Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > alt.comp.software.thunderbird > #21128

Re: Slow message send

From Paul <nospam@needed.invalid>
Newsgroups alt.comp.software.thunderbird
Subject Re: Slow message send
Date 2026-06-26 09:56 -0400
Organization A noiseless patient Spider
Message-ID <111m0f6$it7f$1@dont-email.me> (permalink)
References (6 earlier) <111jb19$3q87m$1@dont-email.me> <111jdtv$3rbru$1@dont-email.me> <111jie3$3st08$1@dont-email.me> <111lup3$ibtj$1@dont-email.me> <111luqb$ic11$1@dont-email.me>

Show all headers | View raw

On Fri, 6/26/2026 9:28 AM, John C. wrote:
> On 26/06/26 06:28 AM, John C. wrote:
>> On 26/06/25 08:45 AM, Paul wrote:
>>> On Thu, 6/25/2026 10:28 AM, Paul wrote:
>>>> On 6/25/2026 9:38 AM, Paul wrote:
>>>>> On 6/23/2026 10:22 AM, John C. wrote:
>>>>>> Nobody wrote:
>>>>>>> Frank Miller wrote:
>>>>>>>> Andy Burns wrote:
>>>>>>>>> Frank Miller wrote:
>>>>>>>>>>
>>>>>>>>>> What the heck is "assembled" there?
>>>>>>>>>
>>>>>>>>> Bolting the headers, body and attachments together.
>>>>>>>>
>>>>>>>> I've never seen a message of that kind.
>>>>>>>
>>>>>>> It's been happening/appearing regularly for me for several versions of
>>>>>>> TB; I can't put a time-frame/version number on it.  It's momentary and
>>>>>>> personally of no concern.
>>>>>>
>>>>>> For me, the delay is from around 5 seconds to as high as 10 (observed)
>>>>>> and maybe even longer.
>>>>>>
>>>>>> That's simply unacceptable to me.
>>>>>>
>>>>>
>>>>> So I suppose I should try a send into this group with TB 152.0 .
>>>>>
>>>>> I have my video recorder running, and my Wireshark running.
>>>>
>>>> From clicking "Sent" to complete, is around 5 seconds.
>>>>
>>>> Ryzen 5700G 8C 16T , lots of RAM,
>>>>   Win10 x86 in VBox, 3072MB RAM, TB152.0 x86 installed.
>>>>
>>>> Unfortunately, the timestamp in Wireshark was not set to real time,
>>>> so I lost my correlation...
>>>>
>>>> <indeterminate text prep time>
>>>> 141.27 to 142.14 seconds
>>>>    AUTHINFO user xxxxxxxx
>>>>    AUTHINFO pass yyyyyyyy
>>>>    281 Authentication succeeded
>>>> 142.14
>>>>    POST
>>>>    340 OK recommended Message-ID <111jb19$3q87m$1@dont-email.me>
>>>>    <Burst of one-sentence-per-packet
>>>>     142.2870 - 142.2930    42 packets in 0.006 seconds
>>>>    >
>>>>    <Burst of ACKS
>>>>     142.4125 - 142.4568    21 ACKS in 0.0443 seconds
>>>>    >
>>>> 143.49
>>>>    240 Article Received
>>>>
>>>> About 2.2 seconds on the wire, using a fairly inefficient
>>>> transmission scheme. Perhaps sending one line per packet
>>>> is some sort of goofy RFC ?
>>>>
>>>> So now we have to figure out where the rest of the time went.
>>>> I'll time this one on the way out.
>>>>
>>>>     Paul
>>>
>>> Imagine a stopwatch is running. Some of the time mentioned
>>> here is "seconds after the top tick on the stop watch".
>>> The transaction starts at 11 seconds on the stop watch and
>>> ends at 18 seconds on the stop watch.
>>>
>>> The longer message took from 11 seconds to 18 seconds, for 7 seconds elapsed.
>>>
>>> Around 16-17.5 seconds is the network phase.
>>> Because I was careful to authenticate before posting,
>>> it did not need to authenticate before the POST request.
>>> At 17.5 seconds it did 240 Article Received.
>>>
>>> OK, that leaves 11-16 for textual preparation.
>>>
>>> Procmon tells me that at 13 seconds, there is the
>>> thunder of hoofbeats as "Avalon Graphics" are called.
>>> In the VBOX video recording, the graphics box appears at 16 seconds.
>>> This means "assembling" might actually correlate better with
>>> the wire phase of the thing.
>>>
>>> For procmon, there is radio silence up to 13 seconds, nothing (as usual)
>>> to tell me when I clicked. Procmon is not Windbg, and
>>> only a limited set of events are recorded.
>>>
>>> At 16.1 seconds,
>>>
>>>    C:\users\bullwinkle\AppData\Local\Temp\nsemail.eml   2386 byte write
>>>
>>> so the message preparation is hidden somewhere else even
>>> more temporary than that. Whether the text is "gathered"
>>> before the 13 second mark, or some time between 13 and 16,
>>> I cannot determine as there are no marks.
>>>
>>> So that is the file which will eventually make it to the wire.
>>> At 16.16 seconds, that's when the POST packet is sent for the
>>> wire phase of the operation. And roughly around that
>>> same time point, the captured video of the VM claims
>>> the status box pops out.
>>>
>>> It's possible that the "baggage" of calling
>>> Avalon Graphics is having a high impact. Maybe in
>>> threading-land, they should have "warmed up" the
>>> Avalon Graphics at 11 seconds, even if there
>>> was no status to write to it at that point.
>>>
>>> I would guess a craftsman could optimize this a tiny
>>> bit, or, just dump the status box.
>>>
>>> I don't have any evidence to convince a jury. I can't
>>> see everything using three tools, and that's about
>>> all I can describe of what happens.
>>>
>>> I consider it a bit weird, that a message is sent, with
>>> a packet per sentence (when a packet could be as big
>>> as 1500 bytes). When I've seen black hole routing problems
>>> caused by my other email tool sending 1500 byte packets.
>>> But sending the small packets, does not cost anything
>>> for messages this small. It would not be a good way to
>>> send a 20MB email :-)
>>
>> Thanks for all this info, Paul. Very interesting.
>>
>> Strange thing is that this morning, several of my messages have sent
>> very quickly.
> 
> But not this last one. It took around 6 seconds to "assemble and send".
> 

