Re: Moving code from octave-forge to octave [Was: polyderiv problem?]

[Paul Kienzle]

Patches are from the early, early days of octave-forge, and I no longer 
maintain them in any sense.  I want everything in octave-forge to work 
on an unpatched version of octave.  Feel free to delete.

As for automatic duplication, yes I would like to see an updated 
version of this patch included in Octave.  We should count the number 
of times it occurs in octave-forge to see if it is worthwhile.  I find 
the alternative hard to read and slower than it ought to be.

- Paul

On Feb 9, 2005, at 10:50 AM, David Bateman wrote:

> I looked at the patches directory of octave-forge and it seems that 
> all of the patches here, except one are either already included in the 
> main part of octave or superseded by other changes to octave. The only 
> remaining patch to consider is M-v-ops-2.1.31.patch that allows the 
> extension of ./, .*, etc to allow Matrix by vector arguments. Its not 
> clear how this ports to NDArrays, since in fact the NDArray 
> generalization means that singleton dimensions in one of the args 
> should force the matrix to be duplicated allow that dimension of the 
> other arg.
>
> Do we want to do that? In any case, matlab doesn't support this 
> behavior, and it might hid some dimensioning errors from the since 
> something like "randn(N,1)+randn(1,N)" would create an N-by-N 
> matrix... It is clear however, that the existing patch is of no use as 
> a basis for this as it is against an octave version prior to the 
> introduction of NDArrays. So all of the patches in octave-forge might 
> be dumped...