Re: assert.m and sparse.tst

[Rik]

On 12/03/2013 07:05 AM, address@hidden wrote:
> Message: 3
> Date: Tue, 3 Dec 2013 10:50:28 +0000
> From: Michael Goffioul <address@hidden>
> To: Rik <address@hidden>
> Cc: Octave Maintainers List <address@hidden>
> Subject: Re: assert.m and sparse.tst
> Message-ID:
>       <address@hidden>
> Content-Type: text/plain; charset="iso-8859-1"
>
> On Mon, Dec 2, 2013 at 8:47 PM, Rik <address@hidden> wrote:
>
>> > On 11/28/2013 06:14 PM, address@hidden wrote:
>>> > > Message: 4
>>> > > Date: Thu, 28 Nov 2013 18:59:33 -0500
>>> > > From: Michael Goffioul <address@hidden>
>>> > > To: Octave Maintainers List <address@hidden>
>>> > > Subject: assert.m and sparse.tst
>>> > > Message-ID:
>>> > >       <
>> > address@hidden>
>>> > > Content-Type: text/plain; charset="iso-8859-1"
>>> > >
>>> > > (Maybe this problem is related to the other assert behavior reported in
>> > my
>>> > > other mail. But as the way to reproduce it is quite different, I put it
>>> > > into another mail.)
>>> > >
>>> > > Another strange assert behavior is that after running the tests in
>>> > > sparse.tst, I get the same behavior as above: assert is not functional
>>> > > until I do "clear assert", at which point it becomes functional again.
>> > Can
>>> > > anybody reproduce that under MinGW build? I'm using a MSVC build based
>>> > > on afeb233ea443.
>> > 12/2/13
>> >
>> > Michael,
>> >
>> > This works for me on MinGW using a cross-built binary from today (12/2/13).
>> >
> If you don't experience bug #40743 while running sparse.tst, then you won't
> have the issue. I believe there might be a difference on my MSVC-built
> octave where some results are NaN instead of Inf, triggering the bugs.

For reference, the code sample in bug #40743 fails for both Linux and
MinGW.  At the same time, running sparse.tst passes on both operating
systems.  There may be an additional reason why sparse.tst is failing on
MSVC ...

--Rik