Re: octave_value_list and octave_value subsasgn

[Thomas Treichl]

John W. Eaton schrieb:
> On 16-Jul-2008, Thomas Treichl wrote:
>
> | I don't know if this is a bug or if I am doing something wrong, that's why I 
> | decided to post this email on the help list. I've attached two files and I want 
> | to get the same result from the DLD-function as I get from the m-function. If I 
> | am using the subsasgn function in C++ then the type of the octave_value is 
> | converted to double. What I would expect is that the class of the 
> | octave_value_list is still logical like in the m-file function. The attached 
> | codes are snipped out of my current implementations.
> | 
> | If someone might want to know what for I need this then the answer is that I'm 
> | currently reworking arrayfun (and maybe cellfun and structfun and spfun later) 
> | because I need even more compatibility and performance of these functions and I 
> | don't want to go through the list of all class types that are available for an 
> | octave_value in C++ but I simple want to use the "(" and "{" assignment.
>
> Then you might as well write it in the scripting language, because you
> probably won't get much speedup using subsasgn and octave_value objects.

I already did, but the DLD implementation is still much faster because I don't 
know any other way to split the input arguments and to sort the results as to 
use two for-loops (one inside the other one)...

  Thomas