Weird mutt problem.

[Thom May]

* Matt Zimmerman (mdz at canonical.com) wrote :
> On Sun, Oct 10, 2004 at 05:40:49PM +1000, Luke Yelavich wrote:
> 
> > I seem to be having a slight problem with mutt, and mailboxes that have
> > already been checked, and alterations made. For example, I receive mail
> > in my inbox, which gets read, and deleted if necessary. I then use the c
> > command to change folders to the next mailbox with new mail. I continue
> > through a varying number of mailboxes, and for some reason find myself
> > back in my inbox again, even though changes have been made to the file.
> > If I press c and enter again, I find myself in the last modified
> > mailbox, even though it also has had modifications made to the file. To
> > stop this, I have to manually touch my mailboxes to update modification
> > info.
> > 
> > This has happened only since running my laptop on battery. I am running 
> 
> When running on battery, your filesystems are remounted with the "noatime"
> option.  This is because, without this option, every file data access (even
> a read) triggers a write (and thus eventually causes the disk to spin up) in
> order to update the access time of the file.
> 
> I stumbled upon the effect in mutt as well, and I was confused for a minute,
> even though I knew this was happening. :-)
> 
> I have a feeling that perhaps this behaviour is not quite as useful when
> we're already using laptop-mode in the kernel when on battery.  Thom?

I tend to agree - laptop-mode on its own should be enough to reduce the
frequency of spin ups without needing to not write out atimes.
To disable this feature, put:
DO_REMOUNT_NOATIME=0
as a new line in /etc/default/laptop-mode and confirm your disks are not
spun up every time you touch a file :-)
I'll set this as the default option if no-one reports big problems.
-Thom