From Earl.Ruby@CellNet.com

# uname -a
SunOS fserv1 5.5.1 Generic_103640-17 sun4u sparc SUNW,Ultra-Enterprise

I am using most of the default options in the install. The only differences
being the default installation path (/home/local/perl rather than
/usr/local) and the use of gcc rather than cc. (I tried cc and had the same
results.) The install would lock up immediately after the line "You do not
have union semun in <sys/sem.h>."

Checking if your memcmp() can compare relative magnitude...
Yes, it can.

select() found.

semctl() found.

semget() found.

semop() found.

You have the full sem*(2) library.

You do not have union semun in <sys/sem.h>.

I tried the install on a Sun Enterprise Ultra 3000 server running Solaris
2.6 and it worked​:

# uname -a
SunOS rdserv1 5.6 Generic_105181-15 sun4u sparc SUNW,Ultra-Enterprise

The only difference between the two machines is the version of Solaris
running and that the first machine has 256MB RAM and the second machine has
1.25GB RAM. Other than that, they're identical.

Any ideas what might be wrong here?

Earl C. Ruby III (eruby@​cellnet.com) Database Engineer, CellNet San Ramon
925/543-1072 (desk) 925/766-2599 (cell/beeper) 114 on NexTel 2-way radio

From [Unknown Contact. See original ticket]

Earl C . Ruby III <Earl.Ruby@​CellNet.com> writes​:

I am trying to install the latest stable Perl release, perl5.005_03. I am
installing it on a Sun Enterprise Ultra 3000 server running Solaris 2.5.1​:

# uname -a
SunOS fserv1 5.5.1 Generic_103640-17 sun4u sparc SUNW,Ultra-Enterprise

I am using most of the default options in the install. The only differences
being the default installation path (/home/local/perl rather than
/usr/local) and the use of gcc rather than cc. (I tried cc and had the same
results.) The install would lock up immediately after the line "You do not
have union semun in <sys/sem.h>."

Most likely the one of machines had actually used some of these SysV
shared memory things. Once upon a time that fact that the id of 1st
such after a machine booted was 0 (which perl treated as false)
was a problem - but I thought that just made tests fail.
--
Nick Ing-Simmons