[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Nmh-workers] IMAP, again

From: Lyndon Nerenberg
Subject: Re: [Nmh-workers] IMAP, again
Date: Sat, 26 Oct 2013 20:05:22 -0700

On Oct 26, 2013, at 7:11 PM, Ken Hornstein <address@hidden> wrote:

>> Beyond that, email messages are generally large (compared to calendar
>> entries, anyway) collections of text. Changes to messages mostly involve
>> small deltas to metadata.
> Note that messages in IMAP are immutable; we don't have to worry about
> changes to a message.

And thanks for quoting that previous text.  I got a bit aggressive with the 
delete key and lost the message I was going to reply to :-P

My comment about the incremental metadata changes being a small part of the 
equation was something I did not explain clearly.  What I was getting at – and 
should have spelled out to being with – is that the major operation in the 
context of IMAP is the operation of moving a message from one folder to 
another.  This is a full-on COPY+DELETE operation; you are instantiating an 
entirely new message, then destroying the original.  Under the hood, this is 
considerably more than just fiddling with the meta-data.  There is no 
equivalent to UNIX's link(2) that MH can use to shortcut the process.  I.e., 
you cannot reuse things like inode numbers, stat(2) data, or most anything 
else, other than the immutability property of the message itself.

And because IMAP messages are immutable, anno(1) fails miserably in the 
presence of an IMAP message store.  Yes, there are some IMAP extensions that 
can help here, but any solution we implement would have to work with a server 
that implements only the base 3501 specification.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]