[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20011022: McIDAS-X 7.8 build failure on IRIX at Kwaj



>From:  Rick Kohrs <address@hidden>
>Organization:  SSEC
>Keywords:  200110221850.f9MIog122926 McIDAS-X netCDF build

Rick,

>Quick background:
>SGI indigo/2
>McIDAS 7.6 Compiled fine.
>McIDAS 7.8 fails when compiling attr.c (netcdf routine).
>Compiler and OS levels are the same as our SGI inhouse.
>
>Couple of things we have noticed:
>1) Core McIDAS builds do not show attr.c getting compiled.

Right.  attr.c is part of the netCDF portion of the McIDAS-X bundle.

>2) Dave Glowaki said the netcdf compile output would not show up in the
>   log.

It shouldn't show up in the makelog file, but it will scroll by on
the screen during the build.

>3) The log from Kwaj does show the compile.

Actually, the output you sent is what is being written to the screen
during the build.  It is not a snippit from makelog.

>I have requested a copy of attr.c from Kwaj.

I don't think that will be necessary.  I recommend that whoever is building
7.8 define the following in their environment _before_ starting the McIDAS
build (C shell syntax):

# Defines for netCDF build
setenv CC "/bin/cc -n32"
setenv CPPFLAGS -DNDEBUG
setenv CFLAGS -O
setenv FC "/bin/f77 -n32"
setenv FFLAGS -O
setenv CXX

The point is to use 'cc' not 'c89'.  The output from Kwaj below shows that
their build is using 'c89':

>> > > >         /bin/c89 -c -O -I.  -DNDEBUG attr.c

I have run into the exact same thing building on IRIX, and the solution is
always to specify the use of 'cc'.  The absolute easiest way to do this
is to set the CC environment variable as I have indicated.

The steps I would send to Kwaj at this point is:

<login as 'mcidas'>
<edit their shell specific environment definition file: .cshrc for C shell;
.profile or .kshrc for Korn shell; etc>
<make the changes active: . .profile for Bourne/Korn shell; source .cshrc
for C shell>
cd ~/mcidas7.8/src
make clobber
make

>Thanks,

Later...

Tom

>--------------CF00D032CADA89BF670F6631
>Rick Kohrs wrote:
>
>> Hi All,
>>
>> Our McIDAS site at Kwajalein is getting some compiler errors (compiling
>> McIDAS 7.8) that we are not seeing.  They are running the supported
>> version of SGI (6.5) and compilers (7.2.1).  There system has been
>> static since 7.6 and that compiled just fine.  I'd appreciate any help
>> you can give.  Here's the error messages:
>>
>> > > > compile nvprep.for:             done
>> > > > update  libmcidas.a:            done
>> > > > compile calb11.for:             done
>> > > > update  libsdi.a:                  done
>> > > > compile BitUtil.c:               done
>> > > > compile Bk11Data.c:          done
>> > > > compile crc16.c:                  done
>> > > > compile GVAR.c:               done
>> > > > compile MSAT.c:               done
>> > > > compile NCDF.c:                done
>> > > > compile nexrutil.c:              done
>> > > > compile POES.c:                done
>> > > > compile SDIConv.c:           done
>> > > > compile SDIUtil.c:             done
>> > > > compile servutil.c:              done
>> > > > compile TIP.c:                    done
>> > > > update  libsdi.a:                  done
>> > > >         cd ../netcdf && make
>> > > >
>> > > > Making `all' in directory /local/vendor/ssec/mcidas/mcidas7.8/netcdf/l
> ibsrc
>> > > >
>> > > >         m4 -B10000 attr.m4 >attr.c
>> > > >         /bin/c89 -c -O -I.  -DNDEBUG attr.c
>> > > >
>> > > > "/usr/include/stddef.h", line 37: error(1084): invalid combination of 
> type
>> > > >           specifiers
>> > > >   typedef int   ptrdiff_t;
>> > > >                 ^
>> > > >
>> > > > "/usr/include/sys/types.h", line 218: error(1084): invalid combination
>  of type
>> > >
>> > > >           specifiers
>> > > >   typedef int    ssize_t;
>> > > >                  ^
>> > > >
>> > > > 2 errors detected in the compilation of "attr.c".
>> > > > *** Error code 2 (bu21)
>> > > > *** Error code 1 (bu21)
>> > > > *** Error code 1 (bu21)
>> > > > *** Error code 1 (bu21)
>> > > > ./mcidas7.8.sh: ERROR: 'make' failed
>
>--------------CF00D032CADA89BF670F6631--
>
--
+-----------------------------------------------------------------------------+
* Tom Yoksas                                             UCAR Unidata Program *
* (303) 497-8642 (last resort)                                  P.O. Box 3000 *
* address@hidden                                   Boulder, CO 80307 *
* Unidata WWW Service                             http://www.unidata.ucar.edu/*
+-----------------------------------------------------------------------------+


NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.