RE: [Evolution] options on download

[Gregory Leblanc <GLeblanc cu-portland edu>]

> -----Original Message-----
> From: Jeffrey Stedfast [mailto:fejj helixcode com]
> Sent: Thursday, June 01, 2000 7:42 PM
> To: Sejal Patel; Tom Musgrove
> Cc: evolution helixcode com
> Subject: Re: [Evolution] options on download
>
> I don't think this should be implemented, it's just silly
> I personally prefer it to download oldest to newest (the 
> order in which
> they are listed by the pop server)
> It just makes more sense, I mean... if we made it download 
> this way, then
> someone else might be left unhappy
> it makes more sense to me to download older messages first
> if you do it the way you suggest, then there may be old 
> messages that you
> may not get to retrieve for days/weeks/months/whatever if you 
> don't always
> let the download finish (which you don't need to do)
>
> you see why that's a bad idea now? it's more important to get the old
> messages first than it is to delay downloading the big ones, at least
> that's the case for 99.999999% of the users
>
> it would also be "un-necessary bloat" in the words of Matt Loper, our
> fearless Evolution leader
>
> but if someone feels it's necessary, here's the simplest way to do it:
>
> send a LIST command to the pop server, record the message 
> sizes for each
> message, sort on size (smallest to largest) and then download in the
> resulting order

ok, the simplest way, in this case, is not the right way.  The message
headers and bodies should be downloaded first, regardless of anything else.
Any binary attachments should be downloaded after all of the header and
plain-text mime portions are downloaded.  That way, I can read ALL of my
mail, from oldest to newest, and not have the silly scheduling problem that
you mentioned above.  

> no need to grab a header and no need to check for mime attachments

Without doing those, you don't know how to download things, and get into
scheduling messes.  

> if anyone feels up to patching 
> /evolution/camel/providers/pop3, feel free
> but I don't think anyone here has any intent on implementing 
> that feature

No, not going to volunteer personally, nobody likes Pascal.  8^)
        Greg