[daip] AIPS Installation

[Eric Greisen]

Greg Castanon writes:

 > In file included from
 > /pkg/gnu/lib/gcc-lib/sparc-sun-solaris2.5/2.7.2/include/time.h:90,
 >                  from /usr/include/sys/time.h:410,
 >                  from /usr/include/sys/select.h:17,
 >                  from
 > /pkg/gnu/lib/gcc-lib/sparc-sun-solaris2.5/2.7.2/include/sys/types.h:278,
 >                  from F2PS.c:2:
 > /usr/include/sys/siginfo.h:74: parse error before `pthread_attr_t'
 > /usr/include/sys/siginfo.h:74: warning: no semicolon at end of struct or
 > union
 > /usr/include/sys/siginfo.h:76: parse error before `}'

If I understand this out code asks for sys/types.h and that somehow
loads
  /pkg/gnu/lib/gcc-lib/sparc-sun-solaris2.5/2.7.2/include/sys/types.h
which loads sys/select.h which loads sys/time.h which loads
  /pkg/gnu/lib/gcc-lib/sparc-sun-solaris2.5/2.7.2/include/time.h

and that loads sys/siginfo.h at line 90 and that has errors.  Try
looking at that file /usr/include/sys/siginfo.h around line 74

Our Solaris file seems okay and says 

#ifndef _SIGEVENT
#define _SIGEVENT
struct sigevent {
        int             sigev_notify;   /* notification mode */
        int             sigev_signo;    /* signal number */
        union sigval    sigev_value;    /* signal value */
        void            (*sigev_notify_function)(union sigval);
        pthread_attr_t  *sigev_notify_attributes;
        int             __sigev_pad2;
};
#endif  /* _SIGEVENT */

where the pthread line is line 74.  What version of GNU compiler do
you have?  2.7.2 is not one I recognize and would be antique.

 > Strangely, PRINTENV compiled.  NEWEST, F2TEXT and F2PS didn't.
 > 
 > Direct compilation of the files produces the same results, and the
 > system doesn't seem to recognize AIGREP as a command.

    You must have $SYSUNIX and $SYSLOCAL in your path to find these -

source $AIPS_ROOT/LOGIN.CSH    or . $AIPS_ROOT/LOGIN.SH

to make this happen.

 > 
 > I'm beginning to wonder if working on a strange platform (This is,
 > sadly, my first experience with unix) is the best of ideas.  Is there
 > an easier way than AIPS to decode the data in the files downloaded
 > from the VLA?  I'm trying to implement a new data interpolation
 > algorithm...it means I don't necessarily need the data processed, I
 > just need the data.  Any suggestions?

You could get the NRAO data analysts to read the data in with FILLM
and write it back out with FITTP.  This would put the data into FITS
format but I suspect that that would be only a bit less opaque to you
than the very very peculiar form the data are in off the real-time
system.

I suspect that there are problems with the OS installation beyond the
silly behavior of removing the better compilers.  Perhap the GNU
version is really old:

gcc --version

will tell a lot.

Eric Greisen