* configure.in (stdint.h): check if presence.

* configure.in (uint32_t): check if defined.

* string.c (hash): fix for portability.  [ruby-dev:34020]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@15760 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

4 files changed, 125 insertions, 10 deletions

diff --git a/ChangeLog b/ChangeLog
index 05f8170fff..6487b72ead 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+Thu Mar 13 01:45:25 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
+
+	* configure.in (stdint.h): check if presence.
+
+	* configure.in (uint32_t): check if defined.
+
+	* string.c (hash): fix for portability.  [ruby-dev:34020]
+
 Wed Mar 12 17:33:34 2008  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 
 	* object.c (rb_cstr_to_dbl): fix for a mere underscore.
diff --git a/configure.in b/configure.in
index c17821011f..a33b5a9fc7 100644
--- a/configure.in
+++ b/configure.in
@@ -582,7 +582,7 @@ AC_CHECK_HEADERS(stdlib.h string.h unistd.h limits.h sys/file.h sys/ioctl.h sys/
 		 fcntl.h sys/fcntl.h sys/select.h sys/time.h sys/times.h sys/param.h\
 		 syscall.h pwd.h grp.h a.out.h utime.h memory.h direct.h sys/resource.h \
 		 sys/mkdev.h sys/utime.h xti.h netinet/in_systm.h float.h ieeefp.h pthread.h \
-		 ucontext.h intrinsics.h langinfo.h locale.h)
+		 ucontext.h intrinsics.h langinfo.h locale.h stdint.h)
 
 dnl Check additional types.
 AC_CHECK_SIZEOF(rlim_t, 0, [
@@ -620,6 +620,18 @@ AC_CHECK_TYPES(struct timespec)
 
 AC_CHECK_TYPE(fd_mask, [AC_DEFINE(HAVE_RB_FD_INIT, 1)])
 
+test ${rb_cv_type_uint32_t+set} && ac_cv_type_uint32_t=yes
+AC_CHECK_TYPE(uint32_t)
+if test ${ac_cv_type_uint32_t} != yes; then
+    AC_CACHE_CHECK([unsigned 32bit int],
+	rb_cv_type_uint32_t,
+	[for type in short int long; do
+	    AC_COMPILE_IFELSE(AC_LANG_BOOL_COMPILE_TRY([], [sizeof($type) == 4]),
+		[rb_cv_type_uint32_t=$type; break], [])
+	done])
+    AC_DEFINE(uint32_t, $rb_cv_type_uint32_t)
+fi
+
 AC_CACHE_CHECK(for stack end address, rb_cv_stack_end_address,
 [rb_cv_stack_end_address=no
 for addr in __libc_stack_end _SEND; do
diff --git a/string.c b/string.c
index 71cc090859..e64d440b77 100644
--- a/string.c
+++ b/string.c
@@ -25,6 +25,10 @@
 #include <unistd.h>
 #endif
 
+#if HAVE_STDINT_H
+#include <stdint.h>
+#endif
+
 VALUE rb_cString;
 VALUE rb_cSymbol;
 
@@ -1685,6 +1689,13 @@ rb_str_concat(VALUE str1, VALUE str2)
     return rb_str_append(str1, str2);
 }
 
+#if defined __i386__ || defined _M_IX86 || defined __ia64
+#define UNALIGNED_WORD_ACCESS 1
+#endif
+#ifndef UNALIGNED_WORD_ACCESS
+#define UNALIGNED_WORD_ACCESS 0
+#endif
+
 /* MurmurHash described in http://murmurhash.googlepages.com/ */
 unsigned int
 hash(const unsigned char * data, int len, unsigned int h)
@@ -1694,22 +1705,106 @@ hash(const unsigned char * data, int len, unsigned int h)
 
     h += 0xdeadbeef;
 
-    while(len >= 4) {
-	h += *(unsigned int *)data;
-	h *= m;
-	h ^= h >> r;
+    if (len >= 4) {
+#if !UNALIGNED_WORD_ACCESS
+	int align = (VALUE)data & 3;
+	if (align) {
+	    uint32_t t = 0, d = 0;
+	    int sl, sr, pack;
+
+	    switch (align) {
+#ifdef WORDS_BIGENDIAN
+	      case 1: t |= data[2];
+	      case 2: t |= data[1] << 8;
+	      case 3: t |= data[0] << 16;
+#else
+	      case 1: t |= data[2] << 16;
+	      case 2: t |= data[1] << 8;
+	      case 3: t |= data[0];
+#endif
+	    }
+
+#ifdef WORDS_BIGENDIAN
+	    t >>= (8 * align) - 8;
+#else
+	    t <<= (8 * align);
+#endif
+
+	    data += 4-align;
+	    len -= 4-align;
+
+	    sl = 8 * (4-align);
+	    sr = 8 * align;
+
+	    while (len >= 4) {
+		d = *(uint32_t *)data;
+#ifdef WORDS_BIGENDIAN
+		t = (t << sr) | (d >> sl);
+#else
+		t = (t >> sr) | (d << sl);
+#endif
+		h += t;
+		h *= m;
+		h ^= h >> r;
+		t = d;
 
-	data += 4;
-	len -= 4;
+		data += 4;
+		len -= 4;
+	    }
+
+	    pack = len < align ? len : align;
+	    d = 0;
+	    switch (pack) {
+#ifdef WORDS_BIGENDIAN
+	      case 3: d |= data[2] << 8;
+	      case 2: d |= data[1] << 16;
+	      case 1: d |= data[0] << 24;
+	      case 0:
+		h += (t << sr) | (d >> sl);
+#else
+	      case 3: d |= data[2] << 16;
+	      case 2: d |= data[1] << 8;
+	      case 1: d |= data[0];
+	      case 0:
+		h += (t >> sr) | (d << sl);
+#endif
+		h *= m;
+		h ^= h >> r;
+	    }
+
+	    data += pack;
+	    len -= pack;
+	}
+	else
+#endif
+	{
+	    do {
+		h += *(uint32_t *)data;
+		h *= m;
+		h ^= h >> r;
+
+		data += 4;
+		len -= 4;
+	    } while (len >= 4);
+	}
     }
 
     switch(len) {
+#ifdef WORDS_BIGENDIAN
+      case 3:
+	h += data[2] << 8;
+      case 2:
+	h += data[1] << 16;
+      case 1:
+	h += data[0] << 24;
+#else
       case 3:
 	h += data[2] << 16;
       case 2:
 	h += data[1] << 8;
       case 1:
 	h += data[0];
+#endif
 	h *= m;
 	h ^= h >> r;
     };
diff --git a/version.h b/version.h
index fb3967317e..032939d883 100644
--- a/version.h
+++ b/version.h
@@ -1,7 +1,7 @@
 #define RUBY_VERSION "1.9.0"
-#define RUBY_RELEASE_DATE "2008-03-12"
+#define RUBY_RELEASE_DATE "2008-03-13"
 #define RUBY_VERSION_CODE 190
-#define RUBY_RELEASE_CODE 20080312
+#define RUBY_RELEASE_CODE 20080313
 #define RUBY_PATCHLEVEL 0
 
 #define RUBY_VERSION_MAJOR 1
@@ -9,7 +9,7 @@
 #define RUBY_VERSION_TEENY 0
 #define RUBY_RELEASE_YEAR 2008
 #define RUBY_RELEASE_MONTH 3
-#define RUBY_RELEASE_DAY 12
+#define RUBY_RELEASE_DAY 13
 
 #ifdef RUBY_EXTERN
 RUBY_EXTERN const char ruby_version[];