Have_Round macro defined during configure process

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Have_Round macro defined during configure process

Clinton Chee-2

Why does the configure process of octave-2.1.63 thinks there is a native
function called 'round' when there doesn't seem to be one.

This results in compiling error:
  cc-1020 CC: ERROR File = lo-mappers.cc, Line = 83
   The identifier "round" is undefined.

      return round (x);
*****            ^

Looking in the config.h file (which is auto created during configure
process), there is:
  /* Define to 1 if you have the `round' function. */
#define HAVE_ROUND 1

Looking in the configure.in file, which I suspect check for the 'round'
AC_CHECK_FUNCS(atexit basename bcopy bzero dup2 endgrent endpwent execvp \
   fcntl fork getcwd getegid geteuid getgid getgrent getgrgid \
   getgrnam getpgrp getpid getppid getpwent getpwuid gettimeofday \
   getuid getwd _kbhit kill link localtime_r lstat memmove mkdir \
   mkfifo mkstemp on_exit pipe poll putenv raise readlink rename \
   rindex rmdir round select setgrent setpwent setvbuf sigaction \
   siglongjmp sigpending sigprocmask sigsuspend stat strcasecmp strdup \
   strerror strftime stricmp strncasecmp strnicmp strptime symlink \
   tempnam umask unlink usleep vfprintf vsprintf vsnprintf waitpid)

What went wrong that resulted in HAVE_ROUND being defined as 1?

I'm on an SGI IRIX-6.5 system, with MIPSpro compilers 7.4.

I know that when I manually COMMENT OUT the "#define HAVE_ROUND 1", the
compilation does not break at the above mentioned point anymore.



Clinton Chee
Computational Scientist
High Performance Computing Unit
Room 2075, Red Centre
University of New South Wales
Australia 2035
chee at parallel stop hpc stop unsw stop edu stop au
Tel: 61 2 9385 6915