24 lines
1.4 KiB
Plaintext
24 lines
1.4 KiB
Plaintext
MBOX-Line: From tss at iki.fi Tue Jun 15 14:30:11 2010
|
|
To: imap-protocol@u.washington.edu
|
|
From: Timo Sirainen <tss@iki.fi>
|
|
Date: Fri Jun 8 12:34:44 2018
|
|
Subject: [Imap-protocol] MOVE is a pipeline
|
|
In-Reply-To: <4C17D434.5080807@aol.com>
|
|
References: <1372616189.4386.1276624386644.JavaMail.root@dogfood.zimbra.com> <1276626345.2916.50.camel@kurkku.sapo.corppt.com> <484642146.4458.1276626832297.JavaMail.root@dogfood.zimbra.com>
|
|
<4C17CBDA.6070604@gulbrandsen.priv.no> <4C17D434.5080807@aol.com>
|
|
Message-ID: <60C60CAD-FD75-4744-BDC0-4FB3F412CB80@iki.fi>
|
|
|
|
On 15.6.2010, at 20.27, John Snow wrote:
|
|
|
|
> How does everyone handle a large folder, say 500,000 messages? In order to process sequence numbers, you have to
|
|
> know which messages are in the set. The few implementations I've seen do this by loading the messages into a cache.
|
|
> This makes some command slow while the cache is loaded. There is also a limit on how many messages can be loaded
|
|
> in the cache. If UID were the only way to reference a message, then I wouldn't need to cache the message list, I
|
|
> could just work directly from the database.
|
|
|
|
Many clients just fetch all of the flags from a mailbox each time they select it. I don't think getting rid of MSNs would help your performance with them.
|
|
|
|
If a client accesses only the most recent messages, then you can just fetch their data (MSN-UID mapping, flags, ..?) from database. No need to waste memory/bandwidth/etc for messages that are never accessed.
|
|
|
|
|