Horton

BUG #15519: Casting float4 into int4 gets the wrong sign instead of "integer out of range" error

The Post Office <noreply@postgresql.org>

View thread View on postgr.es
From: PG Bug reporting form <noreply@postgresql.org>
To: pgsql-bugs@lists.postgresql.org
Cc: victor@magic.io
Date: 2018-11-23T22:03:56Z
Lists: pgsql-bugs 
The following bug has been logged on the website:

Bug reference:      15519
Logged by:          Victor Petrovykh
Email address:      victor@magic.io
PostgreSQL version: 10.5
Operating system:   Gentoo Linux 4.14.20
Description:        

Offending examples:
SELECT ((2147483647::float4) - 1.0::float4)::int4;
SELECT ((2147483590::float4) - 1.0::float4)::int4;
SELECT ((2147483647::float4) + 1.0::float4)::int4;

They all produce the same result: -2147483648

I understand that a float4 cannot represent large integers with the same
precision as int4, that's OK. What surprised me is that instead of getting
an "overflow error" or "integer out of range" I simply got a negative result
for a value that is actually close to maximum int4. To contrast this, the
query:
SELECT ((2147483647::float4) + 200.0::float4)::int4;
The above produces the expected "ERROR:  integer out of range"

Commits

  1. Update additional float4/8 expected-output files.

  2. Fix float-to-integer coercions to handle edge cases correctly.

  3. Adjust new test case for more portability.