[sage-devel] Re: Command line processing sage-python
I am happy to report that boost_1_35_0, pycuda 0.9 (plus my hacks) are all now working and tested against this build. Thanks to David for his help and patches. Game on. On Sep 12, 1:28 am, Simon Beaumont [EMAIL PROTECTED] wrote: Ok - got there in the end... I did a fresh build (OS X 10.5.4 intel) with 3.1.1 sources. Applied David's patches and fixed up the externs in cvxopt sources as he indicated. All built just fine. All test pass apart from trace - which is no doubt due to sage command line not being able to load readline - which is odd since it has been built ok - I guess it's the python framework that is not loading this. I'm sure there's an easy fix. Now onto patching boost and seeing if this wont all play nicely together. Be nice is this was the default (supported) build for OS X since nothing else is a starter if packages need to link against the python. I only want to run one python and I want all my python code and any packages I use to play in sage. That's the whole point for me - extensibility in all directions. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
Glad it eventually worked for you. Sorry I forgot to mention the readline bug. My brain is not big enough to remember all those steps. It does cause a doctest failure. As you probably have realized, Michael already has a patch for it in the pipeline, it will be fixed in 3.1.2 I believe. If you want sage to have a working command line, you could use the newer python spkg. D M: +61 423 535 397 W: +61 2 6125 8260 On 12/09/2008, at 19:52, Simon Beaumont [EMAIL PROTECTED] wrote: I am happy to report that boost_1_35_0, pycuda 0.9 (plus my hacks) are all now working and tested against this build. Thanks to David for his help and patches. Game on. On Sep 12, 1:28 am, Simon Beaumont [EMAIL PROTECTED] wrote: Ok - got there in the end... I did a fresh build (OS X 10.5.4 intel) with 3.1.1 sources. Applied David's patches and fixed up the externs in cvxopt sources as he indicated. All built just fine. All test pass apart from trace - which is no doubt due to sage command line not being able to load readline - which is odd since it has been built ok - I guess it's the python framework that is not loading this. I'm sure there's an easy fix. Now onto patching boost and seeing if this wont all play nicely together. Be nice is this was the default (supported) build for OS X since nothing else is a starter if packages need to link against the python. I only want to run one python and I want all my python code and any packages I use to play in sage. That's the whole point for me - extensibility in all directions. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
Thanks for that... someone might care to fix it... I think it used to work. It calls sage -python passing all the command line. More importantly what would be involved in using the system framework based python - would it just be a case of setting $PYTHON to be /usr/ bin/python? Cheers, Simon On Aug 29, 1:34 am, Carl Witty [EMAIL PROTECTED] wrote: On Aug 28, 4:34 pm, Simon Beaumont [EMAIL PROTECTED] wrote: Well any python code fails to work as expected: try: sage-python -c import sys File string, line 1 import ^ SyntaxError: invalid syntax ... -c print;print # will print two line feeds however this came up when trying to configure boost for sage, wherein configure has: $PYTHON -c import sys; print (\%d.%d\ % (sys.version_info[0], sys.version_info[1]))` sage-python is a shell script that sends its arguments to the Sage python (assuming sage is in your path). Evidently it is buggy. However, you don't want to be using sage-python anyway; instead, you want to arrange things so that $PYTHON points to $SAGE_ROOT/local/bin/ python . Carl --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
On Thu, Aug 28, 2008 at 1:19 PM, Simon Beaumont [EMAIL PROTECTED] wrote: got a rather weird one - os x 10.5.2 sage 3.1.1 (binary build) I am trying to build boost (python) for this version on os x - so I can get pycuda going against the OS X 2.0 cuda. well the python (python-sage) seems not to handle the -c command like regular python that is import fails with a syntax error and print 'foo' doesn't... Any ideas? For clarification can you post an actual log of what you input and what happens? William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
