visualpython-users

Re: [Visualpython-users] Speed?!

[Anton S. <br...@po...>]

Susanne Moelbert wrote:
> I was very amazed yesterday when I wanted to calculate
> a simple thing with Python. I typed
> 
> for a in range(200):
>     for b in range(200):
>         for c in range(200):
>             pass
> 
> and it took over 10 seconds! . . .

range() works by creating a list [0, 1, ... 199].
The above code appears to do that 40201 times.

This saves about half the time:

	spam = range(200)
	for a in spam:
	  for b in spam:
	    for c in spam:
	      pass

-- 
Anton Sherwood  --  br...@p0...  --  http://ogre.nu/

Re: [Visualpython-users] Speed?!

[Markus G. <gr...@iu...>]

Susanne Moelbert wrote:

> Hi,
>
> I was very amazed yesterday when I wanted to calculate a simple thing
> with Python. I typed
>
> for a in range(200):
>     for b in range(200):
>         for c in range(200):
>             pass
>
> and it took over 10 seconds! I typed the same thing in Borland Delphi
> (on the same machine, a Pentium III) and it was immediate. One wouldn't
> even realize that it was calculating. In C++ it was slightly slower than
> in Delphi, but by far not as slow as Python. Why is it so terribly slow,
> even if it doesn't need to calculate anything? Is there a possibility
> for accelerating this?

Well, I guess the situation is this:

Delphi trys to optimize the code at compile time.  It recognizes, that
there is really nothing to do, so the whole loop construct is optimized
away.

C++ would probably be as fast as Delphi if you compile with -O3 or
something like this.

Python does no optimizations, so Python is really performing the loops,
even if there is nothing to do.  So in a real world example there would of
course be something to do in the inner loop, so Delphi and C++ won't return
emmediate.  The other point is that Python is interpreted (even if the
script is translated to bytecode before), and scripting languages tend to
be 20 to 30 times slower than compiled ones.  But if you want short
turn-around times, and want to have fun during programming, ...

Markus

[Visualpython-users] Speed?!

[Susanne M. <sus...@un...>]

Hi,

I was very amazed yesterday when I wanted to calculate a simple thing
with Python. I typed

for a in range(200):
    for b in range(200):
        for c in range(200):
            pass

and it took over 10 seconds! I typed the same thing in Borland Delphi
(on the same machine, a Pentium III) and it was immediate. One wouldn't
even realize that it was calculating. In C++ it was slightly slower than
in Delphi, but by far not as slow as Python. Why is it so terribly slow,
even if it doesn't need to calculate anything? Is there a possibility
for accelerating this? 

thanks
Susanne

-- 
-------------------------------------------------------
Susanne Moelbert
Universite de Fribourg
Departement of Theoretical Physics
CH-1700 Fribourg	
Switzerland		
			

Phone: ++41 (0)26 300 91 46  (office)
       ++41 (0)26 424 19 33  (home)
e-mail: sus...@un...

my homepage: http://ptheosg00.unifr.ch/~moelbert/
-------------------------------------------------------

RE: [Visualpython-users] (no subject)

[David A. <dm...@an...>]

> --On Friday, April 13, 2001 5:06 PM -0400 David Scherer 
<dsc...@vy...> wrote:

> What is desperately needed to fix this is someone with a Red Hat system, 
and
> the time and minimal expertise required to build packages.

Would my Mandrake installation be close enough to Red Hat to serve?     Ari 
could borrow my notebook for a day, or I could try to aquire the minimal 
expertise required ...

[dscherer@milliway.kssacct.com: Re: [Visualpython-users] scene range]

[Ari H. <ahe...@an...>]

I assume Dave meant to forward this to the list.

