Fix for bug #31207: Test "join_nested" shows different strategy on IA64

CPUs / Intel's ICC compile

The bug is a combination of two problems:

1. IA64/ICC MySQL binaries use glibc's qsort(), not the one in mysys.

2. The order relation implemented by join_tab_cmp() is not transitive,
i.e. it is possible to choose such a, b and c that (a < b) && (b < c)
but (c < a). This implies that result of a sort using the relation
implemented by join_tab_cmp() depends on the order in which
elements are compared, i.e. the result is implementation-specific. Since
choose_plan() uses qsort() to pre-sort the
join tables using join_tab_cmp() as a compare function, the results of
the sorting may vary depending on qsort() implementation.

It is neither possible nor important to implement a better ordering
algorithm in join_tab_cmp(). Therefore the only way to fix it is to
force our own qsort() to be used by renaming it to my_qsort(), so we don't depend
on linker to decide that.

This patch also "fixes" bug #20530: qsort redefinition violates the
standard.


include/my_sys.h:
  Renamed qsort() and qsort2() to my_qsort() and my_qsort2(). Since
  previously we relied on stdlib.h to provide a declaration for qsort(), a
  separate declaration for my_qsort() is now required.
libmysql/Makefile.shared:
  Added mf_qsort.c to libmysql, since my_lib.c now uses my_qsort() instead of qsort().
myisam/ft_boolean_search.c:
  Replaced qsort2() with my_qsort2().
myisam/ft_nlq_search.c:
  Replaced qsort2() with my_qsort2().
myisam/myisampack.c:
  Replaced qsort() with my_qsort().
myisam/sort.c:
  Replaced qsort2() with my_qsort2().
mysys/mf_keycache.c:
  Replaced qsort() with my_qsort().
mysys/mf_qsort.c:
  Renamed qsort() to my_qsort() and qsort2() to my_qsort2().
mysys/mf_sort.c:
  Replaced qsort2() with my_qsort2().
mysys/my_lib.c:
  Replaced qsort() with my_qsort().
mysys/queues.c:
  Replaced qsort2() with my_qsort2().
sql/item_cmpfunc.cc:
  Replaced qsort2() with my_qsort2().
sql/item_cmpfunc.h:
  Replaced qsort2() with my_qsort2().
sql/opt_range.cc:
  Replaced qsort() with my_qsort().
sql/records.cc:
  Replaced qsort() with my_qsort().
sql/sql_acl.cc:
  Replaced qsort() with my_qsort().
sql/sql_array.h:
  Replaced qsort() with my_qsort().
sql/sql_help.cc:
  Replaced qsort() with my_qsort().
sql/sql_select.cc:
  Replaced qsort() with my_qsort().
sql/examples/ha_tina.cc:
  Replaced qsort() with my_qsort().
sql/sql_table.cc:
  Replaced qsort() with my_qsort().

1 files changed, 15 insertions, 3 deletions

diff --git a/sql/sql_select.cc b/sql/sql_select.cc
index 1d11f23d854..b519c06420f 100644
--- a/sql/sql_select.cc
+++ b/sql/sql_select.cc
@@ -3658,7 +3658,7 @@ update_ref_and_keys(THD *thd, DYNAMIC_ARRAY *keyuse,JOIN_TAB *join_tab,
   {
     KEYUSE key_end,*prev,*save_pos,*use;
 
-    qsort(keyuse->buffer,keyuse->elements,sizeof(KEYUSE),
+    my_qsort(keyuse->buffer,keyuse->elements,sizeof(KEYUSE),
 	  (qsort_cmp) sort_keyuse);
 
     bzero((char*) &key_end,sizeof(key_end));    /* Add for easy testing */
@@ -4371,8 +4371,9 @@ choose_plan(JOIN *join, table_map join_tables)
       Apply heuristic: pre-sort all access plans with respect to the number of
       records accessed.
   */
-  qsort(join->best_ref + join->const_tables, join->tables - join->const_tables,
-        sizeof(JOIN_TAB*), straight_join?join_tab_cmp_straight:join_tab_cmp);
+  my_qsort(join->best_ref + join->const_tables,
+           join->tables - join->const_tables, sizeof(JOIN_TAB*),
+           straight_join ? join_tab_cmp_straight : join_tab_cmp);
   
   if (straight_join)
   {
@@ -4421,6 +4422,17 @@ choose_plan(JOIN *join, table_map join_tables)
     ptr1 pointer to first JOIN_TAB object
     ptr2 pointer to second JOIN_TAB object
 
+  NOTES
+    The order relation implemented by join_tab_cmp() is not transitive,
+    i.e. it is possible to choose such a, b and c that (a < b) && (b < c)
+    but (c < a). This implies that result of a sort using the relation
+    implemented by join_tab_cmp() depends on the order in which
+    elements are compared, i.e. the result is implementation-specific.
+    Example:
+      a: dependent = 0x0 table->map = 0x1 found_records = 3 ptr = 0x907e6b0
+      b: dependent = 0x0 table->map = 0x2 found_records = 3 ptr = 0x907e838
+      c: dependent = 0x6 table->map = 0x10 found_records = 2 ptr = 0x907ecd0
+     
   RETURN
     1  if first is bigger
     -1 if second is bigger