New VM behaves odd?

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

New VM behaves odd?

Göran Krampe
Hi fellas!

I use IRCe under Linux in a 3.8 image and have noticed that with the new
VM the UI "blocks" when doing operations like connecting or typing a
line of text in the IRC channel etc.

For example, I type in a line and press enter - it "hangs" for a second
obviously trying to send the line or something, and when it succeeds I
get the UI back. I can't say it behaved like this before.

Same weird thing happens when I move the IRCe window I think, but I am
guessing that is also related to communication - perhaps IRCe tries to
poll for more data or something, I dunno.

So my question is: Have anyone else noticed Squeak behaving differently
with the UI having "blocks" like this with the newest VM? And of course
- happen to know why? ;)

I haven't had time to figure out a test case yet, just wanted to check
my sanity first.

regards, Göran

Reply | Threaded
Open this post in threaded view
|

Re: New VM behaves odd?

Matej Kosik-2
[hidden email] wrote:

> Hi fellas!
>
> I use IRCe under Linux in a 3.8 image and have noticed that with the new
> VM the UI "blocks" when doing operations like connecting or typing a
> line of text in the IRC channel etc.
>
> For example, I type in a line and press enter - it "hangs" for a second
> obviously trying to send the line or something, and when it succeeds I
> get the UI back. I can't say it behaved like this before.
>
> Same weird thing happens when I move the IRCe window I think, but I am
> guessing that is also related to communication - perhaps IRCe tries to
> poll for more data or something, I dunno.
>
> So my question is: Have anyone else noticed Squeak behaving differently
> with the UI having "blocks" like this with the newest VM? And of course
> - happen to know why? ;)
>
> I haven't had time to figure out a test case yet, just wanted to check
> my sanity first.
>
> regards, G�ran
Hi,

I wanted to test it (to install the IRCe package) but the server which should provide it
(kilana) is down. I do not have offline copy of that package(s).
--
Matej Kosik




signature.asc (196 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

RE: New VM behaves odd?

Gary Chambers

I've noticed regular ui pauses in general with the new vm too.


Reply | Threaded
Open this post in threaded view
|

Re: New VM behaves odd?

timrowledge

On 3-May-06, at 11:17 AM, Gary Chambers wrote:

>
> I've noticed regular ui pauses in general with the new vm too.
>
Which platform(s), which vm, which image?


tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
"Wibble" said Pooh the stress beginning to show.



Reply | Threaded
Open this post in threaded view
|

Re: New VM behaves odd?

johnmci
I'd be quite curious if this problem shows on the mac os-x platform  
(powerpc or macintel) with the carbon VM, versus say the unix VM.
There were changes to how Sockets are serviced when a pending  
interrupt  happens, also in how the sleep logic works. The unix VM
uses nano-sleep, the carbon vm uses select(), which seems to make  
difference in cpu usage. But the question is does that affect the  UI.

Readers may recall lately I changed the carbon VM and moved the  
interpreter processing to the main UI thread which means events are  
technically polled for, versus being serviced immediately on another  
pthread. You can see that behaviour by accessing the mac squeak menu  
and holding the menu open and watching squeak stop running the VM as  
we block servicing the os-x menu.  Screams and $$ might fix that  
problem, but who uses the mac menu bar anyway.   Noting there is code  
and vm support to enable a mac  squeak image to alter the mac menu  
bar how one sees fit and to trigger actions in smalltalk as a result  
of picking said menu item.  That said no-one reported any weird  
issues with UI interactions on the mac carbon vm so far as I know.


On 3-May-06, at 1:01 PM, tim Rowledge wrote:

>
> On 3-May-06, at 11:17 AM, Gary Chambers wrote:
>
>>
>> I've noticed regular ui pauses in general with the new vm too.
>>
> Which platform(s), which vm, which image?
>
>
> tim
> --
> tim Rowledge; [hidden email]; http://www.rowledge.org/tim
> "Wibble" said Pooh the stress beginning to show.
>
>
>

--
========================================================================
===
John M. McIntosh <[hidden email]> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
========================================================================
===


Reply | Threaded
Open this post in threaded view
|

Re: New VM behaves odd?

Colin Putney

On May 3, 2006, at 4:13 PM, John M McIntosh wrote:

> I'd be quite curious if this problem shows on the mac os-x platform  
> (powerpc or macintel) with the carbon VM, versus say the unix VM.
> There were changes to how Sockets are serviced when a pending  
> interrupt  happens, also in how the sleep logic works. The unix VM
> uses nano-sleep, the carbon vm uses select(), which seems to make  
> difference in cpu usage. But the question is does that affect the  UI.
>
> Readers may recall lately I changed the carbon VM and moved the  
> interpreter processing to the main UI thread which means events are  
> technically polled for, versus being serviced immediately on  
> another pthread. You can see that behaviour by accessing the mac  
> squeak menu and holding the menu open and watching squeak stop  
> running the VM as we block servicing the os-x menu.  Screams and $$  
> might fix that problem, but who uses the mac menu bar anyway.    
> Noting there is code and vm support to enable a mac  squeak image  
> to alter the mac menu bar how one sees fit and to trigger actions  
> in smalltalk as a result of picking said menu item.  That said no-
> one reported any weird issues with UI interactions on the mac  
> carbon vm so far as I know.

Funny, I just noticed UI issues in 3.8.11. Sometimes the Mac menu bar  
doesn't respond for quite some time (10s of seconds), during which,  
Morphic *is* responsive.

I'll see if I can reproduce reliably and get you a better error report.

Colin