  I have a folder hierarchy of maildirs, though only the nodes (or leaves) 
  are 
  actually maildirs, eg:
  
  fol2
 | -Fol2
 || -maildira
 || -maildirb  
 | -Fol3
 || -Fol4
 ||| -maildirc
 ||| -maildird
 || -maildire
 | -Fol5
 || -etc..
  
  This mailboxes command works for me:
  
  mailboxes `find /path/to/mail -type d -name cur printf '%h '`


Firstly sorry, I missed the '-' off printf
 
 Sure, but (my find doesn't have printf):
 
 ~$ time find ~/Mail -type d -name cur -execdir pwd \;  /dev/null
 
 real0m54.973s
 user0m0.447s
 sys 0m54.159s
 ~$ time find ~/Mail -type -d \( \( -name cur -o -name new -o -name tmp \) 
 -prune -o -print \)  /dev/null
 find: -type: -d: unknown type
 
 real0m0.401s
 user0m0.001s
 sys 0m0.011s

Thanks for that, I've created this command based on the above:

mailboxes `find ~/Mail -type d \( \( -name cur -o -name new -o -name tmp \) 
-prune -o -printf '+%P ' \)`

It's cut the time from ~1s to ~0.3s.

If the set of maildirs rarely changes, would it be faster to generate a file 
and just source that in your .muttrc? I can't do that as I use procmail to 
sometimes create new maildirs in some situations.

Also, I've looked up why my script is no longer used, it took ~8-9s to 
produce the list. I'm sure it could be improved upon generally, but as find 
does the job, it'd purely be an academic exercise.

Cheers,
Adam