Menu
▾
▴
visualpython-users
You can subscribe to this list here.
| 2000 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(4) |
Jul
(1) |
Aug
|
Sep
(15) |
Oct
(32) |
Nov
(35) |
Dec
(48) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2001 |
Jan
(46) |
Feb
(22) |
Mar
(65) |
Apr
(49) |
May
(22) |
Jun
(29) |
Jul
(51) |
Aug
(34) |
Sep
(32) |
Oct
(46) |
Nov
(30) |
Dec
(32) |
| 2002 |
Jan
(48) |
Feb
(4) |
Mar
(20) |
Apr
(28) |
May
(13) |
Jun
(34) |
Jul
(51) |
Aug
(15) |
Sep
(15) |
Oct
(35) |
Nov
(15) |
Dec
(20) |
| 2003 |
Jan
(31) |
Feb
(111) |
Mar
(41) |
Apr
(28) |
May
(36) |
Jun
(29) |
Jul
(27) |
Aug
(29) |
Sep
(47) |
Oct
(28) |
Nov
(7) |
Dec
(26) |
| 2004 |
Jan
(44) |
Feb
(9) |
Mar
(17) |
Apr
(26) |
May
(58) |
Jun
(13) |
Jul
(44) |
Aug
(64) |
Sep
(30) |
Oct
(11) |
Nov
(21) |
Dec
(28) |
| 2005 |
Jan
(29) |
Feb
(11) |
Mar
(11) |
Apr
(22) |
May
(85) |
Jun
(46) |
Jul
(17) |
Aug
(18) |
Sep
(14) |
Oct
(22) |
Nov
(1) |
Dec
(45) |
| 2006 |
Jan
(20) |
Feb
(36) |
Mar
(18) |
Apr
(24) |
May
(21) |
Jun
(48) |
Jul
(23) |
Aug
(20) |
Sep
(10) |
Oct
(41) |
Nov
(46) |
Dec
(40) |
| 2007 |
Jan
(40) |
Feb
(20) |
Mar
(13) |
Apr
(6) |
May
(24) |
Jun
(31) |
Jul
(30) |
Aug
(11) |
Sep
(11) |
Oct
(10) |
Nov
(56) |
Dec
(64) |
| 2008 |
Jan
(64) |
Feb
(22) |
Mar
(63) |
Apr
(28) |
May
(25) |
Jun
(36) |
Jul
(11) |
Aug
(9) |
Sep
(14) |
Oct
(41) |
Nov
(46) |
Dec
(130) |
| 2009 |
Jan
(95) |
Feb
(41) |
Mar
(24) |
Apr
(35) |
May
(53) |
Jun
(67) |
Jul
(48) |
Aug
(48) |
Sep
(86) |
Oct
(75) |
Nov
(64) |
Dec
(52) |
| 2010 |
Jan
(57) |
Feb
(31) |
Mar
(28) |
Apr
(40) |
May
(25) |
Jun
(42) |
Jul
(79) |
Aug
(31) |
Sep
(49) |
Oct
(66) |
Nov
(38) |
Dec
(25) |
| 2011 |
Jan
(29) |
Feb
(18) |
Mar
(44) |
Apr
(6) |
May
(28) |
Jun
(31) |
Jul
(36) |
Aug
(24) |
Sep
(30) |
Oct
(23) |
Nov
(21) |
Dec
(27) |
| 2012 |
Jan
(14) |
Feb
(11) |
Mar
(2) |
Apr
(48) |
May
(7) |
Jun
(32) |
Jul
(22) |
Aug
(25) |
Sep
(31) |
Oct
(32) |
Nov
(21) |
Dec
(17) |
| 2013 |
Jan
(44) |
Feb
(27) |
Mar
(3) |
Apr
(1) |
May
|
Jun
|
Jul
(3) |
Aug
(4) |
Sep
(1) |
Oct
(7) |
Nov
(5) |
Dec
(5) |
| 2014 |
Jan
|
Feb
|
Mar
|
Apr
(3) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(3) |
Dec
(2) |
| 2015 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
|
Dec
|
| 2017 |
Jan
(7) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
| 2019 |
Jan
|
Feb
|
Mar
(1) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: steve m. <mwi...@gm...> - 2010-02-08 17:53:35
|
Hi all, after painful trials I lastly was able to install it. It installed in non standard location so I had to add the .pth provided in /src/gtk.. So it detected. But trying to import i get exception. Anyone with experience? I have Mandriva 2010 installed Thanks ImportError: /usr/local/lib/python2.6/site-packages/cvisualmodule.so: undefined symbol: _ZNSt24__default_alloc_templateILb1ELi0EE12_S_free_listE File "/home/stefa/Documents/Career & Hobby/testvpy.py", line 1, in <module> from visual import * File "/usr/local/lib/python2.6/site-packages/visual/__init__.py", line 59, in <module> import cvisual |
|
From: Lars K. <li...@su...> - 2010-02-08 02:48:51
|
Hi, first of all: I just started to use vpython for an existing project and I am impressed how easy it was to get a good output. But sadly I encountered a problem, when trying to display the vpython window along with a pygtk window. If I just show the vpython window (by commenting out the "gtk.main()" call for pygtk), everything runs fine. I can rotate and zoom the 3d view with the mouse. On the other hand I can disable the vpython initialization and the GTK window works fine. But if both are enabled, then the two windows show up as expected, but they don't get repainted and don't respond to any input. I need to send a kill signal to cancel the program. Now I am wondering, if vpython has a problem with a concurrently running GTK main loop. But I could not find any indications for this in vpython's documentation or by searching the web. Does anybody have any ideas about this? Below you find my currently installed versions of gtk and vpython. I am using Debian testing. cheers, Lars Versions: python-visual: 5.12 python-gtk2: 2.16.0 libgtk2.0-0: 2.18.6 |
