Hi,
the following patch applied to fsolve.cc (as distributed in
octave-forge-2003.06.02.tar.gz) makes it possible to compile it with
gcc-3.3 (tested on x86 under Debian GNU/Linx testing).
*** fsolve.cc 2003-09-09 12:02:11.000000000 +0200
--- fsolve-patched.cc 2003-09-09 12:04:32.000000000 +0200
*************** are passed directly to @var{fcn}.\n\
*** 230,236 ****
}
typedef void (NLEqn_options::*d_set_opt_mf) (double);
! typedef double (NLEqn_options::*d_get_opt_mf) (void);
#define MAX_TOKENS 1
--- 230,236 ----
}
typedef void (NLEqn_options::*d_set_opt_mf) (double);
! typedef double (NLEqn_options::*d_get_opt_mf) (void) const;
#define MAX_TOKENS 1
Additional note:
My personal C++-knowledge tends to zero, so if there are problems with
this patch, I won't be able to help. The change (it's only one word) was
done by someone with some experience in C++. He gave me permission to
submit this patch.
I have tested the compiled oct-file, it seems to give correct results
(read: the same as fsolve-function distributed with Octave).
Regards,
Thomas
|