However, as I can tell by the lack of linebreaks in the original, it was
sent using his Blackberry, so he was probably sending it while running
around saving the planet and can be forgiven for the oversight :)

----- Forwarded message from David Scherer <dsc...@mi...> -----

> From: "David Scherer" <dsc...@mi...>
> Date: Fri, 13 Apr 2001 15:31:54 -0400
> Subject: Re: [Visualpython-users] scene range
> To: ahe...@an...
> 
> The limit Arthur is hitting is probably not double precision. It is most
> likely the z buffer range. Essentially, there is a limit on the ratio
> between the near and far clipping planes. As you increase the size of the
> scene, visual increases the far clipping plane to include the whole scene.
> When the far clip plane gets VERY far away, the near clip plane may start
> to get so distant from the camera that it engulfs most of the scene.
> 
> Dave

----- End forwarded message -----

RE: [Visualpython-users] (no subject)

[David S. <dsc...@vy...>]

> It's hard for this bystander to comprehend how Linux users put up
> with this!

The problem isn't really the OS, but our pathetic packages.  On Windows, the
(fairly complex) job of installing shared libraries, registry entries, etc
for Python, OpenGL, Tcl/Tk, cvisual, etc is handled by the Python 2.0
installer and our own.  On Linux distributions, it should be handled by the
package manager, based on dependency information in the packages.
Unfortuantely, Ari is trying to build rpm packages without access to a Red
Hat-like distribution, which is impossible.  Everyone else is therefore
trying to do things like compile from source, which is about as painful as
you would expect.

The only thing hard to understand is that anyone likes VPython enough to put
up with this ;)

> These dialogs sound like bad jokes.

That's just because you aren't used to the particular jargon being used, so
that /usr/lib/libgtkgl.so.4 sounds slightly "worse" to your ears than, say,
c:\windows\system\opengl32.dll.  These problems don't seem a lot worse than
the installation headaches that forced us to build our own Python
distribution for 1.5.2.

What is desperately needed to fix this is someone with a Red Hat system, and
the time and minimal expertise required to build packages.  I would also
advocate static linking against a few things like libstdc++ if possible,
even though that goes against normal practice on Linux.

Dave

> --On Thursday, April 12, 2001 20:13 -0400 Ari Heitner
> <ahe...@an...> wrote:
>
> > On Thu, Apr 12, 2001 at 04:26:14PM -0700, Gary Huber wrote:
> >> Yeah, mine also has glXUseXFont, according to the nm command.  It also
> >> works fine in my C++ code.  I have Mesa 3.4-1. In addition to being in
> >> /usr/X11R6/lib, I've copied libGL.so to /usr/lib and /usr/local/lib, to
> >> no avail.
> >>
> > Ok. Then do an
> >   ldd /usr/lib/libgtkgl.so.4
> > (or whatever version of libgtkgl you have cvisualmodule.so linked
> > against). are all the dependencies resolved? your libGL.so must be in a
> > dir that's in /etc/ld.so.conf ...

[Visualpython-users] RH Linux 6.2 problem, formerly (no subject)

[Gary H. <gh...@bi...>]

I copied the new cvisualmodule.so module from the website, and now things
work just fine.  Thanks to Andrew Morrison for this very simple
suggestion, and thanks also to Ari Heitner and the rest.

Gary

[Visualpython-users] scene range

[<Art...@rs...>]

>>>We're talking *big* scales here :)

Well yes.

In the simplest case, I might define two lines
each through two pickable points, and another point
as the intersection of the two lines.

As I manipulate the pickable points, the relationship
between the two lines can move towards parallel,
so the point of intersection between them is way out
there somewhere, at least momentarily. 

But the artifact is the scene going dead blank at
some point. If the point is still being dragged 
the scene re-appears once my intersection point
returns to some less extreme place.

Under that particular scenario the artifact is actually 
nothing more that a screen blink.  

But in some of what I am doing, the artifact presents
a more serious stumbling block. 

 I have tried some filtering, to have vectors essentially 
scaled down if any of x,y,or z exceed a max. 
The inaccuracy I can live with,  but the mechanism 
I have come up with so far has other implications - 
anything in the scene that depends on - let's say -
the intersection point loses its dependency once
the filtering kicks in, and permanently, i.e. even once
the point reattains a postion where the filtering
is not happening, other objects in the scene stop
"seeing" it as changing postion.

