Adjust string comparison so that only bitwise-equal strings are considered

equal: if strcoll claims two strings are equal, check it with strcmp, and
sort according to strcmp if not identical.  This fixes inconsistent
behavior under glibc's hu_HU locale, and probably under some other locales
as well.  Also, take advantage of the now-well-defined behavior to speed up
texteq, textne, bpchareq, bpcharne: they may as well just do a bitwise
comparison and not bother with strcoll at all.

NOTE: affected databases may need to REINDEX indexes on text columns to be
sure they are self-consistent.

diff --git a/src/backend/access/hash/hashfunc.c b/src/backend/access/hash/hashfunc.c
index 2ffca5efe6a7ea8069ace1ba494f195dc22a9e03..6cf9eb6eee241f3607ffa3c4ebd1faa36fdb0a70 100644 (file)
--- a/src/backend/access/hash/hashfunc.c
+++ b/src/backend/access/hash/hashfunc.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/access/hash/hashfunc.c,v 1.45 2005/10/15 02:49:08 momjian Exp $
+ *   $PostgreSQL: pgsql/src/backend/access/hash/hashfunc.c,v 1.46 2005/12/22 22:50:00 tgl Exp $
  *
  * NOTES
  *   These functions are stored in pg_amproc.  For each operator class
@@ -138,9 +138,9 @@ hashtext(PG_FUNCTION_ARGS)
    Datum       result;
 
    /*
-    * Note: this is currently identical in behavior to hashvarlena, but it
-    * seems likely that we may need to do something different in non-C
-    * locales.  (See also hashbpchar, if so.)
+    * Note: this is currently identical in behavior to hashvarlena, but
+    * keep it as a separate function in case we someday want to do something
+    * different in non-C locales.  (See also hashbpchar, if so.)
     */
    result = hash_any((unsigned char *) VARDATA(key),
                      VARSIZE(key) - VARHDRSZ);

diff --git a/src/backend/utils/adt/varchar.c b/src/backend/utils/adt/varchar.c
index 1377e7cc6d2bd9c6b8907c655fd0bb282d94cafd..8fd46531b849c6536b3375bef51e6581362c3e1e 100644 (file)
--- a/src/backend/utils/adt/varchar.c
+++ b/src/backend/utils/adt/varchar.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/utils/adt/varchar.c,v 1.113 2005/10/15 02:49:30 momjian Exp $
+ *   $PostgreSQL: pgsql/src/backend/utils/adt/varchar.c,v 1.114 2005/12/22 22:50:00 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -614,11 +614,14 @@ bpchareq(PG_FUNCTION_ARGS)
    len1 = bcTruelen(arg1);
    len2 = bcTruelen(arg2);
 
-   /* fast path for different-length inputs */
+   /*
+    * Since we only care about equality or not-equality, we can avoid all
+    * the expense of strcoll() here, and just do bitwise comparison.
+    */
    if (len1 != len2)
        result = false;
    else
-       result = (varstr_cmp(VARDATA(arg1), len1, VARDATA(arg2), len2) == 0);
+       result = (strncmp(VARDATA(arg1), VARDATA(arg2), len1) == 0);
 
    PG_FREE_IF_COPY(arg1, 0);
    PG_FREE_IF_COPY(arg2, 1);
@@ -638,11 +641,14 @@ bpcharne(PG_FUNCTION_ARGS)
    len1 = bcTruelen(arg1);
    len2 = bcTruelen(arg2);
 
-   /* fast path for different-length inputs */
+   /*
+    * Since we only care about equality or not-equality, we can avoid all
+    * the expense of strcoll() here, and just do bitwise comparison.
+    */
    if (len1 != len2)
        result = true;
    else
-       result = (varstr_cmp(VARDATA(arg1), len1, VARDATA(arg2), len2) != 0);
+       result = (strncmp(VARDATA(arg1), VARDATA(arg2), len1) != 0);
 
    PG_FREE_IF_COPY(arg1, 0);
    PG_FREE_IF_COPY(arg2, 1);
@@ -789,7 +795,9 @@ bpchar_smaller(PG_FUNCTION_ARGS)
  * bpchar needs a specialized hash function because we want to ignore
  * trailing blanks in comparisons.
  *
- * XXX is there any need for locale-specific behavior here?
+ * Note: currently there is no need for locale-specific behavior here,
+ * but if we ever change the semantics of bpchar comparison to trust
+ * strcoll() completely, we'd need to do something different in non-C locales.
  */
 Datum
 hashbpchar(PG_FUNCTION_ARGS)

diff --git a/src/backend/utils/adt/varlena.c b/src/backend/utils/adt/varlena.c
index 69544ea90f62851b1008091e6f66be4f17642e92..20b40fe2e5da54d4953ce872931e82ad6a41a3eb 100644 (file)
--- a/src/backend/utils/adt/varlena.c
+++ b/src/backend/utils/adt/varlena.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *   $PostgreSQL: pgsql/src/backend/utils/adt/varlena.c,v 1.141 2005/11/22 18:17:23 momjian Exp $
+ *   $PostgreSQL: pgsql/src/backend/utils/adt/varlena.c,v 1.142 2005/12/22 22:50:00 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -938,6 +938,15 @@ varstr_cmp(char *arg1, int len1, char *arg2, int len2)
 
        result = strcoll(a1p, a2p);
 
+       /*
+        * In some locales strcoll() can claim that nonidentical strings are
+        * equal.  Believing that would be bad news for a number of reasons,
+        * so we follow Perl's lead and sort "equal" strings according to
+        * strcmp().
+        */
+       if (result == 0)
+           result = strcmp(a1p, a2p);
+
        if (a1p != a1buf)
            pfree(a1p);
        if (a2p != a2buf)
@@ -984,11 +993,15 @@ texteq(PG_FUNCTION_ARGS)
    text       *arg2 = PG_GETARG_TEXT_P(1);
    bool        result;
 
-   /* fast path for different-length inputs */
+   /*
+    * Since we only care about equality or not-equality, we can avoid all
+    * the expense of strcoll() here, and just do bitwise comparison.
+    */
    if (VARSIZE(arg1) != VARSIZE(arg2))
        result = false;
    else
-       result = (text_cmp(arg1, arg2) == 0);
+       result = (strncmp(VARDATA(arg1), VARDATA(arg2),
+                         VARSIZE(arg1) - VARHDRSZ) == 0);
 
    PG_FREE_IF_COPY(arg1, 0);
    PG_FREE_IF_COPY(arg2, 1);
@@ -1003,11 +1016,15 @@ textne(PG_FUNCTION_ARGS)
    text       *arg2 = PG_GETARG_TEXT_P(1);
    bool        result;
 
-   /* fast path for different-length inputs */
+   /*
+    * Since we only care about equality or not-equality, we can avoid all
+    * the expense of strcoll() here, and just do bitwise comparison.
+    */
    if (VARSIZE(arg1) != VARSIZE(arg2))
        result = true;
    else
-       result = (text_cmp(arg1, arg2) != 0);
+       result = (strncmp(VARDATA(arg1), VARDATA(arg2),
+                         VARSIZE(arg1) - VARHDRSZ) != 0);
 
    PG_FREE_IF_COPY(arg1, 0);
    PG_FREE_IF_COPY(arg2, 1);