ddd
[Top][All Lists]
Advanced

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

Re: Printing Fortran muti-dimensional array elements


From: Andrew Gaylard
Subject: Re: Printing Fortran muti-dimensional array elements
Date: Mon, 20 Dec 2004 08:47:06 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4.1) Gecko/20031114

Claude N Williams wrote:
Our center is migrating from SUN workstations to Linux RedHat Enterprise (RHE) servers. Our current RHE (i386-redhat-linux) version contains:

GDB  6.1
G77  3.2.3
DDD  3.3.1

We have the problem of being an R&D shop using 4 and 5 dimension arrays. We are having the problem that I have found (since 1999) up to 2002 about the dimension size swapping and confusion when trying to examine an array element. For instance:

Defining an array(21,101,13) and attempting to examine the (2,1,13) element dies with a segmentation fault!

I realize, through the mail lists, this is (was?) due to the translation to C before compilation in G77 and how the gtabs are (were?) handled by GDB. But I have not seen anything recent.....

Has this problem been fixed? Can we just update the apps. to the current releases and the problem is fixed?

If not, is there a work-around? Other than emulating a one dimensional array and typing out the arithmetic calculation of the offset of the element?

If not, will any of the future releases of any of the apps. above fix this problem (or Gfortran)?

Please point me in the right direction - I am the point man for about 20 scientific programmers and I want them to have a good experience with Linux!

Thanks,
Claude Williams
Research Scientist
National Climatic Data Center
Asheville, NC.

Hi there Claude,

The best release is 3.3.9, but I doubt that it fixes your problem.
Nonetheless, I can recommend upgrading anyway, since problems on older
releases won't be fixed.  It's just the way we work.

Then, having upgraded to 3.3.9 or .10, produce a small testcase consisting
of Fortran source and ddd/gdb commands, and then we can dig into the
problem further.

Cheers,
Andrew.





reply via email to

[Prev in Thread] Current Thread [Next in Thread]