Any chance you can run valgrind on this command? Garbage in output
suggests use of uninitialised variable and valgrind will tell us.
Tim
On Wed, 21 Jun 2006, Eduardo Unda-Sanzana wrote:
> Dear all,
>
> I am doing photometry using PHOTOM and CURSA. PHOTOM works OK but I have
> found some strange behaviour when running the CURSA routine catphotomfit.
> Basically, if I set the parameter FIXED to FALSE, the routine should fit two
> parameters ("zero point" and "atmospheric extinction"). However, running the
> routine in this way gives a constant "zero point" = 0.000000, and what seems
> to be a random "atmospheric extinction". In other words, if I run the command
> repeatedly, I don't obtain the same value for the second parameter. That
> should not be the case, I guess.
>
> A workaround I have found is to set FIXED=TRUE, then setting both FZEROP and
> FATMOS to FALSE (which should be equivalent to setting FIXED=FALSE). Then the
> routine does a fit which looks reasonable in most cases but I'm not happy
> about this because it makes me feel suspicious about the results I get. Could
> anybody please have a look into this, please? If needed I can provide
> suitable tables of magnitudes and airmasses (in CURSA format) to use as a
> test.
>
> Thanks in advance,
>
>
--
Tim Jenness
JAC software
http://www.jach.hawaii.edu/~timj
|