|
From: Symion <kn...@ip...> - 2010-02-06 05:08:21
|
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=windows-1252"
http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Craig Berry wrote:
<blockquote
cite="mid:8de...@ma..."
type="cite">
<pre wrap="">On Wed, Feb 3, 2010 at 14:49, Craig Berry <a class="moz-txt-link-rfc2396E" href="mailto:cd...@gm..."><cd...@gm...></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">On Wed, Feb 3, 2010 at 13:53, Bruce Sherwood <a class="moz-txt-link-rfc2396E" href="mailto:Bru...@nc..."><Bru...@nc...></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Could you describe briefly what it is that you want to do?
</pre>
</blockquote>
<pre wrap="">I want to allow the user to control the camera with middle- and
right-drag mouse gestures, but also, if they click on an object, slew
the camera to point at that object from a particular distance. I can
slew the camera's aim point by changing scene.center, but I'm having
the difficulties described previously with putting it at a particular
(apparent) distance from the selected object.
</pre>
</blockquote>
<pre wrap=""><!---->
So...anybody have a clue on how to do that elegantly?
</pre>
</blockquote>
Hi there,<br>
I don't know if this program is elegant or not but it does give you
control over Range, Center and Forward.<br>
It also enables you to defines the Range upper and lower boundaries.<br>
This requires that Userzoom be turned off and scene.range driven
directly.<br>
<br>
A Feedback loop is setup between scene.range and scene.mouse.pos that
drives the zoom effect.<br>
Careful positioning of the mouse pointer gives you more control.<br>
<br>
<br>
''Simple zoom routine GNU \xa9 Symion MMX<br>
<br>
"r" will Reset the scene range, scene.center and scene.forward<br>
<br>
Middle Mouse key and drag will Zoom<br>
Hint:<br>
Find the spot (any corner) and draw mouse with middle key pressed<br>
toward the scene center and back to the corner!<br>
<br>
Right Mouse key will Spin<br>
Left Mouse key will change scene.center<br>
<br>
"x" will exit<br>
'''<br>
<br>
from visual import *<br>
<br>
userzoom = False<br>
nexit = False<br>
scene.userzoom = userzoom<br>
scene.autoscale = False<br>
srange = 10.0<br>
scene.range = srange<br>
rangemin = 1.0<br>
rangemax = 100.0<br>
scenter = vector(scene.center)<br>
forward = vector(scene.forward)<br>
<br>
col = (0,1,0)<br>
work = list()<br>
work.append(sphere(color=col))<br>
work.append(box(pos=(1,0,0),color=(1,0,0)))<br>
scene.visible = True<br>
print __doc__<br>
while not nexit:<br>
post = scene.mouse.pos<br>
if scene.kb.keys>0:<br>
km = scene.kb.getkey()<br>
if km == 'x':<br>
nexit = True<br>
elif km == 'r':<br>
srange = 10.0<br>
scene.range = srange<br>
scenter = vector(0,0,0)<br>
scene.center = scenter<br>
scene.forward = forward<br>
elif scene.mouse.events>0:<br>
mk = scene.mouse.getevent()<br>
if mk.drop == 'middle':<br>
userzoom = False<br>
print 'Zoom Off: %s' %(srange)<br>
elif mk.drag == 'middle':<br>
userzoom = True<br>
print 'Zoom On: %s' %(srange)<br>
elif mk.press == 'left':<br>
if mk.pick:<br>
post = vector(mk.pick.pos)<br>
scene.center = scenter = post<br>
print 'Center %s' %(scenter)<br>
if userzoom:<br>
# print 'range (%0.3f,%0.3f,%0.3f) = %s'
%(post.x,post.y,post.z,mag(post))<br>
# Limit zoom! This line controls the Entire process<br>
srange = min(rangemax,max(rangemin,mag(post)))<br>
scene.range = srange<br>
<br>
for thing in work:<br>
thing.visible = False<br>
del work<br>
scene.visible = False<br>
<br>
<br>
Symion<br>
</body>
</html>
|
|
From: Kadir H. <kha...@ya...> - 2010-02-05 21:05:14
|
