Smalltalk › Squeak › Squeak - Dev

slow display update on debugger proceed

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

12 messages Options

Eliot Miranda-2

slow display update on debugger proceed

Hi All, Hi Bert,

a while ago I complained about the, now seemingly more freq

best,

Eliot

Tobias Pape

Re: slow display update on debugger proceed

On 27.03.2015, at 17:45, Eliot Miranda <[hidden email]> wrote:

> Hi All, Hi Bert,
>
> a while ago I complained about the, now seemingly more freq

ack. here too.

Eliot Miranda-2

Re: slow display update on debugger proceed

In reply to this post by Eliot Miranda-2

Hi All, Hi Bert,

a while ago I complained about extremely slow display update following proceed in the debugger. I can now confirm that it is nothing to do with the debugger simulating code and is simply the system being out of the defer updates state. Are there any fixes for this yet? Ah... I see a reply from Tobias, maybe all I need to do is update ;-)

On Fri, Mar 27, 2015 at 9:45 AM, Eliot Miranda <[hidden email]> wrote:

Hi All, Hi Bert,

a while ago I complained about the, now seemingly more freq

--
best,
Eliot

best,

Eliot

Tobias Pape

Re: slow display update on debugger proceed

On 27.03.2015, at 17:47, Eliot Miranda <[hidden email]> wrote:

> Hi All, Hi Bert,
>
> a while ago I complained about extremely slow display update following proceed in the debugger. I can now confirm that it is nothing to do with the debugger simulating code and is simply the system being out of the defer updates state. Are there any fixes for this yet? Ah... I see a reply from Tobias, maybe all I need to do is update ;-)

Sadly, no.
I have next to no knowledge of that part of Squeak.

Best
-Tobias

Eliot Miranda-2

Re: slow display update on debugger proceed

If there are any experts here I beg you to take a look. This is really impacting ,y productivity with VM development. I lose ~ 30seconds on almost every proceed in the debugger and I'm spending most of my time there-in simulating ARM execution right now. This is really painful. The ARM JIT and hence much faster Squeak on RPi and Android will arrive all the sooner if this is fixed.

AdvThanksance

On Fri, Mar 27, 2015 at 10:31 AM, Tobias Pape <[hidden email]> wrote:

On 27.03.2015, at 17:47, Eliot Miranda <[hidden email]> wrote:

> Hi All, Hi Bert,
>
> a while ago I complained about extremely slow display update following proceed in the debugger. I can now confirm that it is nothing to do with the debugger simulating code and is simply the system being out of the defer updates state. Are there any fixes for this yet? Ah... I see a reply from Tobias, maybe all I need to do is update ;-)

Sadly, no.
I have next to no knowledge of that part of Squeak.

Best
-Tobias

best,

Eliot

Levente Uzonyi-2

Re: slow display update on debugger proceed

Does it help if you revert WorldState >> interCyclePause: to its previous
version (al 7/31/2007 16:12)?

Levente

On Fri, 27 Mar 2015, Eliot Miranda wrote:

> If there are any experts here I beg you to take a look. This is really impacting ,y productivity with VM development. I lose ~ 30seconds on almost every proceed in the debugger and I'm spending most of my
> time there-in simulating ARM execution right now. This is really painful. The ARM JIT and hence much faster Squeak on RPi and Android will arrive all the sooner if this is fixed.
> AdvThanksance
>
> On Fri, Mar 27, 2015 at 10:31 AM, Tobias Pape <[hidden email]> wrote:
>
> On 27.03.2015, at 17:47, Eliot Miranda <[hidden email]> wrote:
>
> > Hi All, Hi Bert,
> >
> > a while ago I complained about extremely slow display update following proceed in the debugger. I can now confirm that it is nothing to do with the debugger simulating code and is simply
> the system being out of the defer updates state. Are there any fixes for this yet? Ah... I see a reply from Tobias, maybe all I need to do is update ;-)
>
> Sadly, no.
> I have next to no knowledge of that part of Squeak.
>
> Best
> -Tobias
>
>
>
>
> --
> best,Eliot
>
>

Eliot Miranda-2

Re: slow display update on debugger proceed

On Fri, Mar 27, 2015 at 2:37 PM, Levente Uzonyi <[hidden email]> wrote:

Does it help if you revert WorldState >> interCyclePause: to its previous version (al 7/31/2007 16:12)?

No :(. But that's at least information.

Levente

On Fri, 27 Mar 2015, Eliot Miranda wrote:

If there are any experts here I beg you to take a look. This is really impacting ,y productivity with VM development. I lose ~ 30seconds on almost every proceed in the debugger and I'm spending most of my
time there-in simulating ARM execution right now. This is really painful. The ARM JIT and hence much faster Squeak on RPi and Android will arrive all the sooner if this is fixed.
AdvThanksance

On Fri, Mar 27, 2015 at 10:31 AM, Tobias Pape <[hidden email]> wrote:

On 27.03.2015, at 17:47, Eliot Miranda <[hidden email]> wrote:

> Hi All, Hi Bert,
>
> a while ago I complained about extremely slow display update following proceed in the debugger. I can now confirm that it is nothing to do with the debugger simulating code and is simply
the system being out of the defer updates state. Are there any fixes for this yet? Ah... I see a reply from Tobias, maybe all I need to do is update ;-)

Sadly, no.
I have next to no knowledge of that part of Squeak.

Best
-Tobias

--
best,Eliot

best,

Eliot

marcel.taeumel (old)

Re: slow display update on debugger proceed

I will try to debug it if I can reproduce it. Trying "self halt" in a workspace followed by a proceed in the debugger did not make my image slow.

So, how to reproduce that bug?

Best,
Marcel

Eliot Miranda-2

Re: slow display update on debugger proceed

Hi Marcel,

On Sat, Mar 28, 2015 at 9:49 AM, Marcel Taeumel <[hidden email]> wrote:

I will try to debug it if I can reproduce it. Trying "self halt" in a
workspace followed by a proceed in the debugger did not make my image slow.

So, how to reproduce that bug?

Try and create *lots* of large overlapping windows, including a transcript. Then write some code that writes something to the transcript and put a halt in it. Then proceed. That seems close to what I'm doing. HTH and thanks for taking a look!

best,

Eliot

marcel.taeumel (old)

Re: slow display update on debugger proceed

busy-image.png

I tried this. No slowdown detected. :-/

Best,
Marcel

marcel.taeumel (old)

Re: slow display update on debugger proceed

I fixed something: http://forum.world.st/The-Trunk-Morphic-mt-793-mcz-td4816096.html

There was the possibility, that two UI processes were running after proceed, which may have caused the slow display update.

Now, it's more like the MVC-version (#mvcResumeProcess:) of resuming the process, which first makes the redraw and them resumes the interrupted process.

Eliot, does this work for you?

Best,
Marcel

Eliot Miranda-2

Re: slow display update on debugger proceed

Hi Marcel,

On Mon, Mar 30, 2015 at 6:49 AM, Marcel Taeumel <[hidden email]> wrote:

I fixed something:
http://forum.world.st/The-Trunk-Morphic-mt-793-mcz-td4816096.html

There was the possibility, that two UI processes were running after proceed,
which may have caused the slow display update.

Now, it's more like the MVC-version (#mvcResumeProcess:) of resuming the
process, which first makes the redraw and them resumes the interrupted
process.

Eliot, does this work for you?

YES!!! Yay! Thank you so much Marcel!

Best,
Marcel

--
View this message in context: http://forum.world.st/slow-display-update-on-debugger-proceed-tp4815619p4816097.html

Sent from the Squeak - Dev mailing list archive at Nabble.com.

best,

Eliot