Merge xiphis.org:/usr/home/antony/work2/mysql-4.1

into  xiphis.org:/usr/home/antony/work2/merge-5.0

4 files changed, 92 insertions, 42 deletions

diff --git a/strings/Makefile.am b/strings/Makefile.am
index ad652862222..c43cf0f290a 100644
--- a/strings/Makefile.am
+++ b/strings/Makefile.am
@@ -22,7 +22,7 @@ pkglib_LIBRARIES =	libmystrings.a
 # Exact one of ASSEMBLER_X
 if ASSEMBLER_x86
 ASRCS		= strings-x86.s longlong2str-x86.s my_strtoll10-x86.s
-CSRCS		= bfill.c bmove.c bmove512.c bchange.c strxnmov.c int2str.c str2int.c r_strinstr.c strtod.c bcmp.c strtol.c strtoul.c strtoll.c strtoull.c llstr.c strnlen.c ctype.c ctype-simple.c ctype-mb.c ctype-big5.c ctype-cp932.c ctype-czech.c ctype-eucjpms.c ctype-euc_kr.c ctype-gb2312.c ctype-gbk.c ctype-sjis.c ctype-tis620.c ctype-ujis.c ctype-utf8.c ctype-ucs2.c ctype-uca.c ctype-win1250ch.c ctype-bin.c ctype-latin1.c my_vsnprintf.c xml.c decimal.c ctype-extra.c str_alloc.c
+CSRCS		= bfill.c bmove.c bmove512.c bchange.c strxnmov.c int2str.c str2int.c r_strinstr.c strtod.c bcmp.c strtol.c strtoul.c strtoll.c strtoull.c llstr.c strnlen.c ctype.c ctype-simple.c ctype-mb.c ctype-big5.c ctype-cp932.c ctype-czech.c ctype-eucjpms.c ctype-euc_kr.c ctype-gb2312.c ctype-gbk.c ctype-sjis.c ctype-tis620.c ctype-ujis.c ctype-utf8.c ctype-ucs2.c ctype-uca.c ctype-win1250ch.c ctype-bin.c ctype-latin1.c my_vsnprintf.c xml.c decimal.c ctype-extra.c str_alloc.c longlong2str_asm.c
 else
 if ASSEMBLER_sparc32
 # These file MUST all be on the same line!! Otherwise automake
@@ -45,7 +45,7 @@ EXTRA_DIST =		ctype-big5.c ctype-cp932.c ctype-czech.c ctype-eucjpms.c ctype-euc
 			ctype-gb2312.c ctype-gbk.c ctype-sjis.c ctype-utf8.c \
 			ctype-ucs2.c ctype-uca.c ctype-tis620.c ctype-ujis.c \
 			xml.c decimal.c strto.c strings-x86.s \
-			longlong2str.c longlong2str-x86.s \
+			longlong2str.c longlong2str-x86.s longlong2str_asm.c \
 			my_strtoll10.c my_strtoll10-x86.s \
 			strxmov.c bmove_upp.c strappend.c strcont.c strend.c \
 			strfill.c strcend.c is_prefix.c strstr.c strinstr.c \
diff --git a/strings/longlong2str-x86.s b/strings/longlong2str-x86.s
index fcc57810224..1840bab3f47 100644
--- a/strings/longlong2str-x86.s
+++ b/strings/longlong2str-x86.s
@@ -16,26 +16,26 @@
 # Optimized longlong2str function for Intel 80x86  (gcc/gas syntax) 
 # Some set sequences are optimized for pentuimpro II 
 
-	.file	"longlong2str.s"
-	.version "1.01"
+	.file	"longlong2str-x86.s"
+	.version "1.02"
 
 .text
 	.align 4
 
-.globl	longlong2str
-	.type	 longlong2str,@function
+.globl	longlong2str_with_dig_vector
+	.type	 longlong2str_with_dig_vector,@function
 	
-longlong2str:
+longlong2str_with_dig_vector:
 	subl $80,%esp
 	pushl %ebp
 	pushl %esi
 	pushl %edi
 	pushl %ebx
 	movl 100(%esp),%esi	# Lower part of val 
-	movl 104(%esp),%ebp	# Higher part of val 
-	movl 108(%esp),%edi	# get dst 
 	movl 112(%esp),%ebx	# Radix 
+	movl 104(%esp),%ebp	# Higher part of val 
 	movl %ebx,%eax
+	movl 108(%esp),%edi	# get dst 
 	testl %eax,%eax
 	jge .L144
 
@@ -69,6 +69,8 @@ longlong2str:
 
 .L150:
 	leal 92(%esp),%ecx	# End of buffer 
+	movl %edi, 108(%esp)    # Store possible modified dest
+	movl 116(%esp), %edi    # dig_vec_upper
 	jmp  .L155
 	.align 4
 
@@ -83,7 +85,7 @@ longlong2str:
 	divl %ebx
 	decl %ecx
 	movl %eax,%esi		# quotent in ebp:esi 
-	movb _dig_vec_upper(%edx),%al   # al is faster than dl 
+	movb (%edx,%edi),%al    # al is faster than dl 
 	movb %al,(%ecx)		# store value in buff 
 	.align 4
 .L155:
@@ -91,20 +93,22 @@ longlong2str:
 	ja .L153
 	testl %esi,%esi		# rest value 
 	jl .L153
-	je .L10_mov		# Ready 
+	je .L160		# Ready 
 	movl %esi,%eax
-	movl $_dig_vec_upper,%ebp
 	.align 4
 
 .L154:				# Do rest with integer precision 
 	cltd
 	divl %ebx
 	decl %ecx
-	movb (%edx,%ebp),%dl	# bh is always zero as ebx=radix < 36 
+	movb (%edx,%edi),%dl	# bh is always zero as ebx=radix < 36 
 	testl %eax,%eax
 	movb %dl,(%ecx)
 	jne .L154
 