I have at least one program where I can run both pyOpenGL and VPython together. I mean I have two different output displays, one for pyOpenGL and one for Vpython. They both work OK. Of course this does not mean EVERY import you are doing will be working correctly, as I may not be using some of those import scripts you are using. Kadir Haldenbilen ________________________________ From: Craig Berry <cd...@gm...> To: vis...@li... Sent: Fri, February 5, 2010 7:54:37 PM Subject: [Visualpython-users] Vpython broke pyopengl Thanks, everyone who helped with my vpython questions. I've concluded that I really need to do what I'm planning directly in opengl, so I'm dropping vpython (for this project, anyway). I found that when I tried to run pyopengl scripts after insalling vpython, they didn't work; various imports couldn't be found. I ended up uninstalling Python entirely and starting over from scratch, reinstalling pyopengl but not vpython. Is this a known issue? Seems pretty nasty. -- Craig Berry - http://www.cine.net/~cberry/ "Lots of things in the universe don’t solve any problems, and nevertheless exist." -- Sean Carroll ------------------------------------------------------------------------------ The Planet: dedicated and managed hosting, cloud storage, colocation Stay online with enterprise data centers and the best network in the business Choose flexible plans and management services without long-term contracts Personal 24x7 support from experience hosting pros just a phone call away. http://p.sf.net/sfu/theplanet-com _______________________________________________ Visualpython-users mailing list Vis...@li... https://lists.sourceforge.net/lists/listinfo/visualpython-users |
|
From: Jon S. <js...@gm...> - 2010-02-05 18:39:05
|
http://igm.rit.edu/jobs [image: tree_wrapper_bg.jpg] <goog_1265393866833> Employment Opportunities with the RIT Interactive Games and Media Department<http://www.python.org/> Rochester Institute of Technology’s B. Thomas Golisano College of Computing and Information Sciences (GCCIS) invites applications and nominations for a tenure-track faculty position in the Interactive Games and Media (IGM) department in the areas of computer games and interactive entertainment. The department has three degree programs, a B.Sc. in Game Design and Development, an M.Sc. in Game Design and Development and a B.Sc. in New Media Interactive Development. The IGM program is internationally recognized as a leader in its degree areas. Information on all of these programs is available here on our website. We are particularly interested in candidates with demonstrated excellence in teaching and scholarship in one or more of the following areas: - Game Design (Graduate and Undergraduate) - 2D and 3D Animation (Undergraduate) - Introductory programming in Flash, Flex, C# and/or XNA (Undergraduate) In addition to the IGM department, the Golisano College of Computing and Information Sciences is home to the Computer Science, Software Engineering, Information Sciences and Technologies, and Networking, Security, and Systems Administration departments, as well as a Ph.D. program in Computer and Information Sciences. The college has 105 faculty and over 2400 undergraduate and 600 graduate students. More info: http://igm.rit.edu/jobs <http://igm.rit.edu/jobs> |
|
From: Craig B. <cd...@gm...> - 2010-02-05 17:55:04
|
Thanks, everyone who helped with my vpython questions. I've concluded that I really need to do what I'm planning directly in opengl, so I'm dropping vpython (for this project, anyway). I found that when I tried to run pyopengl scripts after insalling vpython, they didn't work; various imports couldn't be found. I ended up uninstalling Python entirely and starting over from scratch, reinstalling pyopengl but not vpython. Is this a known issue? Seems pretty nasty. -- Craig Berry - http://www.cine.net/~cberry/ "Lots of things in the universe don’t solve any problems, and nevertheless exist." -- Sean Carroll |
|
From: Bruce S. <Bru...@nc...> - 2010-02-05 17:20:13
|
I have no way of knowing. The only distribution with which I have personal experience is Ubuntu. Bruce Sherwood Gary Pajer wrote: > On Fri, Feb 5, 2010 at 10:40 AM, steve mtangoo <mwi...@gm... > <mailto:mwi...@gm...>> wrote: > > On Thu, Feb 4, 2010 at 10:59 PM, Bruce Sherwood > <Bru...@nc... <mailto:Bru...@nc...>> wrote: > > I've not seen that error before. What Linux distribution are you > using? > > > > Bruce, > Which distros are known to you to compile and run VPython smoothly? > (Other than Ubuntu 9.10, the special case we've discussed before.) > > -gary > |
