- GiNaC-list - ginac.de

Namespace error with is_ex_the_function
by Jacob Garber 20 Oct '15

20 Oct '15

Hello, The following program #include <iostream> #include <ginac/ginac.h> int main() { GiNaC::symbol x("x"); std::cout << GiNaC::is_ex_the_function(x, GiNaC::sin) << std::endl; } gives me a compilation error: bug.cpp:6:24: error: unexpected parenthesis after '::' std::cout << GiNaC::is_ex_the_function(x, GiNaC::sin) << std::endl; ^ /usr/local/include/ginac/function.h:767:43: note: expanded from macro 'is_ex_the_function' #define is_ex_the_function(OBJ, FUNCNAME) (GiNaC::is_the_function<FUNCNAME##_SERIAL>(OBJ)) ^ bug.cpp:6:17: error: no member named 'GiNaC' in namespace 'GiNaC'; did you mean simply 'GiNaC'? std::cout << GiNaC::is_ex_the_function(x, GiNaC::sin) << std::endl; ^ /usr/local/include/ginac/parser.h:31:11: note: 'GiNaC' declared here namespace GiNaC { ^ bug.cpp:6:24: error: expected ';' after expression std::cout << GiNaC::is_ex_the_function(x, GiNaC::sin) << std::endl; ^ /usr/local/include/ginac/function.h:767:90: note: expanded from macro 'is_ex_the_function' #define is_ex_the_function(OBJ, FUNCNAME) (GiNaC::is_the_function<FUNCNAME##_SERIAL>(OBJ)) ^ bug.cpp:6:24: error: expected expression /usr/local/include/ginac/function.h:767:90: note: expanded from macro 'is_ex_the_function' #define is_ex_the_function(OBJ, FUNCNAME) (GiNaC::is_the_function<FUNCNAME##_SERIAL>(OBJ)) ^ 4 errors generated. This is with GiNaC 1.6.5 on Mac OS 10.11, with compiler Apple version 7.0.0. It seems there is an issue with the namespace. Removing `GiNaC::` from `is_ex_the_function` compiles fine. Anyone else have this? Cheers, Jacob Garber

2 1

Linking dl libraries is missing in ginac.pc
by Isuru Fernando 17 Sep '15

17 Sep '15

Hi, 5745deb8 introduced linking with dl libraries, but it was not added to the .pc files. Therefore list of link flags given by pkg-config is incomplete and fails if `-ldl` is not added manually. I have attached a patch which uses CMAKE_DL_LIBS variable ( http://www.cmake.org/cmake/help/cmake2.6docs.html#variable:CMAKE_DL_LIBS) Thanks, Isuru Fernando

1 0

Typo in the documentation
by Jean-Michaël Celerier 01 Aug '15

01 Aug '15

Hello, In the html doc, section 5.15.2 I think that : { parser reader; ex e = reader("2*x+sin(y)"); symtab table = reader.get_syms(); symbol x = reader["x"]; symbol y = reader["y"]; } the two last lines should be symbol x = table["x"]; symbol y = table["y"]; Best regards Jean-Michaël Celerier

2 1

Release 1.6.5
by Richard B. Kreckel 27 Jul '15

27 Jul '15

Hi, A new GiNaC release is out and available. These are the changes: * Fix pow(+(...),2).expand() where the sum contained roots of expressions. * Speed-up pow(+(...),n).expand() where n > 3. * Fix a crash in PolyLogs of rational numbers. Happy hacking -richy. -- Richard B. Kreckel <http://in.terlu.de/~kreckel/>

1 0

Goncharov polylog. leads to Segmentation Fault
by Alexander Hasselhuhn 27 Jul '15

27 Jul '15

Dear all, when trying to evaluate G(I,sqrt(2)) I get "Segmentation fault", both in ginsh as well as in a program: --- begin code --- #include <ginac/ginac.h> using namespace GiNaC; int main () { ex two = 2; ex res = G (lst (I), sqrt(two)); return 0; } --- end code --- Not sure whether this is a different issue, but if instead use try to evaluate G(I,1/sqrt(2)), I get the error-message from cln: --- begin output --- terminate called after throwing an instance of 'cln::notreached_exception' what(): Internal error: statement in file float/conv/cl_F_from_RA.cc, line 27 has been reached!! Please send the authors of the program a description how you produced this error! Aborted --- end output --- I'm using GiNaC v1.6.4 with cln v1.3.3 on OpenSuse 13.2. I compiled GiNaC with gcc v4.8.3. Does this happen on other machines also? Yours, Alex