That might be a pure issue of my coding for the 
filtering, but it's a subtle issue by  my
standards.  Nothing I try seems to work.

Ideas welcome.

ART

Re: [Visualpython-users] scene range

[Ari H. <ahe...@an...>]

On Fri, Apr 13, 2001 at 10:24:52AM -0500, Art...@rs... wrote:
> However, there does seem to be a limiting factor - 
> if an object exists  X distance off the viewable frame,
> the rendering of the scene seems to falter.
> 
> Can anybody help me understand what is the limiting
> X that I seem to be running into?

I'm not sure if this is what you're running into, but there is a limit of
the Visual module's precision on the Primitive::pos attribute. The x,y,z
elements of position are doubles ... i suppose if your scale gets too big,
you'll see weird FP precision artifacts.

We're talking *big* scales here :)




ari

[Visualpython-users] scene range

[<Art...@rs...>]

In my dynamic geometry app, particularly since I am 
playing with projective geometry, it is not unusual
to have a scene with objects that can move to large
distances from one another quickly.

VPython allows me to move quickly between macro
and close-up views of the scene by mouse action.
Some of the effects of being able to do this while 
maintaining  high quality rendering are quite striking - 
allowing me, and hopefully others at some point, - 
to grasp some less than obvious geometric realities 
as - well - realities.

However, there does seem to be a limiting factor - 
if an object exists  X distance off the viewable frame,
the rendering of the scene seems to falter.

Can anybody help me understand what is the limiting
X that I seem to be running into?


ART

 


 

Re: [Visualpython-users] (no subject)

[Ari H. <ahe...@an...>]

On Thu, Apr 12, 2001 at 09:35:55PM -0400, Bruce Sherwood wrote:
> EGO (Eyes Glazing Over)
> 
> It's hard for this bystander to comprehend how Linux users put up with this!
> 
> These dialogs sound like bad jokes.
> 

I disagree.

If the installer were properly built for these linux distributions (as it is
for Windows/Mac) things would work considerably more smoothly. If we had
separate debian, rh6.2, and rh7 packages, everything would work the first
time (and yes i would put my money where my mouth is on that one).

At least when it breaks, however, there's a simple and direct recourse to
examine the situation and determine exactly what's broken. In Windows if
your GL drivers are broken, you try installing new ones from various
sources, and pray :)




Ari

Re: [Visualpython-users] (no subject)

[Bruce S. <ba...@an...>]

EGO (Eyes Glazing Over)

It's hard for this bystander to comprehend how Linux users put up with this!

These dialogs sound like bad jokes.

Bruce Sherwood

--On Thursday, April 12, 2001 20:13 -0400 Ari Heitner 
<ahe...@an...> wrote:

> On Thu, Apr 12, 2001 at 04:26:14PM -0700, Gary Huber wrote:
>> Yeah, mine also has glXUseXFont, according to the nm command.  It also
>> works fine in my C++ code.  I have Mesa 3.4-1. In addition to being in
>> /usr/X11R6/lib, I've copied libGL.so to /usr/lib and /usr/local/lib, to
>> no avail.
>>
> Ok. Then do an
>   ldd /usr/lib/libgtkgl.so.4
> (or whatever version of libgtkgl you have cvisualmodule.so linked
> against). are all the dependencies resolved? your libGL.so must be in a
> dir that's in /etc/ld.so.conf ...

