[Fwd: Re: [Fwd: Re: EigsRealNonSymmetricFunc]]

[Jordi Gutiérrez Hermoso]

There you have it, see below.

- Jordi G. H.

-------- Forwarded Message --------
From: David Bateman <address@hidden>
To: Jordi Gutiérrez Hermoso <address@hidden>
Subject: Re: [Fwd: Re: EigsRealNonSymmetricFunc]
Date: Sat, 16 Aug 2014 18:17:31 +0200


> Le 16 août 2014 à 17:28, Jordi Gutiérrez Hermoso <address@hidden> a écrit :
> 
> Oops, looks like I neglected to CC you in this exchange below.
> 
> - Jordi G. H.
> 
> -------- Forwarded Message --------
> From: Jordi Gutiérrez Hermoso <address@hidden>
> To: Tommi Höynälänmaa <address@hidden>
> Cc: octave maintainers mailing list <address@hidden>
> Subject: Re: EigsRealNonSymmetricFunc
> Date: Sat, 16 Aug 2014 09:16:27 -0400
> 
> Moving this discussion to the maintainers' list...
> 
>> On Fri, 2014-08-15 at 15:58 +0300, Tommi Höynälänmaa wrote:
>> What is the purpose of the postprocessing after call to dneupd in 
>> function EigsRealNonSymmetricFunc in liboctave/numeric/eigs-base.cc?
> 
> I suppose you mean this part of the code?
> 
>    
> http://hg.savannah.gnu.org/hgweb/octave/file/c4ffcc961f3c/liboctave/numeric/eigs-base.cc#l2669
> 
> I don't know, it looks pretty obscure. Looks like the code has been
> mostly untouched since David Bateman first wrote it. You can verify
> this with
> 
>    hg annotate --user --number --ignore-all-space 
> liboctave/numeric/eigs-base.cc
> 
> or, shorter,
> 
>    hg ann -unw liboctave/numeric/eigs-base.cc
> 
> (read `hg help annotate`)
> 
> Let's try asking him, then. David, can you comment on what's going on
> in this part of the code?
> 
> - Jordi G. H.
> 


It was to put the order of the eigenvalues in the same order as Matlab at the 
time I wrote the code. Matlab might have changed the order since so no idea if 
this is still the right thing to do and I don't have access to Matlab yo check

D.
>