March 2007 - GiNaC-list

Problem: Floating point overflow
by Alberto Orlandi 31 Mar '07

31 Mar '07

Hi, I have a question about error handling. I have a problem with the "Floating point overflow". May be is better if I give an example: #include<iostream> #include<fstream> #include<string> #include<ginac/ginac.h> int main(void) { try{ GiNaC::realsymbol x("x"); GiNaC::ex pp("sin(sin(sin(sin(log(-0.760)))))",GiNaC::lst(x)); cout << pp << endl; }catch(exception &p){ cout << p.what() <<endl; } cout << "With or without exceptions\n" <<endl; } (If you are thinking where I found that absurd function...well, this is another story) When run this program it stops and gives : "Floating point overflow". This is related to the fact that Ginac tries to evaluate the numeric expression. So I have two questions: 1) Why does it not throw and exception? 2) Is it possible to stop ginac tries to evaluate the numeric expressions? Thanks in advance, Alberto

5 5

FORTRAN output
by Warren Weckesser 28 Mar '07

28 Mar '07

I am working on a program that uses ginac to generate source code. One part of the program generates fortran-90 code. Currently I am experimenting with using the csrc output format, but with the print function of the power class redefined as in section 6.3.1 of the online tutorial. Is there a better way? Has anyone implemented a full-blown fortran output format? Best regards, Warren

2 1

Bug with evalm()?
by Martin Sandve Alnæs 26 Mar '07

26 Mar '07

I use ginac 1.3.4. I believe there is a bug in the implementation of evalm(). Its behaviour should be to do nothing if it is not applicable to the ex it is called on, right? The minimal example below throws an exception: #include <ginac/ginac.h> using namespace std; using namespace GiNaC; int main(int argc, char **argv) { ex x = symbol("x"); ex e = sin(x); ex s = e.series(x == 0, 3); cout << "This will crash:" << endl; ex m = s.evalm(); return 0; } The error is: terminate called after throwing an instance of 'std::range_error' what(): basic::let_op(): pseries has no operands Aborted (core dumped) The call stack from gdb is: (gdb) where #0 0xffffe410 in __kernel_vsyscall () #1 0xb7b12770 in raise () from /lib/tls/i686/cmov/libc.so.6 #2 0xb7b13ef3 in abort () from /lib/tls/i686/cmov/libc.so.6 #3 0xb7d02520 in __gnu_cxx::__verbose_terminate_handler () from /usr/lib/libstdc++.so.6 #4 0xb7cfff55 in std::set_unexpected () from /usr/lib/libstdc++.so.6 #5 0xb7cfff92 in std::terminate () from /usr/lib/libstdc++.so.6 #6 0xb7d000ca in __cxa_throw () from /usr/lib/libstdc++.so.6 #7 0xb7d9a4a6 in GiNaC::basic::let_op (this=0x80597d8, i=1) at basic.cpp:256 #8 0xb7d96dfd in GiNaC::basic::map (this=0x80595e0, f=@0xb7f8e824) at basic.cpp:316 #9 0xb7d962a9 in GiNaC::basic::evalm (this=0x80595e0) at basic.cpp:482 #10 0x0804b47a in GiNaC::ex::evalm (this=0xbfd820d4) at /usr/include/ginac/ex.h:118 #11 0x0804b214 in main () at test.cpp:12 Of course, it doesn't make much sense to call evalm() on this expression, but it is a problem in an application where we want to call evalm() "just in case" on all ex objects that pass a certain point in the application. Thus a "do nothing" behaviour for evalm() is of great importance. martin

6 20

Bug with matrix subtraction
by Martin Sandve Alnæs 26 Mar '07

26 Mar '07

If I subtract two equal matrices, I get the scalar 0. This messes up later calculations, since a scalar doesn't have op(i), transpose(), etc. If I do a+(-b) instead of a-b, I get a zero matrix as wanted. Note that the code is in Python using the swiginac bindings, which now calls evalm() automatically between operations. In [32]: m = matrix(2,2,[x,0,0,0]) In [33]: m-m Out[33]: 0 In [34]: m+(-m) Out[34]: [[0,0],[0,0]] So we have the silly relation that "a-b != a+(-b)", which can only be called a bug. -- --- Martin Sandve Alnæs PhD student, Cardiac Computations Simula Research Laboratory, Norway

3 4