|
From: Bruce S. <Bru...@nc...> - 2010-02-05 17:18:57
|
So the question is whether there is anyone on this list with experience with Mandriva that can help Steve? Bruce Sherwood steve mtangoo wrote: > On Thu, Feb 4, 2010 at 10:59 PM, Bruce Sherwood <Bru...@nc...> wrote: >> I've not seen that error before. What Linux distribution are you using? > Mandriva 2010 well updated to keep up with repo > Sorry for late reply |
|
From: Gary P. <gar...@gm...> - 2010-02-05 16:05:14
|
On Fri, Feb 5, 2010 at 10:40 AM, steve mtangoo <mwi...@gm...>wrote: > On Thu, Feb 4, 2010 at 10:59 PM, Bruce Sherwood <Bru...@nc...> > wrote: > > I've not seen that error before. What Linux distribution are you using? > Bruce, Which distros are known to you to compile and run VPython smoothly? (Other than Ubuntu 9.10, the special case we've discussed before.) -gary > Mandriva 2010 well updated to keep up with repo > Sorry for late reply > > > ------------------------------------------------------------------------------ > The Planet: dedicated and managed hosting, cloud storage, colocation > Stay online with enterprise data centers and the best network in the > business > Choose flexible plans and management services without long-term contracts > Personal 24x7 support from experience hosting pros just a phone call away. > http://p.sf.net/sfu/theplanet-com > _______________________________________________ > Visualpython-users mailing list > Vis...@li... > https://lists.sourceforge.net/lists/listinfo/visualpython-users > |
|
From: steve m. <mwi...@gm...> - 2010-02-05 15:40:45
|
On Thu, Feb 4, 2010 at 10:59 PM, Bruce Sherwood <Bru...@nc...> wrote: > I've not seen that error before. What Linux distribution are you using? Mandriva 2010 well updated to keep up with repo Sorry for late reply |
|
From: Craig B. <cd...@gm...> - 2010-02-04 20:06:21
|
On Wed, Feb 3, 2010 at 14:49, Craig Berry <cd...@gm...> wrote: > On Wed, Feb 3, 2010 at 13:53, Bruce Sherwood <Bru...@nc...> wrote: >> Could you describe briefly what it is that you want to do? > > I want to allow the user to control the camera with middle- and > right-drag mouse gestures, but also, if they click on an object, slew > the camera to point at that object from a particular distance. I can > slew the camera's aim point by changing scene.center, but I'm having > the difficulties described previously with putting it at a particular > (apparent) distance from the selected object. So...anybody have a clue on how to do that elegantly? -- Craig Berry - http://www.cine.net/~cberry/ "Lots of things in the universe don’t solve any problems, and nevertheless exist." -- Sean Carroll |
|
From: Bruce S. <Bru...@nc...> - 2010-02-04 19:59:35
|
I've not seen that error before. What Linux distribution are you using? I googled "libtool: compile: unable to infer tagged configuration" and the first reference http://www.mail-archive.com/li...@gn.../msg08037.html commented, "just use the latest version of automake/autoconf/libtool and the problem should magically go away". A subsidiary Google reference said, "By passing --tag=CC to libtool invocations. You can probably edit the Makefile(s), or inputs from this project, or simply rerun configure with something like ./configure LIBTOOL='/usr/bin/libtool --tag=CC'." I also found this advice: "You have to modify the Makefile and common/Makefile, add the extra parameter "--tag CC" to the libtool command, everything will be ok." Bruce Sherwood steve mtangoo wrote: > This is my another attempt to get VP installed. I have installed > almost everything I can think of. previously, VP will compile to some > stage but now make ends prematurely. Here is part of build.log > ////////////////////////////////////////////////// > packages/numpy/core/include -fpic -DPIC -ftemplate-depth-120 -DNDEBUG > -MMD -MF atomic_queue.d -MT atomic_queue.d atomic_queue.lo -c -o > atomic_queue.lo ./core/util/atomic_queue.cpp > libtool: compile: unable to infer tagged configuration > libtool: compile: specify a tag with `--tag' > > //////////////////////////////////////////////////////////// > What is wrong? > thanks > |