Re: [Visualpython-users] (no subject)

[Ari H. <ahe...@an...>]

On Thu, Apr 12, 2001 at 04:26:14PM -0700, Gary Huber wrote:
> Yeah, mine also has glXUseXFont, according to the nm command.  It also
> works fine in my C++ code.  I have Mesa 3.4-1. In addition to being in
> /usr/X11R6/lib, I've copied libGL.so to /usr/lib and /usr/local/lib, to no
> avail.
> 
Ok. Then do an
  ldd /usr/lib/libgtkgl.so.4
(or whatever version of libgtkgl you have cvisualmodule.so linked against).
are all the dependencies resolved? your libGL.so must be in a dir that's in
/etc/ld.so.conf ...



ari

Re: [Visualpython-users] (no subject)

[Gary H. <gh...@bi...>]

Yeah, mine also has glXUseXFont, according to the nm command.  It also
works fine in my C++ code.  I have Mesa 3.4-1. In addition to being in
/usr/X11R6/lib, I've copied libGL.so to /usr/lib and /usr/local/lib, to no
avail.

Gary



On Thu, 12 Apr 2001, Ari Heitner wrote:

> On Thu, Apr 12, 2001 at 10:00:55AM -0700, Gary Huber wrote:
> > I have recently downloaded Visual Python; I'm running Red Hat 6.2.
> > I also have Mesa installed.  When I enter python and
> > try to import VPython, I get this:
> > 
> > Python 2.0 (#1, Oct 16 2000, 18:10:03) 
> > [GCC 2.95.2 19991024 (release)] on linux2
> > Type "copyright", "credits" or "license" for more information.
> > >>> import visual
> > Visual-2000-11-26
> > Traceback (most recent call last):
> >   File "<stdin>", line 1, in ?
> >   File "/usr/local/lib/python2.0/site-packages/visual/__init__.py", line
> > 17, in ?
> >     import cvisual
> > ImportError: /usr/lib/libgtkgl.so.4: undefined symbol: glXUseXFont
> > >>> 
> > 
> > I've tried putting my Mesa shared object file in many different
> > places, but with no luck.  Am I missing another piece? I though
> > glXUseXFont is part of OpenGL.
> > 
> It's part of GLX. And it *does* appear in my /usr/lib/libGL.so:
> 
> nimrod:/usr/lib$ nm -D libGL.so | grep glXUseXFont
> 0013eb3c T Fake_glXUseXFont
> 0013bf68 T glXUseXFont
> 
> What version on Mesa are you trying to use here? Are you sure it's a working
> installation? (does other 3d stuff work?)
> 
> 
> ari
> 
> 
> 

Re: [Visualpython-users] (no subject)

[Andrew M. <mo...@tb...>]

Gary:

I think this is the problem that is talked about on the website with
regard to RH6.2.   I was getting similar errors with my installation,
until I copied over the cvisualmodule.so from the vpython website over the
/usr/lib/python1.5/site-packages/cvisualmodule.so

Now, everything is working PERFECTLY for me!

Thank you so much, everyone, especially Ari...

Andrew


On Thu, 12 Apr 2001, Gary Huber wrote:

> I have recently downloaded Visual Python; I'm running Red Hat 6.2.
> I also have Mesa installed.  When I enter python and
> try to import VPython, I get this:
> 
> Python 2.0 (#1, Oct 16 2000, 18:10:03) 
> [GCC 2.95.2 19991024 (release)] on linux2
> Type "copyright", "credits" or "license" for more information.
> >>> import visual
> Visual-2000-11-26
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "/usr/local/lib/python2.0/site-packages/visual/__init__.py", line
> 17, in ?
>     import cvisual
> ImportError: /usr/lib/libgtkgl.so.4: undefined symbol: glXUseXFont
> >>> 
> 
> I've tried putting my Mesa shared object file in many different
> places, but with no luck.  Am I missing another piece? I though
> glXUseXFont is part of OpenGL.
> 
> Thanks
> 
> Gary
> 
> 
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> http://lists.sourceforge.net/lists/listinfo/visualpython-users
> 

Re: [Visualpython-users] (no subject)

[Ari H. <ahe...@an...>]

