On Thu, 26 Oct 2017 20:32:44 -0400 Ken Hornstein wrote:
Ken Hornstein writes:
> Since you say that it "often"
> takes much less than a second, I am wondering if maybe that only happens
> when new mail arrives and Dovecot needs to rescan the directory.
I have no idea. I need to
Ken Hornstein wrote:
the top of thread shows a prediction of likely performance impact from
opening and closing an imap session every time an mh command starts or
stops. i'm arguing against that, because state is necessary in order to
receive push notifications, and so to me, the reason to
On Fri, 27 Oct 2017 01:33:37 +0200 Paul Vixie wrote:
Paul Vixie writes:
>
> i was thinking of this: https://en.wikipedia.org/wiki/Push-IMAP
>
> by which means my GUI imap client gets new mail notifications pushed to
> it, so that it does not have to poll.
Looks like
Hi Ken,
> > min ms avg max mdev
> >128.2.158.29 120.570 123.153 128.260 2.133
> >128.2.105.45 120.830 136.279 249.528 37.794
> >128.2.158.26 116.089 133.622 180.430 20.598
>
> For 128.2.158.29, I get: 22.161/22.480/22.843/0.216 ms
Repeating
>start=1 tls=
>[...]
> Command execution 0.154234 0.158430
> Command execution 9.703771 9.689929
>[...]
> Total elapsed10.504798 10.478854
Interesting! So, this tells me that the RTT, at leasst for the smaller
requests, is not so bad in
Hi Ken,
> I'd be interested why this variance occurs. You would think for the
> same mailbox it would always be the same.
The tests with that public IMAP server are presumably accessing content
that it has readily to hand, e.g. in RAM. I suspect a lot of commercial
IMAP servers, e.g. what
On Fri, 27 Oct 2017 12:21:03 -0400, Ken Hornstein said:
> A better question is ... do you consider a "scan 1-10426" taking 10 seconds
> reasonable?
Well, on my laptop (Dell Latitude, 16G RAM, unfortunately still spinning
oxide), scanning a local folder. 2 runs, first memory-cache cold second
>connect to the imap service and authenticate
>open a unix domain socket and listen() to it
>on each connection:
So, alright, I admit that you're more experienced than I especially at
the "big picture", so I want to understand exactly what you're proposing.
Is this, essentially, a local IMAP
>So some of us are seeing 10+ second scans locally.
>
>I really need to upgrade to a 500G SSD. :) (which I guess answers the
>"is it reasonable/tolerable" question :)
I feel it's only fair to point out that 10 seconds was across the
Atlantic Ocean. To the same server a few states away, that
>The tests with that public IMAP server are presumably accessing content
>that it has readily to hand, e.g. in RAM. I suspect a lot of commercial
>IMAP servers, e.g. what comes thrown in with one's broadband connection,
>are over-subscribed so there's a bit of competition for whose Maildir
Hi Valdis,
> 0 [~] time bash -c '(scan +linux-kernel 1-10426 > /dev/null)'
>
> real 0m17.625s
> user 0m1.610s
> sys 0m3.358s
What does `scan -version' give?
Is 1-10426 actually 10,426 emails? :-)
If you have grep(1) read the files up to the end of the headers instead,
does it still take as
Hi,
I'm tinkering with SPECS/nmh.spec and SPECS/nmh.cygport, in particular
removing the statement a GUI is required if nmh is to be used as one's
main MUA. :-)
nmh.cygport's description suggests the xmh package, and that does exist
in Cygwin, but I don't think exmh is packaged? And nmh.spec,
Ken Hornstein wrote:
>> connect to the imap service and authenticate
>> open a unix domain socket and listen() to it
>> on each connection:
> So, alright, I admit that you're more experienced than I especially at
> the "big picture", so I want to understand
On 28 October 2017 at 1:10, Ralph Corderoy wrote:
> I'm tinkering with SPECS/nmh.spec and SPECS/nmh.cygport, in particular
> removing the statement a GUI is required if nmh is to be used as one's
> main MUA. :-)
>
[snip]
>
> So of the four {Cygwin,Fedora}/{xmh,exmh}, is
>at the moment i want to be able to get the "you have new mail" behaviour
>in my shell (see "mesg y") that delivermail used to offer me (after
>writing into /var/mail/vixie), so i know when to type "inc". but i could
>also imagine automating that so that it would fetch synchronously
>without
Ken Hornstein wrote:
connect to the imap service and authenticate
open a unix domain socket and listen() to it
on each connection:
So, alright, I admit that you're more experienced than I especially at
the "big picture", so I want to understand exactly what you're proposing.
Is this,
Hi Ken,
> But maybe you're bandwith-limited
Yes, in this large case. I can observe `ip -s a show dev ens35' before,
during, and after, and see it causes RX'd bytes to shift from idle to
full pelt.
The frequency of the timestamps of the network packets show they're
fairly even.
3
On Fri, 27 Oct 2017 22:29:32 +0100, Ralph Corderoy said:
> What does `scan -version' give?
> Is 1-10426 actually 10,426 emails? :-)
Linux-kernel is a firehose. 800 to 1200 emails a day (depending on whether or
not Greg KH drops a -stable review patchbomb series, those can be 300 mails
right
18 matches
Mail list logo