Re: mutt cache sensitivity
* Paul Walker [07-03-29 11:23:47 +0100] wrote:
On Wed, Mar 28, 2007 at 05:32:47PM -0600, Kyle Wheeler wrote:
hurry, so I kill mutt (ctrl-\, because ctrl-c won't stop a download).
The cache will be corrupted, because mutt didn't get a chance to
unlink it out of the cache.
That's a good example. It also means mutt won't get a chance to unlink the
file in the temporary location, though. Are you planning to clean it on
startup, or just let the system tidying scripts deal with it?
I know it's not really nice, but I think we can never ever get all
possible situations right to get a valid cache and know that it's valid
without downloading the message again and verifying contents (or compare
checksums). But that breaks the cache idea. :)
For example, messages may get changed server-side (at least for POP
since IIRC there's no exact specification of how the checksum per
message is to be computed, or users may edit IMAP message manually on
the server) without mutt having a chance to take note of it.
Once we find a solution for aborted writes, I think the user is
responsible for maintaince of the cache as he's it now for header cache