VM Crash on Linux

Hi Everyone,

I wish I could give better information as to why this occurred, but I couldn't find a crash log file generated or anything similar.  Any advice or suggestions that you have on how I can gather better information would be appreciated and I will follow up on them.

I was working to save a few hours of work to a Monticello repository and when I clicked the "Save" button to bring up the commit message window the VM crashed dumping me back at the command line with this trace:

sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
sound_Start(default)
soundStart: snd_add_pcm_handler: Function not implemented
I/O possible

I am not sure why it was trying to play a sound, but it looks like that was the cause.  This is the latest Cog VM taken from Eliot, r2361.  Anyway, I was able to recover my changes so it wasn't a huge inconvenience, but I figured that I would report it because I had just worked past the UUID crash and I was feeling good about smalltalking on my linux box again and then this hit me.

Thanks for your help,
Jeff G.

On Wed, 23 Feb 2011, Jeff G wrote:

The symptoms are the same as the UUID bug's. Did you delete UUIDPlugin?
Does evaluating [UUID new] crash the VM?
Are you sure that the console messages appeared after (or around) the
crash?


Levente

>
> Thanks for your help,
> Jeff G.
>

greetings all,

To introduce myself, I first experienced Smalltalk around 1990 as part
of my undergrad in IT Engineering.  It made a great impression on me but
unfortunately I never got a chance to use it professionally, and hadn't
touched it since.  I then worked fifteen years in IT doing
desktop/server/network support of windows/unix/linux systems.  My career
got a bit stale so the past five years I have been doing external
postgrad study at the University of Southern Queensland for a MEngTech
Power Systems.  The past four years I've been working for a small
Electrical Engineering & Contracting firm doing industrial control and
power systems.

My Final Project & Dissertation is now upon me and I've decided to
develop an engineering design tool.  The discovery of Squeak a few weeks
ago and then working through Squeak by Example and the Laser Tutorial
was a major influence in choosing to undertake this project.  Part of it
is a schematic and connection drawing package where changes are live
linked between associated drawings. For example, when wire numbers are
changed on the schematic that will be reflected on the connection diagram.

So now I need to evaluate and choose a framework.  So far the options I
see are:

1. Pure Morphic - might has the advantage of being well supported with
examples.   I can already imagine how things might be implemented -
including maintaining vertical-or-horizontal-only orientations while
moving wires around. It is more likely that unintended behavior can be
understood and prevented.  The model coupling is also well demonstrated
in examples.

2. Ned Konz's Connectors  - I have been playing with Ned Konz's
Connectors framework this week and am quite excited by its potential.  
However I wonder how well it integrates with an underlying model to
maintain consistency across multiple views.

3. GraphViz - seems to be less interactive than Connectors providing
mainly graph auto-generation - though I haven't tried it yet.

4. Mondrian - seems to be less interactive than Connectors providing
mainly graph auto-generation - though I haven't tried it yet.

5. Cobalt, for later addition of 3D equipment layouts and cable
routing.   I could only spend a few hours with it.

In general, the more developed frameworks should provide more
sophisticated features "for free", but the trade offs are time-to-learn,
framework maturity, susceptibility to framework bugs, potential
complexity to trouble-shoot framework internals.  Also there may end up
a mismatch between requirements and framework possibilities, and being
able to extend them.  There also doesn't seem to be a great deal of
consolidated documentation for each framework.  So trying to understand
the high level design and usage philosophy by wading through low-level
code is awkward.

I am interested in your opinion on the choice of suitable frameworks?
Did I miss any?

The basic model will be composed of:
a. Drawings - A3 size with standard client borders. A large electrical
project may have thousands of drawings.
b. Components - with terminals to which wires connect.  These will be
developed custom by the user.  Components have Subparts that scatter
around a drawing.  For example a relay has an input-coil and an
output-contact that appear in different parts of the circuit. .  
Component Subparts Components will have various graphical
representations depending on which type of drawing they are on.
c. Wires - which inter-connect component terminals.  All
short-circuit-connected wires should have the same wire number.
d. Cables - contain multiple wires to connect between drawings and
physical locations.
e. Voltages and currents on components and in wires.