On Thu, Apr 12, 2001 at 10:00:55AM -0700, Gary Huber wrote:
> I have recently downloaded Visual Python; I'm running Red Hat 6.2.
> I also have Mesa installed.  When I enter python and
> try to import VPython, I get this:
> 
> Python 2.0 (#1, Oct 16 2000, 18:10:03) 
> [GCC 2.95.2 19991024 (release)] on linux2
> Type "copyright", "credits" or "license" for more information.
> >>> import visual
> Visual-2000-11-26
> Traceback (most recent call last):
>   File "<stdin>", line 1, in ?
>   File "/usr/local/lib/python2.0/site-packages/visual/__init__.py", line
> 17, in ?
>     import cvisual
> ImportError: /usr/lib/libgtkgl.so.4: undefined symbol: glXUseXFont
> >>> 
> 
> I've tried putting my Mesa shared object file in many different
> places, but with no luck.  Am I missing another piece? I though
> glXUseXFont is part of OpenGL.
> 
It's part of GLX. And it *does* appear in my /usr/lib/libGL.so:

nimrod:/usr/lib$ nm -D libGL.so | grep glXUseXFont
0013eb3c T Fake_glXUseXFont
0013bf68 T glXUseXFont

What version on Mesa are you trying to use here? Are you sure it's a working
installation? (does other 3d stuff work?)


ari

[aheitner@andrew.cmu.edu: Re: [Visualpython-users] What is IDLE?]

[Ari H. <ahe...@an...>]

d'oh, meant for this to go to the list.

----- Forwarded message from Ari Heitner <ahe...@an...> -----

> From: Ari Heitner <ahe...@an...>
> Date: Thu, 12 Apr 2001 14:41:23 -0400
> Subject: Re: [Visualpython-users] What is IDLE?
> To: Anton Sherwood <br...@po...>
>=20
> On Wed, Apr 11, 2001 at 08:38:25PM -0700, Anton Sherwood wrote:
> > Ari Heitner wrote:
> > > Compiling the Linux version is easy. You just need to
> > > get RPMs (from rpmfind.net) for the libraries it needs --
> >=20
> > Tried that.  Got nowhere.
> >=20
> > (Don't ask me for details because since then I managed to install a
> > broken version of rpm, and have failed so far to install a working
> > version.)
> >=20
>=20
> Yeah.
>=20
> See, that's a problem.
>=20
> I can fix problems when I know exactly what the error was. But "I gave =
up,
> It broke, I have no exact error messages" does nothing for me.
>=20
> I do *not* recommend trying to compile VPython if you have no experienc=
e
> compiling stuff before. However, if you *have* worked with C/C++/whatev=
er
> before, I think it's fairly straightforward -- the compiler yells at yo=
u
> about missing header files, and you either go find the appropriate libr=
ary
> on rpmfind.net, or (hopefully as a last resort) post to this list and s=
ay
> "ack, it broke with error X" and I respond and say "Oh no problem juts =
do
> Y".
>=20
>=20
>=20
> ...
>=20
> This doesn't address the real problem: given that RH 6.2, all evil libr=
ary
> versioning problems aside, appears to have a broken C++ compiler and a
> broken kernel, and that RH 7 still has library versioning issues, I sug=
gest
> using the debian packages for debian. What we really need is someone wh=
o
> actually *uses* RH and knows how to build RH packages (can't be that ha=
rd)
> to volunteer to maintain those packages. The alien'd DEBs appear to be =
more
> trouble then they are worth.
>=20
>=20
>=20
> Ari

----- End forwarded message -----

--=20
Ari Heitner
  DC: 703/5733512  PGH: 412/4229837
Non c'=E8 pi=F9 forza nella normalit=E0: c'=E8 solo monotonia.

[Visualpython-users] (no subject)

[Gary H. <gh...@bi...>]

I have recently downloaded Visual Python; I'm running Red Hat 6.2.
I also have Mesa installed.  When I enter python and
try to import VPython, I get this:

Python 2.0 (#1, Oct 16 2000, 18:10:03) 
[GCC 2.95.2 19991024 (release)] on linux2
Type "copyright", "credits" or "license" for more information.
>>> import visual
Visual-2000-11-26
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/local/lib/python2.0/site-packages/visual/__init__.py", line
17, in ?
    import cvisual
ImportError: /usr/lib/libgtkgl.so.4: undefined symbol: glXUseXFont
>>> 

I've tried putting my Mesa shared object file in many different
places, but with no luck.  Am I missing another piece? I though
glXUseXFont is part of OpenGL.