On Aug 28, 2:32 pm, William Stein [EMAIL PROTECTED] wrote: On Thu, Aug 28, 2008 at 1:19 PM, Simon Beaumont [EMAIL PROTECTED] wrote: Hi, got a rather weird one - os x 10.5.2 sage 3.1.1 (binary build) I am trying to build boost (python) for this version on os x - so I can get pycuda going against the OS X 2.0 cuda. David Philip has been playing with building PyCuDA against Sage's Python on OSX. He switched the build to use ucs2 and also builds Sage as a FrameWork. It is not 100% ready for prime time, but check out http://trac.sagemath.org/sage_trac/ticket/3924. well the python (python-sage) seems not to handle the -c command like regular python that is import fails with a syntax error and print 'foo' doesn't... Any ideas? For clarification can you post an actual log of what you input and what happens? Yeah, that would be helpful. You need boost 1.35, too, and the experimental boost package we provide is 1.34. Making one for 1.35 should be trivial. William Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
On 29/08/2008, at 7:56 AM, mabshoff wrote: David Philip has been playing with building PyCuDA against Sage's Python on OSX. No!!! I'd love to do it but I haven't got time for CUDA. I've been using boost.python, for the sake of /other/ C++ code of mine. I originally tried to sort of bodgy them together, but didn't have much luck. But I think you will pretty much need the framework build as per the trac ticket Michael linked. You then build boost python against that---but to build boost python against a non-system python on a mac, you need to patch boost build. See http://svn.boost.org/trac/boost/changeset/48322 I hope you can use CUDA with all that... thus we continue towards SageOS... The alternative I suppose is to redirect the rest of sage to use system python, which is conceivably necessary for CUDA. D == David J Philp Postdoctoral Fellow National Centre for Epidemiology and Population Health Building 62, cnr Mills Rd Eggleston Rd The Australian National University Canberra ACT 0200 Australia T: +61 2 6125 8260 F: +61 2 6125 0740 M: 0423 535 397 W: http://nceph.anu.edu.au/ CRICOS Provider #00120C --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
On Aug 28, 11:22 pm, David Philp [EMAIL PROTECTED] wrote: On 29/08/2008, at 7:56 AM, mabshoff wrote: David Philip has been playing with building PyCuDA against Sage's Python on OSX. No!!! I'd love to do it but I haven't got time for CUDA. I've been using boost.python, for the sake of /other/ C++ code of mine. I originally tried to sort of bodgy them together, but didn't have much luck. But I think you will pretty much need the framework build as per the trac ticket Michael linked. You then build boost python against that---but to build boost python against a non-system python on a mac, you need to patch boost build. Seehttp://svn.boost.org/trac/boost/changeset/48322 I hope you can use CUDA with all that... thus we continue towards SageOS... The alternative I suppose is to redirect the rest of sage to use system python, which is conceivably necessary for CUDA. D Phew! nothing simple on OS X eh? I might just get a 8800 for my linux box ;-) took me about ten minutes to do this on gentoo... boost seems a real pain on OS X - i'll try and build against the standard framework based python for a starter and take it from there. Thanks for the heads up... Simon --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
Well any python code fails to work as expected: try: sage-python -c import sys File string, line 1 import ^ SyntaxError: invalid syntax ... -c print;print # will print two line feeds however this came up when trying to configure boost for sage, wherein configure has: $PYTHON -c import sys; print (\%d.%d\ % (sys.version_info[0], sys.version_info[1]))` On Aug 28, 10:32 pm, William Stein [EMAIL PROTECTED] wrote: On Thu, Aug 28, 2008 at 1:19 PM, Simon Beaumont [EMAIL PROTECTED] wrote: got a rather weird one - os x 10.5.2 sage 3.1.1 (binary build) I am trying to build boost (python) for this version on os x - so I can get pycuda going against the OS X 2.0 cuda. well the python (python-sage) seems not to handle the -c command like regular python that is import fails with a syntax error and print 'foo' doesn't... Any ideas? For clarification can you post an actual log of what you input and what happens? William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
