World: r3wp

Well, it surely does not post back the whole chatset just for adding 
a single message ...

But maybe it does :) Who knows but Carl and those who wrote it :)

that would be carl too  ;-)

It'd be interesting to know the size of Graham's private messages 
file  -- 19.set in his case.  {Andreas, yours would be 107.set]

The files should be in altme/worlds/rebol3/chat

But beware that Windows may be virtualising the files, so the latest 
copy may be well hidden elsewhere,

It seems there's noone who wrote altme ... it just came from the 
outer space :-)

My private messages file is 384K.

Mine is 1.6meg. I suspect Graham can top that!

3 523 969 bytes :-)

I am a politician - I do many behind-the-scenes activity :-)

4.5MB here.

That's large, Maxim.

For comparison, the largest public,active, group is Core -- I see 
that as 3.8 meg in size.

yes... and its why I PM so little now.  its just painfull.

this group is < 1 sec.   PM sometimes even causes my altme to go 
into "internet busy" mode for so long, I loose track of the discussion.

I can see one larger group....REBOL3-old was 4.3meg.

And that was retired and replaced by the REBOL3 precisely because 
of the speed problem.

Doesn't the speed also have to do with the message limit per chat 
settings? I've decreased mine and  it's much faster. Though I do 
wonder how accurate that setting is or what it really measures. I 
set mine to a higher number after a search and warning that it needs 
to be set higher. I've thought that there's no way there could be 
that many messages (several 1000's) in that one topic.

it just limits what is visible in your scrollable view above.   it 
will affect speed so much as it uses less faces to display the list 
and thus less RAM is managed by the GC.

Yeah, I guess I should just search the data directly when I'm looking 
for something really old.

doing an OS (file content) search in the altme directory usually 
is easier and faster too.

Plus the [web-public] groups have a fast search on REBOL.org
    http://www.rebol.org/aga-search.r

19.set is a puny 4.2mb

:) It's one of the larger ones, I'd guess, even if Maxim is just 
ahead.

If you send a private message to yourself -- is that fast or slow?

he was hoping it would be twice as large as mine... but mine is bigger 
 ;-)

16 seconds to message myself

Awesome :)

So, it does appear to be group dependent

reminder to myself .. don't message myself !

Any chance to check if there's heavy network traffic during those 
16 seconds :) ?

I could look at the back off my laptop and count the flashes??

ok, I've got the windows task manager networking tab up now.

that message was not even a blip

same when posting to myself

so presumably it's internal handling that is the error.

now looking at cpu usage

a little spike when posting here ( reaches two bars )

drat ... looking at the wrong core

slightly less cpu use when posting here vs posting to myself

I've had to restart AltMe a couple times today, just to clear the 
unsent message field.

5 seconds for me.  No other bugs with AltME for the most part.

I have learned though that if you have a message ready to be sent, 
and change groups a few times, when you hit send, it is not 100% 
sure it will be sent to the corrrect group.

Maybe there should be some kind of tag written, just as send is pressed, 
so that it appears with guarantee in the correct group.

My guess is that it's the server that takes time to append the message 
in the chat file.

Dunno if it does a simple append to the file but if it needs to load 
it in memory, append the block, then write it down to disk, it could 
explain why bigger chat files take longer to update !!

Test

Looking at my Altme file activity (Filemon) when I post a message 
:

- Nothing happens until the server has done its job and send back 
the new message for the corresponding group/user.

- Then I see that the corresponding chat file is wrote entirely by 
block of 4096B !

- Then the state file (chatMYALTMEID) is updated (wrote by chunk 
of 4096B).

[I'm just guessing here] There are at least three different data 
sets to update when I send a private message to (say) Graham:
-1- my chat local *.set file

-2- whatever data sets are held on the server to record the message 
(I have no idea what these are)
-3- Graham's local *.set file


The first two could contribute to perceived slowness in sending; 
the third to any apparent slowness in updating/downloading messages.


Append/lines to a 4.5 meg file does not take long on my machine (not 
does read/lines) so I'm primed to suspect the problem is on the server.


What would be interesting to know is if the grayness I see when sending 
a message ends when:
-a- the server acknowledges receipt; or
-b- when the server has finished committing all its updates.


If it is -b-, then moving the server response to be -a- could create 
a significant perceived increase in responsiveness.

My guess is that own posts are updated locally like any other post 
: while the server send it back to you as an update. I mean: 1) the 
post is send to the server, 2) it update the DB, then 3) it send 
the post back to you, 4) your altme updates its files.

Anyone hosting some Altme world there ?

there=here

Yes, I have several.

James : It would be interesting for us if you can look the file activity 
of a world server when a new post is send.

I use Filemon from Sysinternals (no included in Process monitor) 
for that.
Can you test this for us and tell us what happens ?

Didier, Process Monitor includes both file and registry monitoring 
- it effectively obsoletes regmon and filemon. You might be thinking 
of Process Explorer.

Also, technically, a given send box should be directly connected 
to the given group, so you can send, and even if it is in transit 
or waiting you can go on and work on other messages.

We do tag messages this way in Qtask, but we have seen this same 
bug, but don't have proof (as I do in AltME) that it is the program 
and not the operator.  But I'm always on the look out for it since 
it can be a very dangerous bug for some people.


This whole class of issue falls to a perceived form of latency, and 
a series of features are needed to suppress it.  For example, instead 
of graying the out going message area, just make it look like it 
was sent, but also instantly add a “message hour glass” in the “New 
message” bar when on that group, and, also create a new tinted empty 
box in the thread saying something like “Waiting for new message 
to post”.


This frees you up to go elsewhere.  If you come back to the same 
group it is clear what state it is in.  If you want to retrieve the 
sent message just click on the “Message hour glass” and it will bring 
it back into the input box.


Lastly, of course, if you try to leave AltME before it is done sending, 
need to offer people a dialog warning them , and giving them access 
to the messages waiting.

There is more, but yeah, it takes a lot to make this work.