[Developers] info lines in post-processed source

Mark Miller mamiller at cyga.jpl.nasa.gov
Fri Jan 23 21:27:24 CST 2004


Cactusers,

I've noticed that post-processed source code has lots of info lines
that look like, e.g.,

# 28 "/u2/mamiller/Cactus_beta13/arrangements/...............

My intel compiler on my laptop and on beowulf clusters have no
problem.  However, I fell off my chair from dizziness at all of the
warnings flashing by from the f90 compiler on lemieux.  Worse, on the
O2K here at JPL, the first Fortran file to get compiled dies
due to what appears to be the compiler misinterpreting these
supposidly "null" lines, e.g.,

f90 -r10000 -TARG:proc=r10000 -64 -mips4 -O3
-OPT:roundoff=3:IEEE_arithmetic=3:const_copy_limit=100000 -c -o
$current_wd/ADM_Constraints.F.o $current_wd/ADM_Constraints.f
f90-995 f90: INTERNAL ADM_CONSTRAINTS, File =
/u2/mamiller/Cactus_beta13/arrangements/SpaceTime_Util/ADM_Constra
, Line = 1
  Unrecoverable error encountered while attempting to print buffered
messages - can not open source file.
f90 INTERNAL ERROR:  /usr/lib32/cmplrs/mfef90 returned non-zero status 1
gmake[3]: *** [ADM_Constraints.F.o] Error 1

Any chance of cleaning this up?  Is a line starting with
a "#" character supposed to be ignored by all compilers?  I would
guess not since this is supposed to indicate a preprocessor directive.
How does one tell the compiler that these are to be ignored?  Or am
I way off base here?

-Mark


------------------------------------------------
Mark Miller
Jet Propulsion Laboratory 238-332
4800 Oak Grove Drive
Pasadena, CA  91109
E-mail: Mark.A.Miller at jpl.nasa.gov
office phone: (818) 393-2657, fax: (818) 393-6890
------------------------------------------------




More information about the Developers mailing list