Re: feval call to a subfunction fails when it is in a subfunction

[Jaroslav Hajek]

On Thu, Oct 1, 2009 at 9:59 PM, Sidney Markowitz <address@hidden> wrote:
> Jaroslav Hajek wrote, On 2/10/09 8:36 AM:
>>
>> The rather general fix for development sources was not transplantable
>> to 3.2.3 due to the binary compatibility policy. A separate fix needs
>> to be created.
>
> So this is a known bug in the 3.2.3 branch? Which bug # is it? I didn't find
> it (or didn't notice it) searching for problems with feval. Is the problem
> with feval just one symptom of something more general?
>

I didn't actually check, but I believe it is an instance of this problem:
http://www.nabble.com/Using-STR2FUNC-in-sub-function-to-refer-to-other-sub-function(s)-td25401966.html#a25432999


> Is this considered serious enough to port to the 3.2.x branch? I don't have
> much experience with Octave yet, but this completely breaks mgene, a large
> gene prediction package that was my reason for trying to get Octave to work,
> and it does look like this use of feval would be quite common.
>

I don't think this ever worked, so it's not really a regression... if
I have the time, I'll invent a patch for 3.2.4.
The relevant patch for 3.3.50+ is
http://hg.savannah.gnu.org/hgweb/octave/rev/8d79f36ebdde
and it may be usable with 3.2.3 sources or just use the development
sources. It can't be transplanted to the 3.2.x tree because a new
member field is added which is a kind of change that generally breaks
binary compatibility.

-- 
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz