20021028: IMGOLAY/FRMOLAY on Solaris SPARC gcc/g77 does not work (cont.)

>From: Robert Mullenax <address@hidden>
>Organization: NSBF
>Keywords: 200210270159.g9R1x6q20772 McIDAS v2002 gcc/g77

Hi Robert,

re: make sure that a 'which mcserv' run from the Unix prompt and from the
McIDAS Text and Command window report the same thing.

>They are the same.
>
>/export/home/weather% which mcserv
>/export/home/mcidas/bin/mcserv
>
>
>           M c I D A S - X     V e r s i o n  2002a Unidata 020810    
>                                                                       
>                                                                       
>SKED SKEDFILE                                                          
>Scheduler starting on file SKEDFILE                                    
>!which mcserv                                                          
>/export/home/mcidas/bin/mcserv 

OK.  I was hoping that the problem was somehow related to McIDAS environment
variable settings in the shell configuration file (e.g., .cshrc for C shell,
.profile for Bourne shell, etc.).

>BTW before I e-mailed you at first I downloaded, built and installed v.2002A. 
>Just to make sure V. 2002 didn't have some issue.

I wouldn't expect this to have made a diffence at all since there was
nothing in the 2002a addendum that should be related to the running of
'mcserv -R'.

Just to make sure that I am understanding things:  you _are_ able to load
both satellite and NEXRCOMP images from a remote server, correct?

>Would it be worthwhile to get later versions of the compilers?  

I don't know.  If I understood why there was a failure in loading mcserv,
I would be better able to answer this question.

>>From address@hidden Mon Oct 28 22:01:50 2002
>>Subject: Re: 20021026: IMGOLAY/FRMOLAY on Solaris SPARC gcc/g77 does not work

>I downloaded gcc-2.95.3 from sunfreeware and install it (it contained the
>upgraded g77 as well).  I then did a make clobber on McIDAS and
>built/installed it again.  Unfortunately I still get the same error.

Well, this certainly answers the compiler upgrade question.

The problem I am having in trying to troubeshoot this is that I do not
get the error you are seeing on any of the platforms I am using gcc/g77
on!  I checked the code that is called from m0addef.for, and it can
return a '-1' status for a variety of reasons, most of which are related
to forking a new process.

Do you have Sun compilers you can use to build?  If not, do you have
f2c installed so you can build using 'make VENDOR=-gcc'?  I am trying
to think of ways of testing to see if the problems you are having
are really related to use of g77.

Tom