Do not use the Perl malloc in IRIX, suggested by

[perl5.git] / hints / irix_6.sh

diff --git a/hints/irix_6.sh b/hints/irix_6.sh
index 3344808..6f4ca17 100644 (file)
--- a/hints/irix_6.sh
+++ b/hints/irix_6.sh
@@ -32,6 +32,14 @@
 # Don't bother with -n32 unless you have the 7.1 or later compilers.
 #     But there's no quick and light-weight way to check in 6.2.
 
+# NOTE: some IRIX cc versions, e.g. 7.3.1.1m (try cc -version) have
+# been known to have issues (coredumps) when compiling perl.c.
+# If you've used -OPT:fast_io=ON and this happens, try removing it.
+# If that fails, or you didn't use that, then try adjusting other
+# optimization options (-LNO, -INLINE, -O3 to -O2, etcetera).
+# The compiler bug has been reported to SGI.
+# -- Allen Smith <easmith@beatrice.rutgers.edu>
+
 # Let's assume we want to use 'cc -n32' by default, unless the
 # necessary libm is missing (which has happened at least twice)
 case "$cc" in
@@ -40,7 +48,13 @@ case "$cc" in
     *) test -f /usr/lib32/libm.so && cc='cc -n32' ;;
     esac       
 esac
-test -z "$cc" && cc=cc
+
+cc=${cc:-cc}
+
+case "$cc" in
+*gcc*) ;;
+*) ccversion=`cc -version` ;;
+esac
 
 case "$use64bitint" in
 $define|true|[yY]*)
@@ -77,9 +91,19 @@ esac
 case "$cc" in
 *"cc -n32"*)
 
-       libscheck='case "`/usr/bin/file $xxx`" in
-*N32*) ;;
-*) xxx=/no/n32$xxx ;;
+       # If a library is requested to link against, make sure the
+       # objects in the library are of the same ABI we are compiling
+       # against. Albert Chin-A-Young <china@thewrittenword.com>
+       libscheck='case "$xxx" in
+*.a) /bin/ar p $xxx `/bin/ar t $xxx | /usr/bsd/head -1` >$$.o;
+  case "`/usr/bin/file $$.o`" in
+  *N32*) rm -f $$.o ;;
+  *) rm -f $$.o; xxx=/no/n32$xxx ;;
+  esac ;;
+*) case "`/usr/bin/file $xxx`" in
+  *N32*) ;;
+  *) xxx=/no/n32$xxx ;;
+  esac ;;
 esac'
 
        # NOTE: -L/usr/lib32 -L/lib32 are automatically selected by the linker
@@ -93,7 +117,7 @@ esac'
        libc='/usr/lib32/libc.so'
        plibpth='/usr/lib32 /lib32 /usr/ccs/lib'
        ;;
-*"cc -64")
+*"cc -64"*)
 
        loclibpth="$loclibpth /usr/lib64"
        libscheck='case "`/usr/bin/file $xxx`" in
@@ -138,12 +162,21 @@ esac
 
 # Settings common to both native compiler modes.
 case "$cc" in
-*"cc -n32"|*"cc -64")
+*"cc -n32"*|*"cc -64"*)
        ld=$cc
 
        # perl's malloc can return improperly aligned buffer
-       # usemymalloc='undef'
-malloc_cflags='ccflags="-DSTRICT_ALIGNMENT $ccflags"'
+       # which (under 5.6.0RC1) leads into really bizarre bus errors
+       # and freak test failures (lib/safe1 #18, for example),
+       # even more so with -Duse64bitall: for example lib/io_linenumtb.
+       # fails under the harness but succeeds when run separately,
+       # under make test pragma/warnings #98 fails, and lib/io_dir
+       # apparently coredumps (the last two don't happen under
+       # the harness.  Helmut Jarausch is seeing bus errors from
+        # miniperl, as was Scott Henry with snapshots from just before
+       # the RC1. --jhi
+       usemymalloc='undef'
+#malloc_cflags='ccflags="-DSTRICT_ALIGNMENT $ccflags"'
 
        nm_opt='-p'
        nm_so_opt='-p'
@@ -207,8 +240,10 @@ esac
 # Don't groan about unused libraries.
 ldflags="$ldflags -Wl,-woff,84"
 
-case "`uname -s`-`$cc -version 2>&1`" in
-IRIX64-*7.2.*) op_cflags='optimize=-O1' ;; # workaround for an optimizer bug
+# workaround for an optimizer bug
+case "`$cc -version 2>&1`" in
+*7.2.*)   op_cflags='optimize=-O1'; opmini_cflags='optimize=-O1' ;;
+*7.3.1.*) op_cflags='optimize=-O2'; opmini_cflags='optimize=-O2' ;;
 esac
 
 # We don't want these libraries.
@@ -299,3 +334,15 @@ EOCBU
 
 # The -n32 makes off_t to be 8 bytes, so we should have largefileness.
 
+# Until we figure out what to be probe for in Configure (ditto for hpux.sh)
+case "$usemorebits" in # Need to expand this now, then.
+$define|true|[yY]*) use64bitint="$define"; uselongdouble="$define" ;;
+esac
+case "$use64bitint" in
+$define|true|[yY]*) ;;
+*) d_casti32='undef' ;;
+esac
+
+# Helmut Jarausch reports that Perl's malloc is rather unusable
+# with IRIX, and SGI confirms the problem.
+usemymalloc=${usemymalloc:-false}