|
From: steve m. <mwi...@gm...> - 2010-02-04 18:43:56
|
This is my another attempt to get VP installed. I have installed almost everything I can think of. previously, VP will compile to some stage but now make ends prematurely. Here is part of build.log ////////////////////////////////////////////////// packages/numpy/core/include -fpic -DPIC -ftemplate-depth-120 -DNDEBUG -MMD -MF atomic_queue.d -MT atomic_queue.d atomic_queue.lo -c -o atomic_queue.lo ./core/util/atomic_queue.cpp libtool: compile: unable to infer tagged configuration libtool: compile: specify a tag with `--tag' //////////////////////////////////////////////////////////// What is wrong? thanks |
|
From: Craig B. <cd...@gm...> - 2010-02-03 22:50:01
|
On Wed, Feb 3, 2010 at 13:53, Bruce Sherwood <Bru...@nc...> wrote: > Could you describe briefly what it is that you want to do? I want to allow the user to control the camera with middle- and right-drag mouse gestures, but also, if they click on an object, slew the camera to point at that object from a particular distance. I can slew the camera's aim point by changing scene.center, but I'm having the difficulties described previously with putting it at a particular (apparent) distance from the selected object. -- Craig Berry - http://www.cine.net/~cberry/ "Lots of things in the universe don’t solve any problems, and nevertheless exist." -- Sean Carroll |
|
From: Bruce S. <Bru...@nc...> - 2010-02-03 21:53:43
|
Could you describe briefly what it is that you want to do? Bruce Sherwood Craig Berry wrote: > Okay, I stayed up way too late last night (er, this morning) > experimenting with vpython camera view-distance control, and have come > to the following conclusions (all relative to a scene containing a > fixed set of static objects): > > 1. There are two fundamental parameters involved: > eyeDistance = mag(scene.mouse.camera - scene.center) > range = scene.range > > 2. eyeDistance can only be changed via user zooming, using > mouse-middle dragging. Nothing the program does directly ever changes > it. > > 3. range can only be changed by the program. Nothing the user does > ever changes it. > > This seems wildly counterintuitive, and I wasted a lot of time trying > to violate point 2 or 3. But all this seems quite solid. > > So, as in your movecamera.py example, the only way to move the eye to > a known apparent viewpoint following user-induced changes to > eyeDistance is to change range proportionately, so that the two change > cancel out. > > Is that a fair summary? Or am I missing something? And if any > vpython developers are listening, why isn't there a programmatic way > to change eyeDistance directly, as for other camera parameters like > "forward"? > > |
|
From: Bruce S. <Bru...@nc...> - 2010-02-03 21:35:56
|
I apologize for the floundering that has occurred with the latest version of VPython for the Mac. Some incompatibilities were introduced into basic libraries by OSX 10.6 (Snow Leopard) and it took various tries to find a way to produce a VPython that would work on OSX 10.5 as well as 10.6. What is now available on the Mac download page at vpython.org (VPython 5.23) has been seen to work on OSX 10.5 and 10.6. It is quite possible it will also work on 10.4, but no one has reported trying that. Many many thanks to Mirko Bordignon for all the help he gave on this problem! Bruce Sherwood |
|
From: Craig B. <cd...@gm...> - 2010-02-03 18:05:47
|
Okay, I stayed up way too late last night (er, this morning)
experimenting with vpython camera view-distance control, and have come
to the following conclusions (all relative to a scene containing a
fixed set of static objects):
1. There are two fundamental parameters involved:
eyeDistance = mag(scene.mouse.camera - scene.center)
range = scene.range
2. eyeDistance can only be changed via user zooming, using
mouse-middle dragging. Nothing the program does directly ever changes
it.
