iraf.net

Hi,I emailed on the 24th October regarding a problem i'm having compiling the
nmisc external package for SuSE 10. I understand if your busy with a backlog
of questions after the Thanksgiving holiday, so sorry to bug you again with
this query.My ability to program in C is limited, never mind SPP, but......Line 78 of t_ccdmask.c refers to a subroutine 'cmmask_' with 14 arguments.
line 159 refers to 'cmmask_' but two of the 14 arguments are the same (&ngood)
line 212 seems to be the 'cmmask_' subroutine, but it only has 13 arguments.The same kinda thing is present in the t_ccdmask.x file on lines 84 with
'cm_mask' having 14 arguments and lines 125 'cm_mask' having 13 arguments.I tried deleting one of the duplicate 'ngood' arguments from
line 84 of 't_ccdmask.x' and re-compiling the nmisc package from scratch.
It seems to have corrected the problem of the missing 'x_nmisc.e' from the
bin.suse directory (details in the email on the 24th).I've done a very quick test of nmisc.ccdmask and it seems to have worked.
Could someone confirm that this was the correct thing to do.Also, what is the significance of the output to the terminal once the
compiling is complete, i.e.[1] done 48.2 1:03 76%should it read 100% ?????One other thing, unpacking the fitsutil tar file yielded no softlink
between bin->bin.generic. bin/ needed to be deleted first and then the
soft link created. The compiling of the FITSUTIL package seemed to go ok with
fxheader/fxsplit working ok.cheersDan

Hi Dan,
Sorry for the unusually slow reply.
> ......
> Line 78 of t_ccdmask.c refers to a subroutine 'cmmask_' with 14 arguments.
> line 159 refers to 'cmmask_' but two of the 14 arguments are the same (&ngood)
> line 212 seems to be the 'cmmask_' subroutine, but it only has 13 arguments.
>
> The same kinda thing is present in the t_ccdmask.x file on lines 84 with
> 'cm_mask' having 14 arguments and lines 125 'cm_mask' having 13 arguments.
>
> I tried deleting one of the duplicate 'ngood' arguments from
> line 84 of 't_ccdmask.x' and re-compiling the nmisc package from scratch.
> It seems to have corrected the problem of the missing 'x_nmisc.e' from the
> bin.suse directory (details in the email on the 24th).
>
> I've done a very quick test of nmisc.ccdmask and it seems to have worked.
> Could someone confirm that this was the correct thing to do. The code change looks correct. Some of the newer GCC compilers
are now throwing errors at argument mis-matches like this where before
they were ignored as warnings. Most of the similar problems have been
removed from the core system long ago but beware you may encounter similar
problems in other external packages in the future (I believe TABLES/STSDAS
have similar issues). If possible using pre-compiled binaries will
generally work, most commonly 'redhat' binaries are available but will run
under newer SuSE so long as you unpack them in the package 'bin.suse'
directory.> Also, what is the significance of the output to the terminal once the
> compiling is complete, i.e.
>
> [1] done 48.2 1:03 76%
>
> should it read 100% ????? These are just estimates of the cpu/user times (i.e. 48.2/63 sec
= 0.76).> One other thing, unpacking the fitsutil tar file yielded no softlink
> between bin->bin.generic. bin/ needed to be deleted first and then the
> soft link created. The compiling of the FITSUTIL package seemed to go ok with
> fxheader/fxsplit working ok. Thanks, I've corrected the tar file in the archive for the future.
Hope this helps, let us know if you still have problems or questions.Regards,
Mike Fitzpatrick--------------------------NOTE: As of Jan 1, 2006 NOAO will no longer be operating a HelpDesk for
IRAF at iraf@noao.edu. Support is being moved to a forum-based system
at https://iraf.net, please post future questions to that site. Thanks.