Mailman 3 October 2002 - Cplusplus-sig

Latest ETA for formal Boost Python 2 release?
by Robert Oschler Oct. 10, 2002

Oct. 10, 2002

Just checking in. What's the latest ETA for formal Boost Python 2 release? thx

3 6

Re: return policies and mipspro errors
by David Abrahams Oct. 10, 2002

Oct. 10, 2002

Stefan Seefeld <stefan.seefeld(a)orthosoft.ca> writes: > David Abrahams wrote: > > > Those aren't link errors as you claimed in your post, but compile > > errors. > > yeah, you are right. But the way mipspro works makes them appear in > the link stage, as that is where (in this case) the templates are > instantiated. > > > Is your CVS state up-to-date? What I'm seeing there looks like > > a bug we fixed yesterday. Probably you're safest getting the RC_1_29_0 > > branch of Boost, as that's nearly what's going to be released in a > > couple of days. > > thanks. For now I just updated from the head branch, and it worked > nicely. > > The new tutorial is *wonderful*. Really a great help. You can thank Joel for that. > It appears you wrote it directly in html. Why don't you use docbook > so the tutorial could be generated for other media (pdf, say), too ? Not the tutorial; Joel has a funky little documentation proccessor which generates the HTML from plaintext. I am looking forward to re-doing the reference documentation in a more-flexible format. However, I'm highly tempted to go with something like reStructuredText (http://docutils.sourceforge.net/rst.html), which is like a much more powerful version of what Joel was using, instead of docbook. I can't stand <TYPING>tags</TYPING> or reading them, either. -- David Abrahams * Boost Consulting dave(a)boost-consulting.com * http://www.boost-consulting.com

2 1

Re: Just about to release
by Greg Hawkins Oct. 9, 2002

Oct. 9, 2002

> I'm not so sure. I think having 'f' as a parameter to a function > called 'f' is confusing, don't you? Or maybe we should change the > function name to 'g', *and* make your suggested fix. Oops! Yes, of course, the change should be in the other direction (make the c++ version use 'y' instead of 'f'). Having f being callable, comparable with 'foo' and the name of the function is not just confusing but impossible isn't it? BTW, a slight worry I have with the tutorial is the emphasis on (b)jam. I don't have any problems with bjam but I don't use it for building v2 extensions and it would be nice to reassure users that (apart from typing 'bjam' to build boost itself) they don't need to fool around with an alien build tool to build an extension. All they need is 1) the python headers in their include path 2) to link with python 3) $BOOST_ROOT in their include path 4) link with BOOST_ROOT/libs/python/build/bin-stage/<boost python> 5) define BOOST_PYTHON_DYNAMIC_LIB and BOOST_PYTHON_V2 (are these still necessary by the way? I haven't fooled around with my build settings for a while) (at least I think that's all you need; a quick glance through my scons scripts doesn't turn up anything else...) There's no need to patronise people but forcing jam on them on page 2 of the tutorial may be offputting... Oh, and seeing that it's almost out of the door now: congratulations Dave (and Joel, Ralf, and everyone else) on Boost.Python v2. Its a marvellous thing. best, Greg David Abrahams wrote: > "Greg Hawkins" <Greg.Hawkins(a)softwire.co.uk> writes: > > >>Sure, >> >>Should minor bits like this go to the list or to you or Joel directly? >>I guess the list is best so you don't get loads of duplicates... > > > To the list is fine, thanks. > > >>Another one, for instance: >> >>------- >> >>first code example in >>boost/libs/python/doc/tutorial/doc/basic_interface.html >> >>the 'y' doesn't correspond to the 'f' in the subsequent cpp >>example. One of them needs fixing. >> >> def f(x, f): >> if (y == 'foo'): <<< this y should be f > > > I'm not so sure. I think having 'f' as a parameter to a function > called 'f' is confusing, don't you? Or maybe we should change the > function name to 'g', *and* make your suggested fix. > > >> x[3:7] = 'bar' >> else: >> x.items += f(3, x) >> return x >> >> def getfunc(): >> return f; >> >>----------- >> >>best, Greg >> > >

3 3

howto convert a python list to a C++ matrix
by Stefan Seefeld Oct. 9, 2002

Oct. 9, 2002

hi there, I'm trying to define a matrix constructor that takes a python list. I want to be able to write > matrix = Matrix([[1,0,0,0],[0,1,0,0],[0,0,1,0],[0,0,0,1]]) where Matrix is a C++ type. Obviously I need a factory such as 'Matrix factory(object);' or similar. But how am I to extract the elements of the list(s) in boost.python style ? And how should I raise a TypeError if I don't find what I'm looking for ? Here is a first attempt: Matrix factory(python::list &l) { float matrix[16]; assert(l.length() == 4); for (size_t i = 0; i != 4; ++i) { python::list row = extract<python::list>(l[i]); assert(row.length() == 4); for (size_t j = 0; j != 4; ++j) { matrix[4*i + j] = row[j]; } } return Matrix(matrix); } But 'length()' doesn't appear to exist, and the asserts should be replaced with some (python) exceptions that indicate a type error or similar.