3. range can only be changed by the program. Nothing the user does
ever changes it.
This seems wildly counterintuitive, and I wasted a lot of time trying
to violate point 2 or 3. But all this seems quite solid.
So, as in your movecamera.py example, the only way to move the eye to
a known apparent viewpoint following user-induced changes to
eyeDistance is to change range proportionately, so that the two change
cancel out.
Is that a fair summary? Or am I missing something? And if any
vpython developers are listening, why isn't there a programmatic way
to change eyeDistance directly, as for other camera parameters like
"forward"?
--
Craig Berry - http://www.cine.net/~cberry/
"Lots of things in the universe don’t solve any problems, and
nevertheless exist." -- Sean Carroll
|
|
From: Bruce S. <Bru...@nc...> - 2010-02-02 19:46:23
|
Basically, by setting scene.range. For some examples of camera manipulation see the example program stonehenge.py (included in the examples installed with VPython) and, in the Contributed section of vpython.org, the programs movecamera.py, look_around.py, and rotate_scene.py. Bruce Sherwood Craig Berry wrote: > Hi, all, > > In the available doc I found information about how my program can set > the camera look-position and direction, but not the zoom level > (distance from look-position). Can that be controlled as well, and if > so, how? > > |
|
From: Craig B. <cd...@gm...> - 2010-02-02 19:16:02
|
Hi, all, In the available doc I found information about how my program can set the camera look-position and direction, but not the zoom level (distance from look-position). Can that be controlled as well, and if so, how? -- Craig Berry - http://www.cine.net/~cberry/ "Lots of things in the universe don’t solve any problems, and nevertheless exist." -- Sean Carroll |
|
From: Bruce S. <Bru...@nc...> - 2010-01-30 21:13:05
|
There is now a tar file for building VPython 5.22 available on the Linux download page at vpython.org, and at sourceforge.net. There is a bug in the standard Boost libraries on Ubuntu 9.10, and to build VPython you must first install Boost 1.38 libraries from https://launchpad.net/~ajmitch/+archive/ppa <https://launchpad.net/%7Eajmitch/+archive/ppa>. Be sure to read and follow ALL of the instructions; click on everything in sight on that web page. If you're content with the older VPython 5.11 but want to run VPython on Ubuntu 9.10, just install these new 1.38 libraries and then install the Ubuntu 9.10 python-visual package. Bruce Sherwood P.S. The problem I had with linking to the Boost libraries turned out that in the build machinery the statement for linking the Boost libraries was -lboost_python -lboost_signals -lboost_thread but had to be changed to state explicitly that we want the multithreaded versions (even though those were the only ones installed, and even though this wasn't necessary in the past): -lboost_python-mt -lboost_signals-mt -lboost_thread-mt I'm hoping/assuming that this will also work okay on other Linux distributions, but I only have access to Ubuntu. |
|
From: Bruce S. <Bru...@nc...> - 2010-01-30 19:50:55
|
Thanks to Gary's comments, I now see that it is possible to run VPython on Ubuntu 9.10. Before installing the "ppa" from https://launchpad.net/~ajmitch/+archive/ppa <https://launchpad.net/%7Eajmitch/+archive/ppa> even the Ubuntu 9.10 python-visual package would not run. But after updating the Boost libraries, the python-visual package DOES run on my Ubuntu 9.10 machine (I realize that Gary has other problems). This is progress, so I'll summarize: If you want to run VPython on Ubuntu 9.10, go through the ppa stuff at the URL listed above (be sure to read and follow ALL of the arcane instructions; click on everything in sight on that web page). Use the package manager to install the standard python-visual package, which is version 5.11. You'll now be able to run VPython. This leaves open the question of how to build a new version from source should you wish to do so. As I reported, I can compile all the files but not link to the Boost libraries. The message in src/build.log is "/usr/bin/ld: cannot find -lboost_python" even though I can see libboost_python-mt-py26.a, libboost_python-mt-py26.so, and libboost_python-mt-py26.so.1.38.0 in /usr/lib, and the package manager thinks they're installed. I have no idea why the link fails. Bruce Sherwood Gary Pajer wrote: > I do have to correct my earlier comment. When I first installed the > package python-visual, it simply segfaulted on start up. When I > installed boost from ppa, it did not segfault, but it did crash. |