Design rule checking (such as preventing mixed voltages) will be done
against the model.  The schematic will be simulated to observe the
effect of push buttons and relay logic.

cheers, Ben

Here are a few links I've saved that might be interesting to you or give you some leads and/or insight:

AR-CAD
<http://www.ar-cad.com/index.html>

FREECAD
<http://www.oopgames.com/freecad/>

StCAD
<http://www.ar-cad.com/stcad.html>

Dr. Geo II
<http://documentation.ofset.org/drgeo/videos/drgeo2-center.mpeg>

Géométrie Interactive Enrichie Et Mosaïque
<http://people.ofset.org/hilaire/drgeo2/demos/4-mosaique/index.html>

ODE - Open Dynamics Engine
<http://g-24.net/squeak/ode.html>
<http://languagegame.org:8080/ggame/15>

UIDesigner
<https://www.hpi.uni-potsdam.de/hirschfeld/trac/SqueakCommunityProjects>
<https://www.hpi.uni-potsdam.de/hirschfeld/trac/SqueakCommunityProjects/wiki>

Thinglab
<http://wiki.squeak.org/squeak/607>

I'm an EE so look forward to following your work.
Good wishes.

   Ken G. Brown

At 1:07 AM +0800 2/25/11, [hidden email] apparently wrote:

Hi Ben and welcome,

lsobcca> was a major influence in choosing to undertake this project.  Part of it
lsobcca> is a schematic and connection drawing package where changes are live
lsobcca> linked between associated drawings. For example, when wire numbers are
lsobcca> changed on the schematic that will be reflected on the connection diagram.

for the model go for Squeak or any other Smalltalk.

lsobcca> 1. Pure Morphic - might has the advantage of being well supported with
lsobcca> examples.   I can already imagine how things might be implemented -
lsobcca> including maintaining vertical-or-horizontal-only orientations while
lsobcca> moving wires around. It is more likely that
lsobcca> unintended behavior can be
lsobcca> understood and prevented.  The model coupling is also well demonstrated
lsobcca> in examples.

If every graphic object should be a Morph I suggest you as soon as
possible calculate how many Morphs you want to be on the screen and
respond to events (like changing a wire number). Then just crate that
number of Morphs at random positions. !!! Be sure to use a throw away
image !!!. Years ago I tried with 30000 points (EllipseMorphs) and a
few thousand lines (PolygonMorphs) which rendered the image
unresponsive. Computers and VM's became faster but I assume there are
still limits.

I had the same issues with BabySRE which is built on Connectors.

Try that with every framework you evaluate.

Otherwise I love Morphic for non standard UI's and handle a few
hundred semi transparent interactive Morphs. But Morphic can be
tricky.

To display 1200 x 300 nodes of a neural network I bitblt to a Form.

Juan Vuletich has done a great job at simplifying cleaning and
speeding up his version of Morphic in Cuis.

Cheers,

Herbert

On 2/24/11 3:07 PM, "[hidden email]"
<[hidden email]> wrote:

> 1. Pure Morphic - might has the advantage of being well supported with
> examples.   I can already imagine how things might be implemented -
> including maintaining vertical-or-horizontal-only orientations while
> moving wires around. It is more likely that unintended behavior can be
> understood and prevented.  The model coupling is also well demonstrated
> in examples.


In http://wiki.squeak.org/squeak/5991 you could see a picture and a old .pr
of a Logic Gates Simulator, all Morphic.
Tomorrow morning I update to 4.2, think you have a good starting point.

Edgar

I've experienced this a couple of times myself, and at the same point
(MC operation).  Makes me wonder if there is a blasted "beep"
somewhere in the code..

I think the same thing happens if you try to run the test-suite.

To be safe, you can launch the VM with the -vm-sound-null argument.


On Thu, Feb 24, 2011 at 12:41 AM, Jeff G <[hidden email]> wrote:

I rescue of the dungeons the Jun 26 2003 SqueakLight complete Logic Gates
Simulator

Here you could download, works in Mac with John non Cog VM

http://www.4shared.com/file/RXI0B2kF/LogicCircusRTS.html

LogicCircus.RTS.image of 3.9 Mb , as you could guess RTS means Run Time
System meaning you could use but could¹t change it.

I convert the project to 4.2 and actualize the wiki info at some point today

Edgar