Re: mutt 1.5.9i: header_cache: Sometimes it does, sometimes it doesn't...
Hi Thomas,
* Thomas Glanzmann <sithglan@xxxxxxxxxxxxxxxxxxxx>:
> > set header_cache="~/.tmp/mutt-header-cache"
>
> Don't forget to
>
> set maildir_header_cache_verify=no
>
> as well. The default of this option is very conservative. And more or
> less only there because one Debian user was stupid[0] enough to violate
> the maildir 'protocol' by running sed on his maildir files to strip
> headers.
oh, I thought it would be Ralf Hildebrandt, but it was someone else. Ralf is
going
to kill me for that ;)
> > This all went very well and at the time I enabled this parameter accessing
> > my
> > folders sped up a lot. But then one day it was back to slow again and I
> > noticed the mutt-header-cache db had been gone without me doing anything.
>
> Please note that header caching is currently only supported for maildir
> and imap style folders. - But I think you know that already.
Yep. This I knew.
> > Just until yesterday no header cache db would be created. I tried to point
> > $header_cache to a dir and to a file, but that wouldn't change anything
> > either.
>
> If header_cache points to a directory. This directory has to be created
> before starting mutt. Otherwise it will silentely fail to use header
> cache. It's time to write a patch to warn the user in such a case.
yep.
> If header_cache points to a file, the directory in which this file will
> resist has to be there, too. There is also no warning if it fails.
yep.
> > Today I noticed that there were two header chache files both dating
> > from yesterday. I've been receiving, reading and sending mail all day
> > long today, so if mutt would use the db I guess the date of the dbs
> > should have todays date.
>
> This should be the case.
>
> (faui03) [/tmp/sithglan-hcache] ls -lart | tail
> -rw------- 1 sithglan icipguru 3680293 Apr 18 16:51
> 76c37b683752469fce2207184cb85249
> -rw------- 1 sithglan icipguru 102095 Apr 18 20:24
> 04ff968b416c48b00ab663e82600c3ac
> -rw------- 1 sithglan icipguru 6751114 Apr 18 20:24
> 0909ee7e8dcd1251aa558bb4a503c62e
> -rw------- 1 sithglan icipguru 814204 Apr 18 20:33
> 63cc6f51b93c5cb3e589755fb9a83056
> -rw------- 1 sithglan icipguru 4130965 Apr 18 20:40
> 0d28aec905aa05f98383360dd269d200
> -rw------- 1 sithglan icipguru 8488592 Apr 18 22:02
> c85bbc6377abf1c6aea7c4d8031e0297
> -rw------- 1 sithglan icipguru 633295 Apr 18 22:14
> a2fc73ba0ea769b4326f23d2925dbbb1
> -rw------- 1 sithglan icipguru 22021450 Apr 18 22:54
> b2932362d26987e33bd5b9ec9653758b
> -rw------- 1 sithglan icipguru 3868220 Apr 18 22:54
> 3d8953a1fd270ecf57bffc785310ec74
> drwxrwxrwt 20 root sys 2555 Apr 18 22:56 ../
>
> Also note. That there must be one file for every maildir you opened in
> mutt. No matter if it contains messages or not.
OK.
> > Am I doing something wrong? Any ideas?
>
> Mutt 1.5.9 has support for IMAP header caching in, but misses the
> support for MAILDIR header caching. From the mutt -v output you
> supplied, I guess you didn't apply the patch before compiling mutt.
This is where we need to talk about my stupidity...
> So you have a mutt without maildir header caching but with header
> caching infrastructure. You can verify my quick guess by putting in the
> "set maildir_header_cache_verify=no" option and mutt should bark at you
> because it does only know this option if it supports maildir header
> caching.
Verified! It barks.
> If my assumtion is correct there are the following ways to build a mutt
> which includes maildir header cache:
>
> - Wait till TLR finds the time to release mutt-1.5.10
> - USE CVS HEAD (maildir header cache is available in CVS HEAD)
> - Or just download the patch[1], apply it and recompile.
>
> For general informations and tips about the header cache please
> consult[2].
Thanks for the quick and detailed reply. I'll do some homework now and let you
know if I succeeded.
p@rick
> [0] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=199052
> [1]
> http://wwwcip.informatik.uni-erlangen.de/~sithglan/mutt/mutt-cvs-header-cache.29
> [2] http://wwwcip.informatik.uni-erlangen.de/~sithglan/mutt/
--
The Book of Postfix
<http://www.postfix-book.com>
SMTP AUTH debug utility:
<http://postfix.state-of-mind.de/patrick.koetter/saslfinger/>