call of overloaded 'fmax(double&, double&)' is ambiguous about adol-c

[svigerske]

Comment by lorenz created at 2010-09-29 07:25:20

What I forgot to mention:

• OS: Ubuntu 10.04 (Linux 2.6.32-24-generic-pae)

[svigerske]

Comment by @kkulshre created at 2010-10-01 12:54:12

We need to look into this for the long run and define these selectively for certain compilers, but for the short run you may use either one of the following two workarounds:

1. use any one of -ansi or -std=c90 switches in your compiler commands. Then gcc does not define fmin and fmax of its own.
2. remove the fmin(const double&, const double&) declaration and defintions from your copy of adouble.h

[svigerske]

Comment by @kkulshre created at 2010-10-01 15:24:22

Please try using the svn trunk revision 169 of adolc

svn co https://projects.coin-or.org/svn/ADOL-C/trunk adolc_trunk

Your little program from above is also included as adolc_trunk/ADOL-C/tests/fminmax.cpp

Hopefully this will get rid of the problem in a portable way.

[svigerske]

Comment by lorenz created at 2010-10-01 19:40:54

Okay, it's fixed in trunk, and the example compiles without any problems. Thank you.

[svigerske]

Comment by @awalther1 created at 2010-10-07 09:26:06

Resolution: fixed