Re: Octave 2.9.12 available for ftp

[Thomas Weber]

Am Donnerstag, 24. Mai 2007 21:18 schrieb David Bateman:
> Thomas Treichl wrote:
> > Hi John,
> >
> > current CVS snapshot for octave-2.9.12 compiles fine for me on MacOSX
> > 10.4.9
> >
> > IA32, I just have 4 fails for './make check', here is the output of the 
fntests.log:
> >  >>>>> processing /Users/Thomas/Development/octave/scripts/set/ismember.m
> >
> >    ***** assert (ismember ('', {'abc', 'def'}), false);
> > !!!!! test failed
> > error: assert (ismember (, {"abc", "def"}),false) expected
> > 0
> > but got
> > [](0x0)
> > Dimensions don't match
> >    ***** fail (ismember ([], {1, 2}), 'error:.*');
> > !!!!! test failed
> > error: expected error <error:.*>
> > but got <parse error near line 183 of file ./fntests.m
> >
> >    syntax error
> >
> >  >>> ;
> >
> >      ^
> >
> >
> >    ***** fail (ismember ({[]}, {1, 2}), 'error:.*');
> > !!!!! test failed
> > error: sort: wrong type argument `cell'
>
> The errors in ismember are expected. See the various threads about them
> over the last month.

For what it's worth, I believe that one of these errors is a bug in the other 
product (either documentation or implementation). Quoting from 
        http://www.nabble.com/ismember()-differences-t3586131.html

==============================================
>> a = ismember( '', { 'abc', 'def' } )
a =
     0 
==============================================

According to 
        www.mathworks.com/access/helpdesk/help/techdoc/ref/ismember.html
"tf = ismember(A, S) returns a vector the same length as A:

Now, size('') = 0x0, but size(false) =1x1.      


About the fail () tests: shouldn't their first argument be quoted? The second 
one works for me then (i.e., it fails)

        Thomas