Tweak the overflow checks in integer division functions to complain if the

machine produces zero (rather than the more usual minimum-possible-integer)
for the only possible overflow case.  This has been seen to occur for at least
some word widths on some hardware, and it's cheap enough to check for
everywhere.  Per Peter's analysis of buildfarm reports.

This could be back-patched, but in the absence of any gripes from the field
I doubt it's worth the trouble.
This commit is contained in:
Tom Lane 2008-10-05 23:18:37 +00:00
parent 1e4b03847c
commit 8acfc7594d
2 changed files with 20 additions and 14 deletions

View File

@ -8,7 +8,7 @@
*
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/utils/adt/int.c,v 1.82 2008/06/17 19:10:56 tgl Exp $
* $PostgreSQL: pgsql/src/backend/utils/adt/int.c,v 1.83 2008/10/05 23:18:37 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -735,9 +735,10 @@ int4div(PG_FUNCTION_ARGS)
/*
* Overflow check. The only possible overflow case is for arg1 = INT_MIN,
* arg2 = -1, where the correct result is -INT_MIN, which can't be
* represented on a two's-complement machine.
* represented on a two's-complement machine. Most machines produce
* INT_MIN but it seems some produce zero.
*/
if (arg2 == -1 && arg1 < 0 && result < 0)
if (arg2 == -1 && arg1 < 0 && result <= 0)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("integer out of range")));
@ -863,9 +864,10 @@ int2div(PG_FUNCTION_ARGS)
/*
* Overflow check. The only possible overflow case is for arg1 =
* SHRT_MIN, arg2 = -1, where the correct result is -SHRT_MIN, which can't
* be represented on a two's-complement machine.
* be represented on a two's-complement machine. Most machines produce
* SHRT_MIN but it seems some produce zero.
*/
if (arg2 == -1 && arg1 < 0 && result < 0)
if (arg2 == -1 && arg1 < 0 && result <= 0)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("smallint out of range")));
@ -1041,9 +1043,10 @@ int42div(PG_FUNCTION_ARGS)
/*
* Overflow check. The only possible overflow case is for arg1 = INT_MIN,
* arg2 = -1, where the correct result is -INT_MIN, which can't be
* represented on a two's-complement machine.
* represented on a two's-complement machine. Most machines produce
* INT_MIN but it seems some produce zero.
*/
if (arg2 == -1 && arg1 < 0 && result < 0)
if (arg2 == -1 && arg1 < 0 && result <= 0)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("integer out of range")));

View File

@ -7,7 +7,7 @@
* Portions Copyright (c) 1994, Regents of the University of California
*
* IDENTIFICATION
* $PostgreSQL: pgsql/src/backend/utils/adt/int8.c,v 1.70 2008/06/17 19:10:56 tgl Exp $
* $PostgreSQL: pgsql/src/backend/utils/adt/int8.c,v 1.71 2008/10/05 23:18:37 tgl Exp $
*
*-------------------------------------------------------------------------
*/
@ -608,9 +608,10 @@ int8div(PG_FUNCTION_ARGS)
/*
* Overflow check. The only possible overflow case is for arg1 =
* INT64_MIN, arg2 = -1, where the correct result is -INT64_MIN, which
* can't be represented on a two's-complement machine.
* can't be represented on a two's-complement machine. Most machines
* produce INT64_MIN but it seems some produce zero.
*/
if (arg2 == -1 && arg1 < 0 && result < 0)
if (arg2 == -1 && arg1 < 0 && result <= 0)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("bigint out of range")));
@ -830,9 +831,10 @@ int84div(PG_FUNCTION_ARGS)
/*
* Overflow check. The only possible overflow case is for arg1 =
* INT64_MIN, arg2 = -1, where the correct result is -INT64_MIN, which
* can't be represented on a two's-complement machine.
* can't be represented on a two's-complement machine. Most machines
* produce INT64_MIN but it seems some produce zero.
*/
if (arg2 == -1 && arg1 < 0 && result < 0)
if (arg2 == -1 && arg1 < 0 && result <= 0)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("bigint out of range")));
@ -1008,9 +1010,10 @@ int82div(PG_FUNCTION_ARGS)
/*
* Overflow check. The only possible overflow case is for arg1 =
* INT64_MIN, arg2 = -1, where the correct result is -INT64_MIN, which
* can't be represented on a two's-complement machine.
* can't be represented on a two's-complement machine. Most machines
* produce INT64_MIN but it seems some produce zero.
*/
if (arg2 == -1 && arg1 < 0 && result < 0)
if (arg2 == -1 && arg1 < 0 && result <= 0)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("bigint out of range")));