20010212: doubly typed u0getcolor in raobcon.pgm

>From: Neil Smith <address@hidden>
>Organization: Texas A&M
>Keywords: 200102130028.f1D0SQL12744 McIDAS-X 7.7x raobcon.pgm

Neil,

>Had problem building on DEC DS20E, TU 5.0. I know DEC isn't
>supported above OSF/1 v4.0, but was hoping this failure on
>compiling raobcon.f was not OS specific. Any guesses?

>./mccomp -O -vendor -I. -I../netcdf/libsrc -I../jpeg -I../zlib -c raobcon.pgm
>rm -f raobcon.o
>f77 -c -O -I. -I../netcdf/libsrc -I../jpeg -I../zlib raobcon.f
>f90: Error: raobcon.f, line 300: This name has already been assigned a data ty
> pe.   [U0GETCOLOR]
>      integer u0getcolor     ! get next color         <<<<< UPC add 981202 >>>
> >>
>--------------^
>compile        raobcon.pgm:            FAILED

The problem is a replication of the line:

integer u0getcolor     ! get next color

in raobcon.pgm.  I had just found this when making another attempt at
getting g77 to work compiling McIDAS-X.

The solution (which I will be inluding in the 7.705 addendum soon) is
to simply delete the second declaration or comment it out.  After that
you can resume your build by typing:

make

or

make all

depending on if you are building McIDAS-X or McIDAS-X and -XCD.

>The environ. was as required for v4.0:
>CC=/bin/cc
>CPPFLAGS=-DNDEBUG
>CFLAGS=-O
>FC=/bin/f77
>FFLAGS=-O
>CXX=

This one was strictly a replication of a declaration line.  The OS/1
4.0 Fortran compiler (and the ones for all other supported OSes) did
not much care about this replication.  It seems that the DEC OSF/1 5.0
one (and g77) does.

Tom

>From address@hidden Tue Feb 13 10:24:03 2001
>Subject: 20010212: doubly typed u0getcolor in raobcon.pgm (cont.)

Neil,

>Thanks Tom. That did the trick.

Good news!

>FYI there were a couple other
>files with the same problem and same solution:
>
>astat.pgm - iniflg - line 2910
>datavail.pgm - size - line 592

Thanks, I should have reviewed my notes before answering your inquiry
yesterday as I had found both of these in my attempt to get things
working with g77.  Looks like the OSF/1 5.0 Fortran compiler will be a
good one to use to clean up some code.

On another note, are you planning on installing/running the remote ADDE
server on this machine?  The reason I ask is that I am troubleshooting
a problem that has cropped up when serving sounding data (Skew-Ts and
Hodographs) through the remote ADDE server on Linux machines (problem
so far seen on RedHat 6.2 and 7.0).  I am curious if this same problem
will show up under OSF/1 5.0.  Please let me know if you are planning
on doing the install (or want help doing the install).

>Thanks again, -Neil

Tom

>From address@hidden Tue Feb 13 10:50:29 2001
>Subject: Re: 20010212: doubly typed u0getcolor in raobcon.pgm (cont.)

Well, I suppose I could. This platform looks like a good
candidate. But it may be a day or two before I can
get to it.  Will keep you informed. (Most likely via
faint screams of stark terror you'll hear out your window
from somewhere down south.) -Neil