Thanks

Gary

Re: [Visualpython-users] Linux users of VPython, unite!

[Andrew M. <mo...@tb...>]

This would be a great help, I think.  Obviously, I have had some problems
getting going with the linux version.  

What was so great about the Windows version is you download two files
(assuming you didn't have python to begin with) which you double click in
the proper order to install.  Then, you use the provided short-cut to
start IDLE, open an included example program, and hit F5 to see if it
works.   That was painless.

For linux, I had to find all the dependencies first.  That wasn't so bad,
except the key dependency that I was missing in the end was labeled as a
RedHat7.0 rpm.  I am running Redhat 6.2, but it turned out the 7.0 rpm for
the libstdc++ worked on my system anyway.  After that, I got the vpython
rpm to install with no worries.

What also confused me was the whole IDLE issue.  When I write perl
programs or C programs or whatever, I usually use pico or vi.  The only
python programming I have done is "Hello World" type stuff, which I also
wrote using pico.  But, in Windows, it was more convenient to use the IDLE
that was included.  So, I thought that was something special about the
visual module that made the python program necessary to run even on a
linux system.  Now, I see I was foolish for thinking that.

I would help critique or edit a Howto or help page for the linux
installation.  Right now, I am still not familiar enough with it to
actually be able to create it myself.  Let me know what I can do to help.

Andrew

On Wed, 11 Apr 2001, Bruce Sherwood wrote:

> I'd like to propose a small Linux VPython community project. I find it 
> frustrating that what look to me like the same Linux questions and problems 
> keep recurring. Ari Heitner has been trying to provide some compiled 
> packages, but it's clear that these often are not usable, depending on the 
> Linux flavor, so there should be a high-quality document explaining how to 
> do a source compile and install.
> 
> I propose that someone knowledgeable draft and post FULLY detailed, 
> comprehensive instructions for how to compile and install on various common 
> flavors of Linux. Other knowledgeable Linux users would critique and debate 
> the document. When everyone had thrashed out agreement, we would post this 
> prominently on the VPython site. The document needs in particular to state 
> clearly and in detail all the dependencies, and all the places where one 
> has to go to get the various pieces. And if we get that far, maybe we could 
> even get all the pieces together in one place!
> 
> Feel free also to tell me that my proposal makes no sense. Since I'm not a 
> Linux user, I may well have misunderstood everything. But watching from the 
> sidelines it sure sounds like utter chaos that one is sent to lots of 
> different repositories to find obscure files and keep trying to compile and 
> find other dependencies and ask some more questions and update the compiler 
> itself to the latest version and then.....
> 
> Bruce Sherwood
> 
> 
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> http://lists.sourceforge.net/lists/listinfo/visualpython-users
> 

Re: [Visualpython-users] Linux users of VPython, unite!

[heafnerj <hea...@vn...>]

On Wed, 11 Apr 2001, Bruce Sherwood wrote:

> I'd like to propose a small Linux VPython community project. I find it 
> frustrating that what look to me like the same Linux questions and problems 
> keep recurring. Ari Heitner has been trying to provide some compiled 
> packages, but it's clear that these often are not usable, depending on the 
> Linux flavor, so there should be a high-quality document explaining how to 
> do a source compile and install.
> 
Excellent idea. From what I read on the VPython website, the problem was
with a library supplied with Red Hat Linux 6.2. From version 6.2 to 7, the
problem was apparently fixed with an updated library.

Personally, I much prefer precompiled rpm's rather than compiling my own
packages.