|
From: Gary P. <gar...@gm...> - 2010-01-30 18:11:53
|
PPA: Once you've done it three or four times you get used to it. :) I do have to correct my earlier comment. When I first installed the package python-visual, it simply segfaulted on start up. When I installed boost from ppa, it did not segfault, but it did crash. I misinterpreted the error message, though. The current error is the glibmm error mentioned here: https://bugs.launchpad.net/python/+bug/457688 (I had just glanced at the error, saw the letters "gl", and thought it was an OpenGL error.) Is that part of the boost bug? I now have two Ubuntus: my regular OS, and one running in a VM. My regular os is the one where Vpython ran, but the file system became corrupted due to the Mesa/kernel/Radeon bug. I simply don't remember if I installed boost packages, or compiled boost from source. I did not have the PPA boost installed ... I didn't know about that until yesterday. I can't tell by looking at what's installed because I did a complete reinstall of Ubuntu in the meantime, hoping that one or more of my bugs would go away. In the VM Ubuntu I installed the visual-python package, and boost from the PPA. That's the setup that allows me to import visual, but crashes with the glibmm error whenever I try to create a visual object. I have another VM, this one running Mandriva. There's no visual rpm for Mandriva that I'm aware of. I'm in the process of installing the prerequisites for visual from standard Mandriva rpm packages, and then I'll compile visual. I'll also see what progress I can make on the Ubuntu VM. I'll let you know what happens. As you very well know, all this takes time. What you may not know is I have a life. :) I'll get it done before long. -gary PS I'm beginning to wonder whether Ubuntu's policy of forcing out a new release every six months is such a good idea after all. We'll probably get visual running on 9.10 just as 10.04 comes out. On Fri, Jan 29, 2010 at 8:37 PM, Bruce Sherwood <Bru...@nc...>wrote: > Gary, I fought my way through the arcaneness of the ppa machinery (Linux > isn't for the faint of heart). I installed the new Boost 1.38 and 1.40, > and tried 1.40 first, but couldn't compile. So I removed 1.40 and found > I could compile with Boost 1.38 but not link: the message in > src/build.log is > > /usr/bin/ld: cannot find -lboost_python > > even though I can see libboost_python-mt-py26.a, > libboost_python-mt-py26.so, and libboost_python-mt-py26.so.1.38.0 in > /usr/lib, and the package manager thinks they're installed. > > Did you do some additional arcane operation to make the library > findable? What am I missing? > > Bruce Sherwood > > Gary Pajer wrote: > > Someone has made new boost packages for Ubuntu: > > https://launchpad.net/~ajmitch/+archive/ppa<https://launchpad.net/%7Eajmitch/+archive/ppa> > > <https://launchpad.net/%7Eajmitch/+archive/ppa> > > > > I installed them on a Ubuntu 9.10 virtual machine, and Vpython did not > > segfault. Python did crash with a graphics error when I attempted to > > create a sphere, but you may have different results. > > > > > ------------------------------------------------------------------------------ > The Planet: dedicated and managed hosting, cloud storage, colocation > Stay online with enterprise data centers and the best network in the > business > Choose flexible plans and management services without long-term contracts > Personal 24x7 support from experience hosting pros just a phone call away. > http://p.sf.net/sfu/theplanet-com > _______________________________________________ > Visualpython-users mailing list > Vis...@li... > https://lists.sourceforge.net/lists/listinfo/visualpython-users > |
|
From: Bruce S. <Bru...@nc...> - 2010-01-30 02:25:42
|
Gary, I fought my way through the arcaneness of the ppa machinery (Linux
isn't for the faint of heart). I installed the new Boost 1.38 and 1.40,
and tried 1.40 first, but couldn't compile. So I removed 1.40 and found
I could compile with Boost 1.38 but not link: the message in
src/build.log is
/usr/bin/ld: cannot find -lboost_python
even though I can see libboost_python-mt-py26.a,
libboost_python-mt-py26.so, and libboost_python-mt-py26.so.1.38.0 in
/usr/lib, and the package manager thinks they're installed.
Did you do some additional arcane operation to make the library
findable? What am I missing?
