[Pkg-octave-devel] Bug#626264: Bug#626264: Bug#626264: Bug#626264: octave-miscellaneous: Please upload new upstream (1.0.11) release

[Jordi Gutiérrez Hermoso]

On 11 May 2011 15:14, Thomas Weber <tweber at debian.org> wrote:
> On Tue, May 10, 2011 at 06:24:41PM -0500, Jordi Gutiérrez Hermoso wrote:
>> On 10 May 2011 14:51, Thomas Weber <tweber at debian.org> wrote:
>> > On Tue, May 10, 2011 at 01:54:24PM +0200, Didier Raboud wrote:
>> >> Package: octave-miscellaneous
>> >> Version: 1.0.9-1+b1
>> >> Severity: minor
>> >>
>> >> Hi,
>> >>
>> >> octave-optim, in it 1.0.16-1 version, uploaded on 23. April, depends on
>> >> octave-miscellaneous >= 1.0.11, so please upload it to allow the installation
>> >> of octave-optim/unstable.
>> >
>> > From miscellaneous 1.0.11:
>> > [map]
>> >  ***** test
>> >  assert(map(@min,[1 2 3 4 5],[5 4 3 2 1]), [1 2 3 2 1])
>> >  assert(map(@min,rand(1,5),[0 0 0 0 0]), [0 0 0 0 0])
>> >  assert(map(@(x,y) (sin(x).^2 + cos(y).^2),-pi:0.5:+pi,-pi:0.5:+pi),ones(1,13))
>> > !!!!! test failed
>> > assert (map (@(x, y) (sin (x) .^ 2 + cos (y) .^ 2);
>> > , -pi:0.5:+pi, -pi:0.5:+pi),ones (1, 13)) expected
>> >   1   1   1   1   1   1   1   1   1   1   1   1   1
>> > but got
>> >   1   1   1   1   1   1   1   1   1   1   1   1   1
>> >
>> >
>> > The problem is that Octave 3.2 isn't consistent in what ones() returns:
>> > ...
>> > GNU Octave, version 3.2.4
>> > ...
>> > octave:1> typeinfo(ones(1,13))
>> > ans = range
>> > octave:2> typeinfo(ones(2,13))
>> > ans = matrix
>> >
>> > I don't have the build power on my notebook to build 3.4. Can you check wether
>> > this has changed there?
>> >
>>
>>      octave:1> typeinfo(ones(1,2))
>>      ans = range
>>      octave:2> typeinfo(ones(2,2))
>>      ans = matrix
>>      octave:3> version
>>      ans = 3.5.0+
>>
>> But I don't think this is a problem; it's a useful optimisation.
>
> I have a bit of a problem accepting
>
> octave:3> assert([1,1,1], ones(1,3))
> error: assert ([1, 1, 1],ones (1, 3)) expected
>   1   1   1
>   but got
>   1   1   1
>   Type matrix != range
>
> as not being a bug. This has nothing to do with map().

Oh, I'm sorry, as I said, the assert doesn't fail if you replace map
with arrayfun in the current dev sources possibly also in 3.4.0, in
particular, your example doesn't fail either anymore. I suppose that
has been fixed, then.

I have a feeling Octave is nearing a 3.4.1 release, just asked the
maintainers list if this is the case. In that case, when we package
that, this bug will go away?

- Jordi G. H.