- GiNaC-list - ginac.de

Cygwin compilation problem
by Giuseppe Brandino 06 Nov '08

06 Nov '08

Hello, i installed both cln 1.2.2 and ginac 1.4.3 on cygwin. The compilation, check and installation of the packages are fine. However when i try to compile an easy tutorial program like #include <ginac/ginac.h> using namespace std; using namespace GiNaC; int main() { symbol v("v"), c("c"); ex gamma = 1/sqrt(1 - pow(v/c,2)); ex mass_nonrel = gamma.series(v==0, 10); cout << "the relativistic mass increase with v is " << endl << mass_nonrel << endl; cout << "the inverse square of this series is " << endl << pow(mass_nonrel,-2).series(v==0, 10) << endl; } i get TONS of error like /usr/local/lib/libginac.a(inifcns_nstdsums.o): In function `_Z41__static_initialization_and_destruction_0ii': /usr/local/include/cln/univpoly.h:454: undefined reference to `global constructors keyed to cl_module__cl_UP_no_ring__firstglobalfun' /usr/local/include/cln/univpoly.h:357: undefined reference to `global constructors keyed to cl_module__cl_UP__firstglobalfun' /usr/local/lib/libginac.a(inifcns_nstdsums.o): In function `_Z41__static_initialization_and_destruction_0ii': /usr/local/include/cln/SV_ringelt.h:47: undefined reference to `global constructors keyed to cl_module__cl_SV_ringelt__firstglobalfun' /usr/local/lib/libginac.a(inifcns_nstdsums.o): In function `_Z41__static_initialization_and_destruction_0ii': /usr/local/include/cln/SV_number.h:46: undefined reference to `global constructors keyed to cl_module__cl_SV_number__firstglobalfun' /usr/local/lib/libginac.a(inifcns_nstdsums.o): In function `_Z41__static_initialization_and_destruction_0ii': /usr/local/include/cln/GV_integer.h:57: undefined reference to `global constructors keyed to cl_module__cl_GV_I__firstglobalfun' /usr/local/lib/libginac.a(inifcns_nstdsums.o): In function `_ZNK5GiNaC9containerISt4listE9imag_partEv': /home/pinupower/ginac-1.4.3/ginac/container.h:(.text$_ZN3cln4cl_ND1Ev[cln::cl_N::~cl_N()]+0x24): undefined reference to `cln::cl_free_heap_object(cln::cl_heap*)' /home/pinupower/ginac-1.4.3/ginac/container.h:(.text$_ZNSt6vectorIN3cln4cl_NESaIS1_EE13_M_insert_auxEN9__gnu_cxx17__normal_iteratorIPS1_S3_EERKS1_[std::vector<cln::cl_N, std::allocator<cln::cl_N> >::_M_insert_aux(__gnu_cxx::__normal_iterator<cln::cl_N*, std::vector<cln::cl_N, std::allocator<cln::cl_N> > >, cln::cl_N const&)]+0x102): undefined reference to `cln::cl_free_heap_object(cln::cl_heap*)' /home/pinupower/ginac-1.4.3/ginac/container.h:(.text$_ZNSt6vectorIN3cln4cl_NESaIS1_EE13_M_insert_auxEN9__gnu_cxx17__normal_iteratorIPS1_S3_EERKS1_[std::vector<cln::cl_N, std::allocator<cln::cl_N> >::_M_insert_aux(__gnu_cxx::__normal_iterator<cln::cl_N*, std::vector<cln::cl_N, std::allocator<cln::cl_N> > >, cln::cl_N const&)]+0x2b4): undefined reference to `cln::cl_free_heap_object(cln::cl_heap*)' /home/pinupower/ginac-1.4.3/ginac/container.h:(.text$_ZNSt6vectorIN3cln4cl_NESaIS1_EE14_M_fill_insertEN9__gnu_cxx17__normal_iteratorIPS1_S3_EEjRKS1_[std::vector<cln::cl_N, std::allocator<cln::cl_N> >::_M_fill_insert(__gnu_cxx::__normal_iterator<cln::cl_N*, std::vector<cln::cl_N, std::allocator<cln::cl_N> > >, unsigned int, cln::cl_N const&)]+0x16b): undefined reference to `cln::cl_free_heap_object(cln::cl_heap*)' /home/pinupower/ginac-1.4.3/ginac/container.h:(.text$_ZNSt6vectorIN3cln4cl_NESaIS1_EE14_M_fill_insertEN9__gnu_cxx17__normal_iteratorIPS1_S3_EEjRKS1_[std::vector<cln::cl_N, std::allocator<cln::cl_N> >::_M_fill_insert(__gnu_cxx::__normal_iterator<cln::cl_N*, std::vector<cln::cl_N, std::allocator<cln::cl_N> > >, unsigned int, cln::cl_N const&)]+0x3cd): undefined reference to `cln::cl_free_heap_object(cln::cl_heap*)' /usr/local/lib/libginac.a(symmetry.o):/usr/local/include/cln/object.h:212: more undefined references to `cln::cl_free_heap_object(cln::cl_heap*)' follow collect2: ld returned 1 exit status and this is just a part of it. I compile using g++ prog.c -L/usr/local/lib -lcln -lginac -o prog.x Can somebody help me? thanks Giuseppe Brandino ---------------------------------------------------------------- SISSA Webmail https://webmail.sissa.it/ Powered by Horde http://www.horde.org/