Bruce Sherwood
Gary Pajer wrote:
> Someone has made new boost packages for Ubuntu:
> https://launchpad.net/~ajmitch/+archive/ppa
> <https://launchpad.net/%7Eajmitch/+archive/ppa>
>
> I installed them on a Ubuntu 9.10 virtual machine, and Vpython did not
> segfault. Python did crash with a graphics error when I attempted to
> create a sphere, but you may have different results.
>
|
|
From: Gary P. <gar...@gm...> - 2010-01-29 19:19:16
|
Here are a few recent observations: Preface: I had visual running on Ubuntu 9.10. I think I compiled it myself ... I don't really remember. But I have another problem: Ubuntu 9.10 has a bug in the kernel and/or Mesa which causes irretrevable damage to the file system if you have Radeon Mobility 7500 graphics, which of course I do have. Hence I'm very reluctant to run Vpython (or Googleearth, or ... ). Notes: Someone has made new boost packages for Ubuntu: https://launchpad.net/~ajmitch/+archive/ppa<https://launchpad.net/%7Eajmitch/+archive/ppa> I installed them on a Ubuntu 9.10 virtual machine, and Vpython did not segfault. Python did crash with a graphics error when I attempted to create a sphere, but you may have different results. I have a workaround that lets me use visual on my Ubuntu machine, and I can think of another that should work for anyone. My University has put a WinXP machine on my desk. I have Vpython installed there, and I access the machine using Remote Desktop Protocol from Ubuntu. Works like a charm. Little or no degradation of performance. Another idea that should work, but I haven't tried it, is to install a distribution of linux for which Vpython *does* work on a virtual machine on the Ubuntu host. VirtualBox <http://www.virtualbox.org/> seems to work perfectly on Ubuntu, and it's (mostly) open source. E.g., install a Fedora guest on the Ubuntu host, then install Vpython on Fedora. There's likely to be a small performance hit in this scheme, but it might be tolerable. Maybe I'll give this a try myself if I have time. -gary |
|
From: Mirko B. <mir...@ie...> - 2010-01-29 17:06:59
|
Actually I don't believe this is the case, since when I compiled on 10.6 I forced 32 bit compilation, and in fact otool tells me that the libs are i386 and not x86_64 binaries, but despite that they were still not working on 10.5. It sounds more likely that the problems were due to the new dyld features of 10.6, binaries compiled there might be different to support those feats in some way. It should in principle be possible to force retrocompatibility using **the -mmacosx-version-min flag, but I wasn't able to do so (I'm not a big expert in the Jam build system that boost uses), luckily we had a 10.5 machine in the lab. Mirko Bordignon On Fri, Jan 29, 2010 at 5:58 PM, Steve Spicklemire <st...@sp...> wrote: > Sounds likely that it's the 64/32 bit difference between 10.5 and > 10.6 is the culprit? > > -steve > > On Jan 29, 2010, at 11:29 AM, Bruce Sherwood wrote: > > > Mirko Bordignon recompiled the Boost libraries on OSX 10.5 and > > verified > > that VPython then works on both OSX 10.5 and OSX 10.6. He's not sure > > whether it will work on OSX 10.4. There is now available a VPython > > 5.22 > > installer at vpython.org based on his libraries. > > > > Many thanks to Mirko for his prompt response to the problem! > > > > Bruce Sherwood > > > > > > > > ---------------------------------------------------------------------- > > -------- > > The Planet: dedicated and managed hosting, cloud storage, colocation > > Stay online with enterprise data centers and the best network in > > the business > > Choose flexible plans and management services without long-term > > contracts > > Personal 24x7 support from experience hosting pros just a phone > > call away. > > http://p.sf.net/sfu/theplanet-com > > _______________________________________________ > > Visualpython-users mailing list > > Vis...@li... > > https://lists.sourceforge.net/lists/listinfo/visualpython-users > > > > ------------------------------------------------------------------------------ > The Planet: dedicated and managed hosting, cloud storage, colocation > Stay online with enterprise data centers and the best network in the > business > Choose flexible plans and management services without long-term contracts > Personal 24x7 support from experience hosting pros just a phone call away. > http://p.sf.net/sfu/theplanet-com > _______________________________________________ > Visualpython-users mailing list > Vis...@li... > https://lists.sourceforge.net/lists/listinfo/visualpython-users > |
33 messages has been excluded from this view by a project administrator.
