Re: [CHANGESET]: First attempt at a single precision type.

[David Bateman]

David Bateman wrote:
>> BTW, my approach to getting single working might have been to convert
>> the current double classes to a use template class instantiated with
>> double, then generate the corresponding single class.  But what you've
>> done also seems reasonable to me, as it gets something working
>> faster.  As I see it, given the changes you've made, we can also work
>> toward a template version by pushing duplicate code to a template base
>> class, and that's something that we can also do incrementally.
>>   
> Ok, in this case let me start on the first increment, and get the rest
> of the code converted to single precision before you import the code
> into your repository. I'd also like to add test code, though I suppose
> that can wait as well, though is essential prior to a release. Give me a
> couple of days..
> 
> D.
> 

Ok I updated my repository with the following

* Respect the assignment rules like "a=zeros(2,2,'single');a(1) = 1"
* Add mixed integer single/integer operators
* Prefer demotion to single precision rather than promotion to double
* Also convert the balance, qr, find, hess, expm functions
* Extract the generalize eigenvalue balancing from balance.cc and move
it to liboctave.

I still have started templatizing some of the Array code or added test
code to the single precion type, though in my brief tests it appears to
be correct. I'll try to add templates for at least the obvious functions
before I'd suggest importing this to the main repository.

D.