Re: classdef problems

[John W. Eaton]

On 09/12/2013 07:05 AM, Richard Crozier wrote:
> On 12/09/2013 11:43, Richard Crozier wrote:
> > Quoting Richard Crozier <address@hidden> on Thu, 12 Sep 2013
> > 11:26:12 +0100:
> >
> > > I'm using 498b2dd1bd56 (classdef).
>
> Ok, apologies, I've realised my errors where largely due to using
> addpath(genpath('.')) when testing rather than the full path name. I
> still think there may be a bug in there somewhere though, would it be
> possible for octave to resolve paths to overcome the ingenuity of fools
> such as myself?

What did addpath (genpath ('.')) do incorrectly?  Did it add + and @ 
directories to the path?  Should it not do that?  If not, is it genpath 
or addpath that should be fixed?  Or should it just warn you?  I can see 
arguments for all possibilities, but probably the only thing that 
matters is WDMD, because if we don't do what Matlab does, we will likely 
see bug reports about the differences.

jwe