Re: Patches
Hi,
* Thomas Roessler [06-04-28 21:58:48 +0200] wrote:
On 2006-04-28 21:45:45 +0200, Thomas Roessler wrote:
Oh well. Now trying to hack this into some shape that
remotely resembles decent code.
Try the change below the .signature, it's in CVS now. Still,
playing with this feature a bit, it's a terrible performance
hog.
That's why I dirty-fixed the symptoms a while a back and reported it.
But it can be turned off completely.
If you use this, it quite simply requires mutt to parse every
individual message when you scroll through an index for the
first time; with a large-ish mailbox, it's quite painful to
use, even on modern machines with fast hard disks.
But people get what they want. I first noticed the impact on IMAP when
my mailbox contained a >10 MB message. Scrolling then isn't fun.
It's
probably ok with a long-running mutt instance that operates,
say, on an inbox where messages slowly come in and the parsing
can be done one by one.
Only when you stay within the same folder.
I'm wondering, is there anybody here who *seriously* uses this
feature?
I use it. But I could live without it, I guess.
If the main use case is to simply know whether the top-level
body part is a multipart, then we can get that a *lot* cheaper,
and with a fraction of the code.
Yeah, but's not what the patch is about. It also adds additional
configuration commands so that the user can define what MIME part mutt
is supposed to count as 1 attachment.
Persoally, I don't think that
this is particularly useful information (and that's why I threw
the feature out years ago -- it's particularly nonsensical for
multipart/{signed,encrypted}),
I didn't quite understand the configuration. But with the default
configuration given as an example, only non-text/plain parts are counted
excluding mulitpart/signed. I'm not sure about encrypted.
The encryption should only be triggered by entering the pager (i.e. a
body handler) which isn't used in this case anyway.
I removed my fix as it's not necessary any longer.
bye, Rocco
--
:wq!