visualpython-users

[Visualpython-users] Release: VPython 2.9.7

[Jonathan B. <jbr...@ea...>]

This release is expected to be the last beta before releasing VPython
3.0 from this series.

All renderable objects contain a new member function, __copy__(), which
produces a duplicate of the old object after applying a user-defined set
of keyword=value argument pairs.

Some internal changes have been made to produce a smaller extension
module.

Visual cooperates with numarray when it is installed.  Numeric is still
required, but both Numeric and numarray may be installed at the same
time.

The build scripts for OSX have been updated.  The source tree contains a
file, mkdist_osx.sh, that will download, build, and install GCC and
Boost.Python from source.  Run the script from its own directory.  You
must have fink and the fink packages python23, numeric-py23, gtkglarea,
pkgconfig, and wget installed to use the script.


Enjoy,
Jonathan Brandmeyer

Re: [Visualpython-users] Release: VPython 2.9.7

[Joe H. <hea...@ct...>]

I am very pleased to report that I was able to build Visual and VPython 
from the latest source on Mac OS 10.3.4 with no major problems. I used 
the mkdist_osx.sh script for the entire process. I now have an extra 
copy of gcc on my system, but it doesn't seem to interfere with 
anything as far as I can tell. The build process seemed to bring my 
entire system to a grinding halt (quite literally) during compilation 
of the vector array stuff. I've no idea why. I shut down and restarted 
and reran the script and all worked just fine. It appears that my 
previous VPython installation was completely overwritten, which is what 
I would have expected.

The help facility doesn't work within IDLE, but I think I have an old 
email somewhere explaining how to fix that. I note that VPython windows 
seem "snappier" and more responsive than in the version I was using. 
The new crystal.py demo is simply amazing. I've not yet tried a program 
with a graph to see if those issues have been cleared up, but I'll try 
tonight or tomorrow.

Cheers,
Joe Heafner -- Astronomy/Physics Instructor (by some definitions)

Re: [Visualpython-users] Release: VPython 2.9.7

[Jonathan B. <jbr...@ea...>]

On Fri, 2004-07-23 at 16:08, Joe Heafner wrote:
> I am very pleased to report that I was able to build Visual and VPython 
> from the latest source on Mac OS 10.3.4 with no major problems. 

Excellent!

> I used 
> the mkdist_osx.sh script for the entire process. I now have an extra 
> copy of gcc on my system, but it doesn't seem to interfere with 
> anything as far as I can tell. 

I used several options to GCC's configure program to ensure that the one
being built would not interfere with the stock compiler.
- It is installed into /usr/local.
- All of the generated programs have the full GCC version number
appended to their name (eg, gcc-3.3.4 rather than gcc).
- The compiler's libraries are installed into version-specific
directories, so the system compiler will not pick up the new GCC's libs
by accident.

> The build process seemed to bring my 
> entire system to a grinding halt (quite literally) during compilation 
> of the vector array stuff. I've no idea why. 

How much memory do you have?  The system might have started swapping out
to disk.  Actually, now that I think about it, I probably should remove
'-j2' from the make commands in that script.  That option causes make to
run two jobs at once, and a few of the source files require ~200 MB of
RAM to compile - my development machine is a dualie.

Is there a way to determine how many processors are installed on a
PowerMac from the command line?

> I shut down and restarted 
> and reran the script and all worked just fine. It appears that my 
> previous VPython installation was completely overwritten, which is what 
> I would have expected.
> 
> The help facility doesn't work within IDLE, but I think I have an old 
> email somewhere explaining how to fix that. I note that VPython windows 
> seem "snappier" and more responsive than in the version I was using. 
> The new crystal.py demo is simply amazing. I've not yet tried a program 
> with a graph to see if those issues have been cleared up, but I'll try 
> tonight or tomorrow.

The only visible issue with graphs may be a bit of a stutter on startup.

Thanks for the feedback,
-Jonathan

Re: [Visualpython-users] Release: VPython 2.9.7

[Joe H. <hea...@ct...>]

On Jul 23, 2004, at 4:42 PM, Jonathan Brandmeyer wrote:

> I used several options to GCC's configure program to ensure that the 
> one
> being built would not interfere with the stock compiler.
> - It is installed into /usr/local.
> - All of the generated programs have the full GCC version number
> appended to their name (eg, gcc-3.3.4 rather than gcc).
> - The compiler's libraries are installed into version-specific
> directories, so the system compiler will not pick up the new GCC's libs
> by accident.

Great work!

>> The build process seemed to bring my
>> entire system to a grinding halt (quite literally) during compilation
>> of the vector array stuff. I've no idea why.
>
> How much memory do you have?  The system might have started swapping 
> out
> to disk.  Actually, now that I think about it, I probably should remove
> '-j2' from the make commands in that script.  That option causes make 
> to
> run two jobs at once, and a few of the source files require ~200 MB of
> RAM to compile - my development machine is a dualie.

I used my iBook G4 with 256MB RAM. I was running some other things in 
the background, including a real time Earth illumination program that 
uses the entire desktop (EarthDesk...<http://www.timepalette.com>) and 
Mail. Window moves took like 15 seconds or longer. Things got a little 
faster after shutting down everything except the build process, but it 
still just sat there forever so I just restarted and all was fine. I 
will attempt to do all this on my iMac this weekend.

I've now got a problem with Fink. It seems that some of the packages in 
unstable don't want to update even though I've enabled the unstable 
tree and there are more recent versions available. I'll deal with that 
on the fink-users list though.

> Is there a way to determine how many processors are installed on a
> PowerMac from the command line?

No idea. I've not played with a PowerMac. I have my heart set on one of 
the upcoming iMacs, provided I survive this current onslaught of 
medical bills.

Thank you for your hard work!

Cheers,
Joe Heafner -- Astronomy/Physics Instructor (by some definitions)