More MacOSX & Python - 2.3
Well thanks again to Ralf for doing some of the MacOSX work... and figuring out how to install the latest Python as a framework. Good news is... * Extensions and libboost_python build and link. That is the ones that compile. * Of the ones that produce an extension they are loaded by the Python interpreter. * Of the ones that load, some of them seem to run to completion, and what looks like without errors. Bad news... * Only a small number of them compile. The single compile problem is this error: In file included from /net/taipan/scratch1/grafik/boost/boost/python/class.hpp:11, from args.cpp:10: /net/taipan/scratch1/grafik/boost/boost/python/bases.hpp:19: parse error before numeric constant /net/taipan/scratch1/grafik/boost/boost/python/bases.hpp:24: confused by earlier errors, bailing out * Of the ones that compile and load, the majority seem to just hang there doing nothing for some time. After some short period of time (~1 minute) I killed the Python interpreter. Don't know why this is, but it could be something as silly as it deciding to open a dialog box I can't see and is waiting for never to arrive feedback. There is one change I had to make to Boost.Python code to even get this far: Index: rvalue_from_python_data.hpp =================================================================== RCS file: /cvsroot/boost/boost/boost/python/converter/rvalue_from_python_data.hpp,v retrieving revision 1.10 diff -b -u -2 -r1.10 rvalue_from_python_data.hpp --- rvalue_from_python_data.hpp 12 Oct 2002 15:37:33 -0000 1.10 +++ rvalue_from_python_data.hpp 31 Oct 2002 04:57:49 -0000 @@ -98,5 +98,5 @@ && !defined(BOOST_PYTHON_SYNOPSIS) /* Synopsis' OpenCXX has trouble parsing this */ // This must always be a POD struct with m_data its first member. - BOOST_STATIC_ASSERT(BOOST_PYTHON_OFFSETOF(rvalue_from_python_storage<T>,stage1) == 0); + //BOOST_STATIC_ASSERT(BOOST_PYTHON_OFFSETOF(rvalue_from_python_storage<T>,stage1) == 0); # endif If that BOOST_STATIC_ASSERT is not commented out the compiler chokes with an can't-do-that/unimplemented error. I'm attachinng the test output just in case it usefull in some form. The setup for compilation is like so: PYTHON_ROOT=/Library/Frameworks/Python.framework/Versions/2.3 PYTHON_VERSION=2.3 -- grafik - Don't Assume Anything -- rrivera@acm.org - grafik@redshift-software.com -- 102708583@icq - Grafik666@AIM - Grafik@jabber.org
--- Rene Rivera <grafik666@redshift-software.com> wrote:
Well thanks again to Ralf for doing some of the MacOSX work... and figuring out how to install the latest Python as a framework.
Good news is...
* Extensions and libboost_python build and link. That is the ones that compile. * Of the ones that produce an extension they are loaded by the Python interpreter. * Of the ones that load, some of them seem to run to completion, and what looks like without errors.
This is fantastic progress! At least we are at the point where we know it works in general.
Bad news...
* Only a small number of them compile. The single compile problem is this error:
As per http://cci.lbl.gov/boost/ the Sun gcc 3.1 compilations are also failing at the moment. The current boost CVS clearly is in need of maintenance. Unfortunately I don't have time for this right now. If someone could help out getting the gcc 3.1 compilations going again this would be great.
* Of the ones that compile and load, the majority seem to just hang there doing nothing for some time. After some short period of time (~1 minute) I killed the Python interpreter. Don't know why this is, but it could be something as silly as it deciding to open a dialog box I can't see and is waiting for never to arrive feedback.
Unfortunately I am not too surprised. One of my (non-Boost.Python) regression tests revealed a fairly bad Apple c++ bug: under certain (?) circumstances objects are double-destructed! I've sent a comprehensive bug report to Apple but have not heard back for a couple of weeks. Ralf __________________________________________________ Do you Yahoo!? HotJobs - Search new jobs daily now http://hotjobs.yahoo.com/
participants (2)
-
Ralf W. Grosse-Kunstleve -
Rene Rivera