1 0

separating expressions into parts
by Krishna Mohan Gundu 02 Nov '08

02 Nov '08

Hi, I would like to represent an expression as a product of a scalar expression and a vector expression. This would enable me to extract these parts of the expression trivially instead of scanning the prod sequence (and hence is more optimal). However I am not clear how this can be done in ginac. For example, when multiplying a scalar expression with a rhs expression, I would like to absorb the scalar expression into the scalar expression of the rhs. This requires modifying operator*(ex,ex). Is there a better way to do it? I would appreciate any help. example use cases: symbol s1, s2, s3; // scalars vector v1, v2, v3; // vectors myex e1(s1, (s2*v3 + s1*v2)); // initialized to: e1.scalar = s1; e1.vector = (s2*v3 + s3*v2) myex e2 = s3 * e1; // should produce: e2.scalar = s3 * s1; e2.vector = (s2*v3 + s3*v2) myex e3 = cross_prod(e, v1); // should produce: e3.scalar = s1; e3.vector = cross_prod(s2*v3 + s3*v2, v1); thanks, Krishna.

1 0

expand conundrum
by Krishna Mohan Gundu 02 Nov '08

02 Nov '08

Hi, I am unable to understand why the following expression is not fully expanded (I manually split it at key points to go easy on eyes). I turn to you for help. I failed to isolate it to a simpler problem. 16*(1.0)*z1*(dz2 . [1, 0])*((1.0)*p2[1]^(-1)*z1*dz2[0]+(1.0)*p2[1]^(-1)*z2*dz1[0]) -(631.6546816697189115)*p1[2]*(1.0)^2*z1^2*z2^2 +16*(1.0)*z2*((1.0)*p2[0]^(-1)*z1*dz2[1]+(1.0)*p2[0]^(-1)*z2*dz1[1])*(dz1 . [0, 1]) +16*(1.0)*z2*((1.0)*p2[1]^(-1)*z1*dz2[0]+(1.0)*p2[1]^(-1)*z2*dz1[0])*(dz1 . [1, 0]) +16*(1.0)*z1*(dz2 . [0, 1])*((1.0)*p2[0]^(-1)*z1*dz2[1]+(1.0)*p2[0]^(-1)*z2*dz1[1]) legend: p1,p2,dz1,dz2,dz3 are vectors, z1, z2, z3 are scalars p2[0],p2[1] etc are components [1,0], [0,1] etc are vectors too (dz1 . dz2) etc are inner products each of the terms that starts as 16*(1.0) ... has an unexpanded part of the from ... * (... + ...) I can consider many possible reasons but could not come up with an answer 1) The part is actually expanded but somewhere it is being gathered again. Not likely, as my code does not override the gather function. 2) GiNaC thinks it cannot be expanded for some reason a) GiNaC refuses to expand due to restrictions on the depth of the tree? (not likely, the depth is too small) b) GiNaC refuses to expand cause it thinks the product may not be distributive? (likely, but i dont know) c) GiNaC thinks its an atom but the output fools me into believing it is not an atom? I would appreciate if any of you can point out the problem. Meanwhile I will keep digging at it. thanks, Krishna.

1 2

extending the symbol
by Krishna Mohan Gundu 27 Oct '08

27 Oct '08

Hi, I am adding a light weight support for vector algebra in ginac. I am aware that it can be done through Clifford algebra, but the output is almost incomprehensible to debug. My problem is this. When a request like inner_product(lhs, rhs) is made, where one of lhs and rhs is not a vector, I would like to replace the inner_product with a regular product. How can I achieve this? I am planning to do it as follows 1) extend a symbol to a vector 2) determine if both lhs and rhs have a hanging vector by traversing the expressions in lhs and rhs and by testing for is_a<vector> Is there a better way to test if a symbol is a vector? If not what is the simplest way to extend symbol? thanks, Krishna.

