Re: unintended consequences of running Matlab script

[Jordi Gutiérrez Hermoso]

On 28 June 2012 09:29, Thomas Scofield <address@hidden> wrote:
> I am working at getting someone else's Matlab programs to run in
> Octave v. 3.4.3, which I obtained from MacPorts.  Now and then this
> appears to involve substantive issues, such as removing the 'qz'
> option from a call to eig().

Why was this necessary? There is a qz command in Octave. I'm currently
investigating one bug related to it. Did you find another? If so,
please report it.

> But the warnings are the most annoying problems. What I really don't
> understand is all the warnings that start appearing after I've run a
> script, particularly when the results of running it seem to be the
> ones it was meant to produce. For instance, before I run it,
> commands like
>
>   help length
>   svds(rand(5,8))
>
> behave normally. However, after I run the script, here is the output
> from the same three commands:
>
> octave:9> help length
> warning: concatenation of different character string types may have
> unintended consequences

The script that you ran at some point did something like

    warning("on")

or

    warning on

which turns on *all* warnings (see "doc warning"). A lot of these
warnings are about using Octave-like syntax, which might be
undesirable if you want your m-files to run in Matlab. Sadly, this
warning also warns about Octave's own m-scripts having Octave syntax
(this really is a bug).

Hunt down the line in your script that turned on all warnings and
eliminate it.

HTH,
- Jordi G. H.