> I propose that someone knowledgeable draft and post FULLY detailed, 
> comprehensive instructions for how to compile and install on various common 
> flavors of Linux. Other knowledgeable Linux users would critique and debate 
> the document. When everyone had thrashed out agreement, we would post this 
> prominently on the VPython site. The document needs in particular to state 
> clearly and in detail all the dependencies, and all the places where one 
> has to go to get the various pieces. And if we get that far, maybe we could 
> even get all the pieces together in one place!
> 
In my own experience, the necessary dependencies were on the supplementary
CD-ROM that came with my distribution. Well, all but one anyway. The
various Linux distributions all maintain repositories for bleeding edge
versions of libraries and programs (e.g. Mandrake's "cooker", Red Hat's
"rawhide"). These packages are usually supplied customized for a
particular distribution and are seldom compatible with other
distributions. To a certain extent, this is also true of source rpm's
because the various distributions sometimes store source files in
different places. This may quickly turn into a collection of HOW-TO's for
each of the various major Linux distributions.

> Feel free also to tell me that my proposal makes no sense. Since I'm not a 
> Linux user, I may well have misunderstood everything. But watching from the 
> sidelines it sure sounds like utter chaos that one is sent to lots of 
> different repositories to find obscure files and keep trying to compile and 
> find other dependencies and ask some more questions and update the compiler 
> itself to the latest version and then.....
> 
Well, it may not be as bad as it seems. There are a handful of well known
Linux repositories. One of the most often used is rpmfind.net, but I've
discovered tuxfinder.com is better for my purposes and has always
performed wonderfully for me. Most Linux users know about these, but the
newbies will probably not. I speak from experience since I've only been
using Linux for a year, and I've found these places usually by accident.
:-)

I have all of the pieces needed for installing VPython on my RH7 system,
but have not yet done so. After I get things installed and running, I will
doucment the dependencies and put together a HOW-TO for RH7. I've not yet
learned how to create rpm packages, but I need to learn anyway.

Now, back to doing my taxes. :-)

Joe Heafner

--
My Book <http://www.willbell.com/new/fundephcomp.htm>
My Home Page <http://users.vnet.net/heafnerj/>
CVAC Home Page <http://users.vnet.net/heafnerj/cvac.html>
PLEASE -- no Microsoft attachments! I will not read them!

Re: [Visualpython-users] What is IDLE?

[Anton S. <br...@po...>]

Ari Heitner wrote:
> Compiling the Linux version is easy. You just need to
> get RPMs (from rpmfind.net) for the libraries it needs --

Tried that.  Got nowhere.

(Don't ask me for details because since then I managed to install a
broken version of rpm, and have failed so far to install a working
version.)

-- 
Anton Sherwood  --  br...@p0...  --  http://ogre.nu/
(returning to lurk mode)

[Visualpython-users] Linux users of VPython, unite!

[Bruce S. <ba...@an...>]

I'd like to propose a small Linux VPython community project. I find it 
frustrating that what look to me like the same Linux questions and problems 
keep recurring. Ari Heitner has been trying to provide some compiled 
packages, but it's clear that these often are not usable, depending on the 
Linux flavor, so there should be a high-quality document explaining how to 
do a source compile and install.

I propose that someone knowledgeable draft and post FULLY detailed, 
comprehensive instructions for how to compile and install on various common 
flavors of Linux. Other knowledgeable Linux users would critique and debate 
the document. When everyone had thrashed out agreement, we would post this 
prominently on the VPython site. The document needs in particular to state 
clearly and in detail all the dependencies, and all the places where one 
has to go to get the various pieces. And if we get that far, maybe we could 
even get all the pieces together in one place!

Feel free also to tell me that my proposal makes no sense. Since I'm not a 
Linux user, I may well have misunderstood everything. But watching from the 
sidelines it sure sounds like utter chaos that one is sent to lots of 
different repositories to find obscure files and keep trying to compile and 
find other dependencies and ask some more questions and update the compiler 
itself to the latest version and then.....

Bruce Sherwood

Re: [Visualpython-users] What is IDLE?

