In message <199501062256.QAA26179@dax.Stoner.COM>,
Bryan Curnutt writes:
>SunOS 4.1.3 has compile problems with the "NO_SETGROUP" stuff in
>wrapper.c, including "POSIX_GID undefined". Adding -DNO_SETGROUP
>to the compile flags gets around this
I changed the test in wrapper.c to include the setgroups only if
SETGROUPS is defined, and I added setgroups to the posix
WRAPPER_FLAGS.
>though I'd feel better on
>general principal about making the setgroups stuff work even on
>non-POSIX systems.
I would feel better as well, but it can't. The euid has to be root to
have setgroups work. Under non-posix systems the euid becomes the
majordomo user immediately, there is never a location where the euid
is root.
>The "install-scripts" target in the Makefile tries to copy a file
>named "test" to the W_BIN directory, but no file named "test" exists
>in the distribution, so "make install" aborts with a fatal error.
>Removing the word "test" from the line in the Makefile fixes this.
Oops, fixed. The test file wasn't being checked out due to a missing
version tag.
>Sun's C compiler (the bundled one) also warns "illegal combination of
>pointer and integer" for line 65 of wrapper.c, which looks like
>
> if (STRCHR(argv[1], '/') != (char *) NULL) {
>
>Adding "#include <string.h>" or "char *strchr();" inside the "#ifndef
>STRCHR" fixes this for SunOS, but probably breaks it for other systems.
I think I will let the bundled compiler complain about that since I am
so close to release, and I don't have a copy of that compiler here to
look for unique CPP defines.
-- John
John Rouillard
Senior Systems Administrator IDD Information Services
rouilj@dstar.iddis.com Waltham, MA (617) 890-1576 x103
Senior Systems Consultant (SERL Project) University of Massachusetts at Boston
rouilj@cs.umb.edu (preferred) Boston, MA, (617) 287-6480
===============================================================================
My employers don't acknowledge my existence much less my opinions.
References:
|
|