On 29/08/2008, at 9:29 AM, Simon Beaumont wrote: On Aug 28, 11:22 pm, David Philp [EMAIL PROTECTED] wrote: On 29/08/2008, at 7:56 AM, mabshoff wrote: David Philip has been playing with building PyCuDA against Sage's Python on OSX. No!!! I'd love to do it but I haven't got time for CUDA. I've been using boost.python, for the sake of /other/ C++ code of mine. I originally tried to sort of bodgy them together, but didn't have much luck. But I think you will pretty much need the framework build as per the trac ticket Michael linked. You then build boost python against that---but to build boost python against a non-system python on a mac, you need to patch boost build. Seehttp://svn.boost.org/trac/boost/changeset/48322 I hope you can use CUDA with all that... thus we continue towards SageOS... The alternative I suppose is to redirect the rest of sage to use system python, which is conceivably necessary for CUDA. D Phew! nothing simple on OS X eh? I might just get a 8800 for my linux box ;-) took me about ten minutes to do this on gentoo... boost seems a real pain on OS X - i'll try and build against the standard framework based python for a starter and take it from there. Well, it's all in the pipeline so the future is a bit rosier. The boost patch won't appear till 1.37 AIUI, and the sage/python/framework won't appear till Michael Abshoff is happy with it. There are also a couple of bugs in the version of cvxopt that ships with sage. Those appear to be fixed in the latest version of cvxopt, so they are also in the pipeline (assuming that sage will upgrade to cvxopt 1.0 and/or make cxvopt optional, there doesn't seem to be a lot of point bothering with a sage-specific patch to cvxopt-0.9.) But if you want my (hacked untested but compiling) cvxopt spkg, let me know. D == David J Philp Postdoctoral Fellow National Centre for Epidemiology and Population Health Building 62, cnr Mills Rd Eggleston Rd The Australian National University Canberra ACT 0200 Australia T: +61 2 6125 8260 F: +61 2 6125 0740 M: 0423 535 397 W: http://nceph.anu.edu.au/ CRICOS Provider #00120C --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
I have pycuda working on OSX 10.5.2 with the standard framework based python and nv CUDA 2.0... now for the hacks! Not at all sure about the right direction for this CUDA stuff yet. Simon On Aug 29, 12:29 am, Simon Beaumont [EMAIL PROTECTED] wrote: On Aug 28, 11:22 pm, David Philp [EMAIL PROTECTED] wrote: On 29/08/2008, at 7:56 AM, mabshoff wrote: David Philip has been playing with building PyCuDA against Sage's Python on OSX. No!!! I'd love to do it but I haven't got time for CUDA. I've been using boost.python, for the sake of /other/ C++ code of mine. I originally tried to sort of bodgy them together, but didn't have much luck. But I think you will pretty much need the framework build as per the trac ticket Michael linked. You then build boost python against that---but to build boost python against a non-system python on a mac, you need to patch boost build. Seehttp://svn.boost.org/trac/boost/changeset/48322 I hope you can use CUDA with all that... thus we continue towards SageOS... The alternative I suppose is to redirect the rest of sage to use system python, which is conceivably necessary for CUDA. D Phew! nothing simple on OS X eh? I might just get a 8800 for my linux box ;-) took me about ten minutes to do this on gentoo... boost seems a real pain on OS X - i'll try and build against the standard framework based python for a starter and take it from there. Thanks for the heads up... Simon --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
[sage-devel] Re: Command line processing sage-python
On Aug 28, 4:34 pm, Simon Beaumont [EMAIL PROTECTED] wrote: Well any python code fails to work as expected: try: sage-python -c import sys File string, line 1 import ^ SyntaxError: invalid syntax ... -c print;print # will print two line feeds however this came up when trying to configure boost for sage, wherein configure has: $PYTHON -c import sys; print (\%d.%d\ % (sys.version_info[0], sys.version_info[1]))` sage-python is a shell script that sends its arguments to the Sage python (assuming sage is in your path). Evidently it is buggy. However, you don't want to be using sage-python anyway; instead, you want to arrange things so that $PYTHON points to $SAGE_ROOT/local/bin/ python . Carl --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://www.sagemath.org -~--~~~~--~~--~--~---