[Ari H. <ahe...@an...>]

On Wed, Apr 11, 2001 at 04:35:22PM -0400, heafnerj wrote:
> On Wed, 11 Apr 2001, Andrew Morrison wrote:
> 
> > Uncelievable!  I went back and re-read your other suggestions from the
> > previous thread.  After looking around on rpmfind, I could on lyfind the
> > libstdc++ for RedHat7.0.   I'm using RH6.2, but I decided to install
> > anyway.  Then, the visual python rpm actually installed cleanly!  So,
> > thanks for you help on that.
> > 
> Yippee! I'm running RH7 too and I've not yet installed VPython on my
> machine. I'm glad to hear that everything installs correctly. I was
> worried that RH7 ships with Python 1.5 rather than Python 2.0, but I guess
> Visual will work with either version.
> 
Um, that's an error on the web page.

the Visual module will indeed work with either version, in a vacuum.

the linux packages specifically install it into the python 1.5.2
site-packages directory. On linux, Visual *should* be used with python
1.5.2.




Ari

Re: [Visualpython-users] What is IDLE?

[heafnerj <hea...@vn...>]

On Wed, 11 Apr 2001, Andrew Morrison wrote:

> Uncelievable!  I went back and re-read your other suggestions from the
> previous thread.  After looking around on rpmfind, I could on lyfind the
> libstdc++ for RedHat7.0.   I'm using RH6.2, but I decided to install
> anyway.  Then, the visual python rpm actually installed cleanly!  So,
> thanks for you help on that.
> 
Yippee! I'm running RH7 too and I've not yet installed VPython on my
machine. I'm glad to hear that everything installs correctly. I was
worried that RH7 ships with Python 1.5 rather than Python 2.0, but I guess
Visual will work with either version.

Now, I guess my main concern is speed. My Linux box is a Pentium 166. 

Cheers,
Joe Heafner

--
My Book <http://www.willbell.com/new/fundephcomp.htm>
My Home Page <http://users.vnet.net/heafnerj/>
CVAC Home Page <http://users.vnet.net/heafnerj/cvac.html>
PLEASE -- no Microsoft attachments! I will not read them!

Re: [Visualpython-users] What is IDLE?

[Ari H. <ahe...@an...>]

On Wed, Apr 11, 2001 at 02:54:34PM -0500, Andrew Morrison wrote:
> 
> Uncelievable!  I went back and re-read your other suggestions from the
> previous thread.  After looking around on rpmfind, I could on lyfind the
> libstdc++ for RedHat7.0.   I'm using RH6.2, but I decided to install
> anyway.  Then, the visual python rpm actually installed cleanly!  So,
> thanks for you help on that.

:)

> 
> Now, I'm still left with my other question about IDLE.  If IDLE is just
> the programming environment, I am guessing it is not included with the
> visual python rpm.  I went to python.org and saw I could download
> IDLE.  But, if I get that installed will it automatically use the visual
> modules?

Um, that's not our IDLE, that's Guido's IDLE. You're welcome to use it -- it
will work for any python program just fine, including Visual programs. But
it will have the same problem as Guido's IDLE on Windows -- programs are run
in the same thread, so if the program crashes/hangs, so does the IDE.

Most Unix people already have their own favorite text editors
(nedit/vi/emacs/SciTE/whatever) so we don't usually hear opinions on that as
it relates to VPython.

> 
> Also, I've tried running some of the demo programs.  Here's the output:
> 
> [morris@domra Demos]$ python1.5 orbit.py 
> Visual-2000-11-26
> Traceback (innermost last):
>   File "orbit.py", line 1, in ?
>     from visual import *
>   File "/usr/lib/python1.5/site-packages/visual/__init__.py", line 12, in ?
>     from Numeric import *
> ImportError: No module named Numeric
> 
> 
> Do I need to get a Numeric module?   

Yeah, sorry.

That's a runtime dependency of the Visual module, so it doesn't yell at you
on install (it just fails to run). fun.

I think the package you want is python-numpy. I see it on rpmfind.




Good luck,


Ari