2 2

Ginac not collecting terms when multiplying powers (bug?)
by Jonathan Mills 20 Jul '15

20 Jul '15

Hello Consider this short programme: #include <iostream> #include <ginac/ginac.h> using namespace std; using namespace GiNaC; int main() { possymbol x("x"); cout << pow(4,x + 1) * pow(4,-1) << endl; } The result produced is: 1/4*4^(1+x), the expected output is 4^x. I can see that the documentation says these types of situations will only be handled if it is "safe", but having specified x as a positive value, I'm confused why this expression is not considered safe? All the best Jonathan

2 1

ginsh - p=1-norm of vector possible?
by Michael Lenz 07 Jul '15

07 Jul '15

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Hello everybody, for an ongoing research project I need to (auto-)generate an equation for a ILP that involves some fiddling with matrices, vectors and scalars, and finally taking the "p=1-norm" (or "1-norm", "sum-norm", ...) in order to retrieve my desired equation. Is there any chance I can trick ginsh into performing this operation or do I have to implement it in C++? Yours, Michael Lenz PS: I'd like to use ginsh in order to have my project's workflow as modular as possible -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAEBCAAGBQJVm5lYAAoJEMIC8hvSujFKvhgH/0prTe2zqjHieyaWM/ECVXBT leEGNVm9LuvS0f1tKDH40S6zlDzUU5qqwyjoMn+EbNzksCDixFiraqn6wp2fBRPI /i8qbsnGzh1QG6uDMkF3mxbM2P8Jor6tTDLlgzDgM9QU/pAKnthbmryrDXjt3bzR /2SfnhEdxMqzypbotCVGKuSu0Fdu+g3zWyp6yggwvMAo1LWtzW9tdO64v0+t+RpL jZpBmTtIei16RvBUxQF307GSybFlCX9kqWXPjATwixloPkyo4yJ3Dz/1uzos49OL JRUuYgP7gxCZrEDuUiOKOLKM3GFu6Z5MobcU5ReqVuYaXsCpxvwZkqF3vl5m/To= =eNjo -----END PGP SIGNATURE-----

1 0

Release 1.6.4
by Richard B. Kreckel 08 May '15

08 May '15

Hi, A new GiNaC release is out and available. These are the changes: * Fixed pow(+(...),2).expand() where the sum contained roots of rationals. * Various improvements for info(info_flags::nonnegative). * Make power::expand() (x*p)^c -> x^c * p^c, if p>0. * Fixed normal() where nested functions occur multiple times. * Add optional explicit derivatives to functions and use it in abs(), imag_part(), real_part(). * Evaluate abs(z)^(2n) -> z^n * conjugate(z)^n. Happy hacking -richy. -- Richard B. Kreckel <http://in.terlu.de/~kreckel/>

1 0

Bug in expand
by Isuru Fernando 07 May '15

07 May '15

Hi, I modified the first benchmark at http://www.ginac.de/csSC-0004015.pdf to, e = expanded value of (a0 + a1 + sqrt(3)*a2 + sqrt(3)*a3 + ... + sqrt(3)*a(n-1))^2 in e, substitute a0 with -sqrt(3)*a2 - sqrt(3)*a3 - ... - sqrt(3)*a(n-1) expand e again to obtain a1^2 For n = 4, I get, -3*a3^2+2*3*a2^2+a1^2-6*a2^2+3*a3^2 Or 3*a3^2-3*a2^2-3*a3^2+a1^2+3*a2^2 which is true, but not the expression I expected. Why are the terms 3*a2^2 and -3*a2^2, not canceled out immediately? Also a term like 2*3*a2^2 means there is some bug right? I am using ginac-1.6.3 on Ubuntu 14.10 with gcc-4.9.1. I am attaching the sample program. Isuru Fernando

2 1

compile_ex with more than two symbols in the expression
by William Silva 01 May '15

01 May '15

Dear all, Is there a way to use "compile_ex" with more than 2 free variables? I am generating several expressions using .diff() and would like to compile them into C functions and attach function pointers. Using the .subs() method is too slow for my code. Unfortunately my expressions contain more than just two symbolic variables. Ideally I would love to have a FUNCP_nP typedef that can accept a list or array of n symbolic variables and use a compile_ex function that accepts an lst of the symbols. I noticed that the FUNCP_CUBA typedef sort of does this, but I am not sure how to access the compiled function. Any help or pointers would be greatly appreciated! Best, Will S.

2 2