4 10

a few simple useful features
by Krishna Mohan Gundu 26 Oct '08

26 Oct '08

Hello All, 1) I find something like GINAC_DECLARE_REGISTERED_CLASS_INLINE to be very useful to define virtual member functions like archive, unarchive etc. inside the class definition. (This macro should not declare these functions, ofcourse) 2) Should numeric and ex have constructors that take std::complex<double> to initialize a complex number? This makes integration of GiNaC into other libraries simpler. cheers, Krishna.

1 0

GiNaC Licensing
by Graham Hemingway 22 Oct '08

22 Oct '08

Hello, I am working on an open source 3D solid modeling kernel ( http://wildcat-cad.blogger.com and http://wildcat-cad.googlecode.com) It allows users to sketch 2D profiles and extrude them into 3D to form mechanical parts. The 2D sketches are constrained with dimensions (lengths, angles, etc.) and these constraints naturally form a system of equations. An unrelated project (http://code.google.com/p/psketcher) has developed a very nice solver for these systems of equations. The psketcher project uses GiNaC as its mathematical engine. I would like to incorporate the psketcher solver into my kernel, but there may be a conflict in licenses. My project is released under the BSD license, while I understand GiNaC to be under the GPL. Is there any way to reconcile this problem other than changing to a different license? I would love to use GiNaC for a couple of other purposes within my project also. Any help would be appreciated. Cheers, Graham

4 7

spcializing print function
by nomad 04 Oct '08

04 Oct '08

Hello, I am trying to define a new algebraic class the easy way. The documentation is a bit confusing in this regard. It mentions typedef structure<sprod_s> sprod; and then proceeds with void sprod::print(const print_context & c, unsigned level) const { ... } How is this possible as structure<> already defines print() member function? Can someone give more specific details? thanks, Krishna. -- View this message in context: http://www.nabble.com/spcializing-print-function-tp19809899p19809899.html Sent from the Ginac - General mailing list archive at Nabble.com.

2 1

multiplying a matrix with 0 makes it become a scalar 0
by dummy-address＠web.de 21 Sep '08

21 Sep '08

Hi! symbol a("a"), b("b"), c("c"); matrix blah(3,1); blah = a,b,c; cout << "blah: " << blah << endl; //output: // blah: [[a],[b],[c]] cout << "3*blah: " << 3*blah << endl; //output: // 3*blah: 3*[[a],[b],[c]] cout << "(3*blah).evalm(): " << (3*blah).evalm() << endl; // output: // [[3*a], [3*b], [3*c]] cout << "0*blah: " << 0*blah << endl; // output: // 0 cout << "(0*blah).evalm(): " << (0*blah).evalm() << endl; // output: // 0 I don't like this behaviour. Is it supposed to be that way or is this a bug? Is there a nice way to prevent a matrix from losing its row and column information? Regards Al Capone

1 0

solving differential equatrions symbolically or numerically with C++
by Alessandro Bernardini 01 Sep '08

01 Sep '08

Hello, I am looking for a C++ library for doing symbolic or numeric computations within C++ programs. I would like a class that provides usefull tools for scientific computing (symbolic and numeric computations). I woluld like tools like in Mathematica (or Maple), but in a C++ class library !! Of particular importance is the solution of differential equations (symbolically or numerically; for ODEs or PDEs), so as variable precision arithmetic. Of course symbolic sympflification is very important. Is GiNaC the right tool for me, for doing symbolic computing and solving differential equations ? Or SymbolicC++ ? Can this libraries be an alternative to the functionality of mathematica (when used in C++ programs of course ... ) ? Thank you very very much for a short answer ... Alex.

2 1

FYI: CLN fork
by Alexei Sheplyakov 27 Aug '08

27 Aug '08

Hello! CLN is a great library, but it has a number of problems. 1. CLN can be built only with GNU C++ compiler. Even worse -- some bits of CLN (in particular, include/cln/modules.h) work only with specific versions of GCC, so CLN can break at any moment (this actually happens with every new [major] release of GCC). 2. The code is a bit obfuscated. 3. The build scripts are extremely user- and hacker-hostile. Since CLN developers keep [silently] refusing patches addressing these issues, I've decided to fork CLN. The code can be obtained from git repository: git://ffmssmsc.jinr.ru:443/varg/cln.git To view the repository on-line point your browser to http://ffmssmsc.jinr.ru/~varg/git/gitweb.cgi?p=varg/cln.git Best regards, Alexei -- All science is either physics or stamp collecting.

2 2