3 2

RE: [C++-sig] Just about to release
by Greg Hawkins Oct. 9, 2002

Oct. 9, 2002

Dave, I think there's a small glitch in the boost/libs/python/doc/tutorial/doc/call_policies.html. Informs Boost.Python that the lifetime of the argument indicated by ward (i.e. the 2nd argument: Z* z) is dependent on the lifetime of the argument indicated by custodian (i.e. the 1st argument: Z* z). should be Informs Boost.Python that the lifetime of the argument indicated by ward (i.e. the 2nd argument: Z* z) is dependent on the lifetime of the argument indicated by custodian (i.e. the 1st argument: Y& y). best, Greg -----Original Message----- From: David Abrahams [mailto:dave@boost-consulting.com] Sent: 09 October 2002 16:10 To: c++-sig(a)python.org Subject: [C++-sig] Just about to release Hi Everyone, The main trunk of Boost.Python is basically ready for release, including documentation and tutorial material. I am just going to move the latest tweaks over to the RC_1_29_0 branch and run a few tests. If you have any burning comments about the current state of the CVS, now would be a good time to make them. In particular, take a look at the documentation and the tutorial. -Dave -- David Abrahams * Boost Consulting dave(a)boost-consulting.com * http://www.boost-consulting.com _______________________________________________ C++-sig mailing list C++-sig(a)python.org http://mail.python.org/mailman/listinfo/c++-sig

3 2

Just about to release
by David Abrahams Oct. 9, 2002

Oct. 9, 2002

Hi Everyone, The main trunk of Boost.Python is basically ready for release, including documentation and tutorial material. I am just going to move the latest tweaks over to the RC_1_29_0 branch and run a few tests. If you have any burning comments about the current state of the CVS, now would be a good time to make them. In particular, take a look at the documentation and the tutorial. -Dave -- David Abrahams * Boost Consulting dave(a)boost-consulting.com * http://www.boost-consulting.com

4 10

Trivial error in tutorial.
by Prabhu Ramachandran Oct. 9, 2002

Oct. 9, 2002

hi, Got the chance to look at the latest docs. The tutorial (which is all i've had time to look at) is very nice! I think I've found a small mistake. http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/*checkout*/boost/boost/libs/… """ ... Will yield the expected result. Finally, calling calling the free function call_f with derived as argument: >>> call_f(derived()) 42 Will also yield the expected result. Here's what's happening: 1. call_f(derived()) is called in Python ... ####### # And finally at the end of the page: Calling call_f, passing in a derived object: >>> call_f(derived()) 42 """ AFAIK, *all* the call_f(derived())'s should read call_f(derived) or call_f(Derived()) Unless you are magically defining a default __call__ method when you wrap the class? I hope I'm not missing something obvious here. cheers, prabhu

3 2

RE: [C++-sig] Just about to release
by Greg Hawkins Oct. 9, 2002

Oct. 9, 2002

Sure, Should minor bits like this go to the list or to you or Joel directly? I guess the list is best so you don't get loads of duplicates... Another one, for instance: ------- first code example in boost/libs/python/doc/tutorial/doc/basic_interface.html the 'y' doesn't correspond to the 'f' in the subsequent cpp example. One of them needs fixing. def f(x, f): if (y == 'foo'): <<< this y should be f x[3:7] = 'bar' else: x.items += f(3, x) return x def getfunc(): return f; ----------- best, Greg -----Original Message----- From: David Abrahams [mailto:dave@boost-consulting.com] Sent: 09 October 2002 20:33 To: c++-sig(a)python.org Subject: Re: [C++-sig] Just about to release "Greg Hawkins" <Greg.Hawkins(a)softwire.co.uk> writes: > Dave, > > I think there's a small glitch in the > boost/libs/python/doc/tutorial/doc/call_policies.html. > > Informs Boost.Python that the lifetime of the argument indicated by > ward (i.e. the 2nd argument: Z* z) is dependent on the lifetime of the > argument indicated by custodian (i.e. the 1st argument: Z* z). > > should be > > Informs Boost.Python that the lifetime of the argument indicated by > ward (i.e. the 2nd argument: Z* z) is dependent on the lifetime of the > argument indicated by custodian (i.e. the 1st argument: Y& y). Thanks, Greg. I'm going to ask Joel to make the appropriate patches when he gets done sleeping (it's 3AM where he is). -Dave -- David Abrahams * Boost Consulting dave(a)boost-consulting.com * http://www.boost-consulting.com _______________________________________________ C++-sig mailing list C++-sig(a)python.org http://mail.python.org/mailman/listinfo/c++-sig

2 1

Re: [C++-sig] static data members (fwd) -- oops
by Francois Ostiguy Oct. 8, 2002

Oct. 8, 2002

There was I typo in my previous message. It should have read as follows: I am stumped. I cannot get the code below to compile. I get the following message from g++ 3.1.1: py-ring.cc: In function `void init_module_ring()': py-ring.cc:26: parse error before `=' token module("ring") .add ( ring_class object = class_<Ring>("Ring", python::args<>() ) .def("doTurn", &Ring::doTurn) ); ring_class.attr("nodesInitialized") = Ring::nodesInitialized; ... where Ring::nodesInitialized is a static variable Ring::doTurn is a static method Ring has both static and non-static data/function members. -Francois ---------------------------------------------------------------------------- Dr. Jean-Francois OSTIGUY voice: (630) 840-2231 Beam Physics Dept MS220 FAX: (630) 840-6039 Fermi National Accelerator Laboratory email: ostiguy(a)fnal.gov Batavia IL 60510-0500 WWW:www-ap.fnal.gov/~ostiguy _______________________________________________ C++-sig mailing list C++-sig(a)python.org http://mail.python.org/mailman/listinfo/c++-sig

2 3

static data members
by Francois Ostiguy Oct. 8, 2002

Oct. 8, 2002

Quick question - How to I expose static class data members ? Can you supply a short example ? -Francois ---------------------------------------------------------------------------- Dr. Jean-Francois OSTIGUY voice: (630) 840-2231 Beam Physics Dept MS220 FAX: (630) 840-6039 Fermi National Accelerator Laboratory email: ostiguy(a)fnal.gov Batavia IL 60510-0500 WWW:www-ap.fnal.gov/~ostiguy

2 2