Re: leasqr error

[Lukas Reichlin]

On 14.11.2012, at 14:05, Martin Helm <address@hidden> wrote:

> I can see that problem when I (as your error message indicates) load the
> control package and the optim package, it does not happen when I only do a
> 
> pkg load optim
> 
> and do not load control in parallel, this looks as if they contain
> conflicting options functions.

On 14.11.2012, at 14:00, Ben Abbott <address@hidden> wrote:

>>> The cell you pass to leasqr is undefined
>>>> {stol,niter,wt,dp,dFdp,options}
>>> What are all these things?
>>> 
>>> The error says that "options" is being parsed as a function.
>> 
>> The error is from line ...
>> 
>>      leasqr(x,y,pin,F,{stol,niter,wt,dp,dFdp,options})
>> 
>> Which is *not* part of the example at the  link you provided.  You have not 
>> defined many of the variables in this call to leasqr.  In particular 
>> "options" has not been defined.
>> 
>> Ben

The control package has a function called "options". Use another name for your 
options argument in leasqr, e.g. "opts", or use "pkg unload control".

But this should be no problem at all if you defined options beforehand, because 
variable "options" would shadow function "options". I think there is another 
problem:
AFAIK the curly brackets in

        leasqr(x,y,pin,F,{stol,niter,wt,dp,dFdp,options})

state that these arguments are optional. It should be sufficient to use

        leasqr(x,y,pin,F)

Even if you want to specify all arguments, you must not write the curly 
brackets:

        leasqr(x,y,pin,F,stol,niter,wt,dp,dFdp,options)

Lukas