Re: Class related regressions in GNU Octave 4.0

[Oliver Heimlich]

On 12.04.2015 16:58, Mike Miller wrote:
> On Sun, Apr 12, 2015 at 15:31:44 +0200, Oliver Heimlich wrote:
> > I have verified that #44334, #44498, and #43925 all are regressions, which
> > have been working in 3.8.2 and are broken in 4.0.
>
> Correction: bug #43925 was reported against 3.8.2, is not a regression
> from 3.8.2 to 4.0.0.

Yes.

> > What is the desired behaviour?
> > 1) A user-defined class must override all methods together to get a
> > consistent behaviour: size, isvector, isscalar, and isequal. Any builtin
> > functions will assume that the object's size is always [1 1].
> > 2) A user-defined class must override size only. The builtin functions will
> > respect the overriden size method and produce a consistent behaviour for the
> > user (like they did in version 3.8.2).
>
> Good question, I don't know the answer. I have not seen any clear
> documentation on the intended interaction between these builtin
> functions and user-defined classes.
>
> I don't have access to Matlab, but I guess a good next step would be to
> put together some minimal @-class examples and get someone to test them
> in Matlab to see what it returns from the functions in question.

I don't have access to Matlab either, the attachment contains a minimal 
@-class example.

Output in octave-3.8.2 (with annotations):
size_ =

   1   1

isvector_ =  1
isscalar_ =  1
isequal_ =  1
size_ =

   3   1

isvector_ =  1
isscalar_ =  1  <<-- bug 43925
isequal_ =  1
size_ =

   3   2

isvector_ = 0
isscalar_ =  1  <<-- bug 43925
isequal_ =  1


Output in octave-4.0 (with annotations):
size_ =

   1   1

isvector_ =  1
isscalar_ =  1
isequal_ =  1
size_ =

   3   1

isvector_ =  1
isscalar_ =  1  <<-- bug 43925
isequal_ = 0    <<-- bug 44434
size_ =

   3   2

isvector_ =  1  <<-- bug 44498
isscalar_ =  1  <<-- bug 43925
isequal_ = 0    <<-- bug 44434

bug44498.tar.gz
Description: application/gzip