It adds only a tiny increment of time, to re-authenticate and then
do the POST and packet-per-sentence thing.

The time to prepare the nsemail file, seems a bit excessive.

Authentication has been busted before, so it's not like authentication
always happens instantly. But then your "time" would be a lot
longer than 6 seconds, if that style of flaw at the server end
was involved.

   Paul


   Paul

Back to alt.comp.software.thunderbird | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-22 02:46 -0700
  Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-22 02:47 -0700
  Re: Slow message send Andy Burns <usenet@andyburns.uk> - 2026-06-22 11:04 +0100
    Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-23 06:31 -0700
      Re: Slow message send "Carlos E. R." <robin_listas@es.invalid> - 2026-06-23 19:23 +0200
        Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-24 04:34 -0700
  Re: Slow message send "Alan K." <alan@invalid.com> - 2026-06-22 07:28 -0400
    Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-23 06:53 -0700
  Re: Slow message send VanguardLH <V@nguard.LH> - 2026-06-22 12:01 -0500
    Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-23 07:18 -0700
      Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-23 07:20 -0700
        Re: Slow message send VanguardLH <V@nguard.LH> - 2026-06-23 12:34 -0500
          Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-24 05:38 -0700
            Re: Slow message send VanguardLH <V@nguard.LH> - 2026-06-24 17:18 -0500
      Re: Slow message send VanguardLH <V@nguard.LH> - 2026-06-23 12:02 -0500
        Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-24 05:48 -0700
          Re: Slow message send Nobody <jock@soccer.com> - 2026-06-24 09:10 -0700
            Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-25 05:20 -0700
              Re: Slow message send Nobody <jock@soccer.com> - 2026-06-25 09:15 -0700
                Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-26 06:24 -0700
                Re: Slow message send Nobody <jock@soccer.com> - 2026-06-26 11:46 -0700
                Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-26 22:45 -0400
                Re: Slow message send Nobody <jock@soccer.com> - 2026-06-27 11:27 -0700
                Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-26 22:50 -0400
                Re: Slow message send VanguardLH <V@nguard.LH> - 2026-06-26 23:44 -0500
                Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-27 06:59 -0400
                Re: Slow message send "Carlos E. R." <robin_listas@es.invalid> - 2026-06-27 14:09 +0200
                Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-27 10:33 -0400
                Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-28 00:55 -0700
          Re: Slow message send VanguardLH <V@nguard.LH> - 2026-06-24 17:15 -0500
            Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-25 05:21 -0700
  Re: Slow message send Frank Miller <miller@posteo.ee> - 2026-06-22 19:44 +0200
    Re: Slow message send Andy Burns <usenet@andyburns.uk> - 2026-06-22 19:03 +0100
      Re: Slow message send Frank Miller <miller@posteo.ee> - 2026-06-22 20:15 +0200
        Re: Slow message send Nobody <jock@soccer.com> - 2026-06-22 12:01 -0700
          Re: Slow message send "Carlos E. R." <robin_listas@es.invalid> - 2026-06-22 21:31 +0200
          Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-23 07:22 -0700
            Re: Slow message send "Mr. Man-wai Chang" <toylet.toylet@gmail.com> - 2026-06-23 22:34 +0800
            Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-25 09:38 -0400
              Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-25 10:28 -0400
                Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-25 11:45 -0400
                Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-26 06:28 -0700
                Re: Slow message send "John C." <r9jmg0@yahoo.com> - 2026-06-26 06:28 -0700
                Re: Slow message send Paul <nospam@needed.invalid> - 2026-06-26 09:56 -0400
          Re: Slow message send "Mr. Man-wai Chang" <toylet.toylet@gmail.com> - 2026-06-23 22:34 +0800
  Re: Slow message send "Mr. Man-wai Chang" <toylet.toylet@gmail.com> - 2026-06-23 03:13 +0800
  Re: Slow message send "Mr. Man-wai Chang" <toylet.toylet@gmail.com> - 2026-06-24 14:51 +0800

csiph-web