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

[McIDAS #WQW-520685]: Fedora



Hi Richard,

re:
> I did as you suggested and modified tcl/unix/configure and
> tk/unix/configure files.

FYI: I bundle the stable latest TclTk release with Unidata McIDAS-X;
SSEC is still bundling an old version (8.4.2). After I grabbed the
latest distribution (8.4.13), I noticed that the TclTk folks have
finally fixed the error in the configure scripts.

> I then restarted the make and got an error in
> program xcdgrib.c.  I went in and determined that I needed to delete the
> word static on line 486 of xcdgrib.c.  I then restarted the make.
> Everything then compiled fine and I then ran the install script.  I had
> AREA files from work on a CD and used them to test McIDAS and everything
> seems to work perfectly under Fedora 5.

Thanks for reminding me that I also had to make a change related to the
declaration of GDSToRowsColumns.  I chose, however, to make the modification
in the includes file, mcgrib.h:

mcgrib.h:static int GDSToRowsColumns(GDS *, int *, int *); /* <<<<< UPC mod 
20050615 >>>>> */

My change log file (file where I log all changes made to the SSEC distribution
when creating the Unidata distribution) indicates that I first ran into
this problem in Fedora Core 4:

CHANGES.2005
 ...

mcgrib.h               SSEC 2005 version with mods
                         changed declaration of GDSToRowsColumns to
                           'static int' to match declaration in xcdgrib.c.
                           Problem uncovered when compiling with gcc 4.0
                           under Fedora Core 4.

> Thanks for the help.

No worries.  Glad to help.

Just so you know, I will be feeding all of the code fixes I make back
to SSEC for (possible) inclusion in core McIDAS (I have CVS access
to their development system).

Another issue:  I and a number of my sites that are running on
'fast' hardware with 'current' Linux installations have noted a problem
whereby not all object modules get correctly added to McIDAS libraries
(e.g., libmcidas.a, libSDI.a).  I traced the problem down to the version
of 'ar' that is being distributed with "current" Linux distributions (e.g.,
FC[45], Slackware 9.0, SuSE 10, etc.).  My "fix" was to add short sleeps
after 'ar' invocations in mcar.sh.

Question: Have you run into this situation? Whether you have or have not,
what kind of hardware are you building McIDAS on (e.g., processor speed and
type, RAM, type and speed of disks).

Thanks for your input!




Cheers,

Tom
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: WQW-520685
Department: Support McIDAS
Priority: Normal
Status: Closed