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

20010212: McIDAS on Red Hat 7

>From:  Wahner Paul Contr CSR4500 <address@hidden>
>Organization:  Cape Canaveral Air Force Station
>Keywords:  200102121310.f1CDAFL23157 McIDAS-X Linux &


>I recall you saying, during the MUG meeting, there were some minor changes
>required to build McIDAS on Red Hat 7.

I did a few things over what SSEC did for building on Linux.  Here are
some of the changes I felt were necessary (the comments about RH 6 apply
to RH 7 as well):

ncgentab.c             SSEC 7.70 version of netCDF 3.4 is beta
                         replaced module with general 3.4 module so it would
                           compile on RedHat 6.0 Linux

ncgenyy.c              SSEC 7.70 version of netCDF 3.4 is beta
                         replaced module with general 3.4 module so it would
                           compile on RedHat 6.0 Linux

decoder.cp             SSEC 7.70 version with mods
                         changed sigset call to signal for Linux; NOTE: not
                           tested, but decoder is only used for SDI boxes
                         defined MAP_FAILED to be ((void *) -1) for OSF/1
                         included if !defined(__FreeBSD__) with Linux

mcar.sh                SSEC 7.70 version with mods
                         don't return error if mcar run with no objects
                         added section for FreeBSD
                         modified ranlib entry for Linux

mccomp.sh              SSEC 7.70 version with mods
                         added sections for FreeBSD
                         modified SSEC sections for Linux
                         added section to trap Sun SC4.0 compilation bug
                         added section to trap Sun SC5.0 compilation bug
                         added passing of loader flag for stripping executables
                           for IRIX and IRIX64

tclcomp.sh             SSEC 7.70 version with mods
                         added section for DEC OSF/1
                         added section for FreeBSD
                         added section for Linux

>Specifically I believe it was something regarding Tcl 8.2.

I did run into a problem when the OS has Tcl/Tk installed and it is a
newer version that what McIDAS uses.  The mod in tclcomp.sh was aimed
at this.

>Do you build your McIDAS with g77 or f2c? I've
>noticed g77 comes as part of my boxed set so I'd like to use that if it is

I have made a number attempts at compiling McIDAS with g77.   I am
getting closer to getting something working, but there are still a
number of problems.  I strongly recommend that you not waste your time
with g77 until all of the coding bugs in McIDAS source are taken care
of (g77 does not initialize varibles to 0 like most "real" Fortrans;
this causes big problems in execution).

>Any information you could share would be greatly appreciated. Be gentle
>though, as I am still a Linux neophyte. 

Since I still have my own distribution of McIDAS, it is not a simple
thing for me to tell you what  you would have to do with the SSEC
distribution in order to get things to build correctly on Linux.  You
can grab my distribution if you like, buth that will put you out of
sync with SSEC mods.

Another thing that does not work on RH 7 is the
installation/configuration of the ADDE remote server.  mcinet7.7.sh
does not work since inetd has been replaced by xinetd in the latest
Linux releases.  Along with this is a change in how configurations are
done.  inetd configuration was done by modifying /etc/inetd.conf and
/etc/services.  xinetd configuration is done by creating two files in
/etc/xinetd.d; one for mcserv (port 500) and the other for mccompress
(port 503) access.  I discuss what has to be done in:


This document, Notes and Warnings, also has other items that sites
wanting to build with Linux should read.

One last thing: just last week I uncovered (actally a site of mine
found this) a problem producing soundings and hodographs on Linux.  The
situation is very specific: the failure occurs when a Linux system is
hosting the ADDE remote server and the client that is requesting the
data is not on the same LAN as the serving machine.  I will be beating
on this problem this week, so I may have more on this later.

Please let me know if you would like to grab my McIDAS distribution. or
if you would like help in getting McIDAS running on your Linux box.  In
fact, it might be a real good test for you to plunge ahead with a build
of the SSEC distribution _without any of my changes_ to see if things
will build.  If/when they don't, we could roll in just the change to
get things compiling/linking and continue.  Eventually, you would get a
system that should run, and then we could test out the UAPLOT/HODO
problems accessing data from a remote server installed on your box.
This would let me know if I have introduced something that is making
things break (I don't think I have, but...).  I would consider this a
BIG favor, so I hope that you will agree!

>Thank you, 
>Paul Wahner
>MIDDS Systems Engineer/ Meteorologist
>CSR 4500, Cape Canaveral Air Force Station
>address@hidden <mailto:address@hidden> 
>(Com) 321.853.8184 (Fax) 321.853.8291
>(DSN) 467.8184


>From address@hidden Wed Feb 14 06:08:15 2001
>Subject: RE: 20010212: McIDAS on Red Hat 7 


Thank you for the information. I will do as you suggest and just work toward
getting the SSEC distribution working. 

I'm actually playing with this in my spare time on one of my PC's at home. I
will try to help you in any way I can though at this point I have a bit of a
learning curve with the nuances of RH7. 

I'll be in touch. 


Paul Wahner
MIDDS Systems Engineer/ Meteorologist
CSR 4500, Cape Canaveral Air Force Station
address@hidden <mailto:address@hidden> 
(Com) 321.853.8184 (Fax) 321.853.8291
(DSN) 467.8184

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.