+.L160:
+	movl 108(%esp),%edi	# get dst 
+	
 .L10_mov:
 	movl %ecx,%esi
 	leal 92(%esp),%ecx	# End of buffer 
@@ -129,7 +133,7 @@ longlong2str:
 	jmp .L165
 
 .Lfe3:
-	.size	 longlong2str,.Lfe3-longlong2str
+	.size	 longlong2str_with_dig_vector,.Lfe3-longlong2str_with_dig_vector
 
 #
 # This is almost equal to the above, except that we can do the final
@@ -137,9 +141,6 @@ longlong2str:
 #	
 
 	.align 4
-.Ltmp:
-        .long 0xcccccccd
-	.align 4
 	
 .globl	longlong10_to_str
 	.type	 longlong10_to_str,@function
@@ -202,8 +203,8 @@ longlong10_to_str:
 
 	# The following code uses some tricks to change division by 10 to
 	# multiplication and shifts
-	movl .Ltmp,%esi		# set %esi to 0xcccccccd
-	
+	movl $0xcccccccd,%esi
+		
 .L10_40:
         movl %ebx,%eax
         mull %esi
diff --git a/strings/longlong2str_asm.c b/strings/longlong2str_asm.c
new file mode 100644
index 00000000000..e38a8328b91
--- /dev/null
+++ b/strings/longlong2str_asm.c
@@ -0,0 +1,33 @@
+/* Copyright (C) 2000 MySQL AB
+
+   This program is free software; you can redistribute it and/or modify
+   it under the terms of the GNU General Public License as published by
+   the Free Software Foundation; either version 2 of the License, or
+   (at your option) any later version.
+
+   This program is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+   GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA */
+
+/*
+  Wrapper for longlong2str.s
+
+  We need this because the assembler code can't access the local variable
+  _dig_vector in a portable manner.
+*/
+
+#include <my_global.h>
+#include "m_string.h"
+
+extern char *longlong2str_with_dig_vector(longlong val,char *dst,int radix,
+                                          const char *dig_vector);
+
+char *longlong2str(longlong val,char *dst,int radix)
+{
+  return longlong2str_with_dig_vector(val, dst, radix, _dig_vec_upper);
+}
diff --git a/strings/my_strtoll10-x86.s b/strings/my_strtoll10-x86.s
index c04384667a7..1d8b8f2bbce 100644
--- a/strings/my_strtoll10-x86.s
+++ b/strings/my_strtoll10-x86.s
@@ -17,21 +17,8 @@
 # For documentation, check my_strtoll.c
 	
 	.file	"my_strtoll10-x86.s"
-	.version "01.01"
-.data
-	.align 32
-	.type	 lfactor,@object
-	.size	 lfactor,36
-lfactor:
-	.long 1
-	.long 10
-	.long 100
-	.long 1000
-	.long 10000
-	.long 100000
-	.long 1000000
-	.long 10000000
-	.long 100000000
+	.version "01.02"
+	
 .text
 	.align 4
 	
@@ -209,14 +196,16 @@ my_strtoll10:
 	jne .L500
 	cmpl -36(%ebp),%esi	# Test if string is less than 18 digits
 	jne .Lend_i_and_j
-	jmp .Lend3		# 18 digit string
+.L499:	
+	movl $1000000000,%eax	
+	jmp .Lgot_factor	# 18 digit string
 
 	# Handle the possible next to last digit and store in ecx
 .L500:
 	movb (%esi),%al
 	addb $-48,%al
 	cmpb $9,%al
-	ja .Lend3
+	ja .L499		# 18 digit string
 
 	incl %esi
 	movzbl %al,%ecx
@@ -315,14 +304,41 @@ my_strtoll10:
 .Lend_i_and_j:
 	movl %esi,%ecx
 	subl -12(%ebp),%ecx	# ecx= number of digits in second part
-	movl lfactor(,%ecx,4),%eax
-	jmp .L523
 
-	# Return -8(%ebp) * $1000000000 + edi
+	# Calculate %eax= 10 ** %cl, where %cl <= 8
+	# With an array one could do this with:
+	# movl 10_factor_table(,%ecx,4),%eax
+	# We calculate the table here to avoid problems in
+	# position independent code (gcc -pic)
+
+	cmpb  $3,%cl
+	ja    .L4_to_8
+	movl  $1000, %eax
+	je    .Lgot_factor	# %cl=3, eax= 1000
+	movl  $10, %eax
+	cmpb  $1,%cl		# %cl is here 0 - 2
+	je    .Lgot_factor	# %cl=1, eax= 10
+	movl  $100, %eax	
+	ja    .Lgot_factor	# %cl=2, eax=100
+	movl  $1, %eax		
+	jmp   .Lgot_factor	# %cl=0, eax=1
+
+.L4_to_8:			# %cl is here 4-8
+	cmpb  $5,%cl
+	movl  $100000, %eax
+	je   .Lgot_factor	# %cl=5, eax=100000
+	movl  $10000, %eax
+	jbe  .Lgot_factor	# %cl=4, eax=10000
+	movl  $10000000, %eax
+	cmpb  $7,%cl
+	je   .Lgot_factor	# %cl=7, eax=10000000
+	movl  $100000000, %eax	
+	ja   .Lgot_factor	# %cl=8, eax=100000000
+	movl  $1000000, %eax	# %cl=6, eax=1000000
+
+	# Return -8(%ebp) * %eax + edi
 	.p2align 4,,7
-.Lend3:
-	movl $1000000000,%eax
-.L523:
+.Lgot_factor:
 	mull -8(%ebp)
 	addl %edi,%eax
 	adcl $0,%edx