Re: [Evolution-hackers] Mailer behavior with slow connections.

[Not Zed <notzed ximian com>]

Ok i'll qualify this, although the explanations probably wont help.

1) certain operations need to run in order, otherwise the ui doesn't
work properly.  e.g. it could show that you're looking at folder "y"
even though it is actually looking at folder "x" because of out-of-order
task completion
2) folders have a single lock, so if you start an operation others can't
run until that opreation is complete.
3) who knows, probably it was busy on a slow link and there's nothing
that can be done about it.

1 & 2 are difficult to fix.  3 is probably not "fixable".


On Mon, 2005-05-16 at 09:49 +0530, Not Zed wrote:
> 
> It's mainly because of threading and locking and scheduling.
> 
> Its too complicated to explain in an email.  Look at the source-code
> (this is the hackers list afterall), or file a bug I suppose.
> 
> 
> On Sat, 2005-05-14 at 14:04 -0400, Nat Friedman wrote: 
> > Hi guys,
> > 
> > This week, for the first time in years, I've been using a modem to
> > access the Internet.  And I've noticed a few behaviors in Evolution's
> > mailer that were pretty annoying in the presence of a slow connection.
> > 
> > It seems that whenever certain mail operations are underway -- like
> > scanning IMAP folders for new messages, expunging messages from IMAP
> > folders, or syncing remote IMAP mail locally, to name three -- many
> > other mail operations block on the completion of the (seemingly
> > unrelated) ongoing network operation.
> > 
> > Examples of operations that blocked for me include:
> > 
> >         1) Switching from my IMAP folder to a local folder, so that I
> >         can read local mail while I wait for my IMAP folder to update.
> >         This blocks until the IMAP folder update is complete.
> >         
> >         2) Switching from an IMAP messages which is not yet downloaded
> >         and locally sync'd to once which I know for sure has been
> >         downloaded and locally sync'd, so I can reread it and perhaps
> >         reply to it.
> >         
> >         3) Closing Evolution.
> > 
> > There might have been others, but these were the ones that stuck out the
> > most.
> > 
> > Is there a design reason for this behavior, or is it just a bug?
> > 
> > Thanks,
> > Nat
> > 
> > 
> > _______________________________________________
> > evolution-hackers maillist  -  evolution-hackers lists ximian com
> > http://lists.ximian.com/mailman/listinfo/evolution-hackers
> _______________________________________________
> evolution-hackers maillist  -  evolution-hackers lists ximian com
> http://lists.ximian.com/mailman/listinfo/evolution-hackers