Re: bitcmp function error?

[Andreas Corneliussen]

Sorry for the long delay in answering.

I have found that the problem were with me expecting the function to do something it were not expected to. I expected the function to calculate the two's complement of the number, but instead it is calculationg the actual complement. Therefore my bug-filing is mute.

Med venlig hilsen / Best Regards / Mit freundlichen Grüßen

Andreas Corneliussen     Ektos A/S

Development Engineer       Farvervej 35

                                        DK-8800 Viborg

Mobile: +45 3135 0886

E-mail: address@hidden

2013/6/11 Michael D. Godfrey <address@hidden>

On 06/11/2013 09:48 AM, PhilipNienhuis wrote:

Andreas Corneliussen wrote

>Hi
>
>I've today used some time on debugging my vhdl system for errors, just to
>find out that the error were not in my VHDL code but in the testbench
>generated by an octave script.
>I have found that the bitcmp.m function of release 3.6.2 installed from
>linux mint package system (think it is ubuntu repos'es) calculates the
>two's complement of a number simply as a bitwise not and not as it should
>be (bitwise not) +1.
>
>I have corrected this on my own PC, but would like to inform you
>developers
>of the packages so that future users can be spared of the frustration of
>having to debug errors that are in the core octave system (which I think
>the majority of users, like me, assume is correct ;-)).

Although in your other thread in the help-octave ML it was suggested that
the developers would pick up from the maintainers list (indeed we sometimes
do), a formal bug report is much better, see:
http://www.gnu.org/software/octave/bugs.html

If you can, please attach a patch there, preferrably as a Mercurial
changeset.

Thanks,

Philip

Before reporting this as a bug you might note that Octave appears to be compatible
with Matlab for bicmp(A,2)  for A = 0,1,2,3.  For values of A < 3 Matlab returns an error.
It is possible that Octave should do this also in order to improve compatibility.

Or, if you have a different case, please explain exactly what case fails.

Michael