[MITgcm-devel] Re: [MITgcm-cvs] MITgcm/tools CVS Commit

[Ed Hill]

On Sun, 2005-12-04 at 11:30 +0100, Martin Losch wrote:
> Ed,
> 
> you new "sigrec" causes not only my g77 tests to crash (I'll repeat 
> them, because I think my tests started half an hour before you fixed 
> the problem), but also my pgf77 test on the XD1, also a few with Sun 
> F77, here's an example error message from the XD1
> mpicc  -c sigreg.c
> PGC-S-0157-A declaration may not appear after a statement (sigreg.c: 41)
> PGC/x86-64 Linux/x86-64 5.2-4: compilation completed with severe errors
> > xd1>> which mpicc
> > /usr/mpich/mpich-1.2.5-pgi522/bin/mpicc
> 
> I have no idea about C, so I cannot give you an idea what's going wrong.

Hi Martin & Jean-Michel,

My sincere apologies for the syntax errors.  I've checked in a fix and
confirmed that its working OK with the PGI compiler.  For faulks, a new
set of PGI results will appear on our testing pages within two hours to
confirm the fix.

And in light of the problems reported by you and Jean-Michel, I will
renew my vows concerning the running of testreport.  That is, I will be
careful to get clean results with testreport for a few experiments [1]
with at least *three* compilers (GNU, Intel, PGI) before committing any
non-trivial [2] changes.

Ed

[1] Running all the test setups with all three compilers simply 
    takes too long.  I'll run the ones defined by our CVS alias 
    "MITgcm_verif_basic".  So if you don't like that set please 
    discuss it amongst yourselves and edit the CVSROOT/modules 
    file as you see fit:

      cvs co CVSROOT

[2] So, how do we define non-trivial?  ;-)

-- 
Edward H. Hill III, PhD
office:  MIT Dept. of EAPS;  Rm 54-1424;  77 Massachusetts Ave.
             Cambridge, MA 02139-4307
emails:  eh3 at mit.edu                ed at eh3.com
URLs:    http://web.mit.edu/eh3/    http://eh3.com/
phone:   617-253-0098
fax:     617-253-4464