Return to Question

I am trying to generate a hash code from two integer inputs. The approach outlined in

Combining Java hashcodes into a "master" hashcode

seems to work well for many input values. However, when one of the input integers is int.MinValue, the behavior seems less than ideal. Specifically I observe

int.MinValue * 1013 == int.MinValue int.MinValue * 1009 == int.MinValue 

but

int.MinValue * 2 == 0 int.MinValue * 20 == 0 

All of this is in an unchecked context.

I would naively (and wrongly) assume that int.MinValue * (something other than 1 or 0) would yield a new bit pattern different than int.MinValue or 0.

Questions

1. Why does multiplying int.MinValue by these constants yield int.MinValue (2 cases) or 0 (2 cases)?
2. Does the behavior of int.MinValue indicate a flaw in the hash algorithm?

I am trying to generate a hash code from two integer inputs. The approach outlined in

Combining Java hashcodes into a "master" hashcode

seems to work well for many input values. However, when one of the input integers is int.MinValue, the behavior seems less than ideal. Specifically I observe

int.MinValue * 1013 == int.MinValue int.MinValue * 1009 == int.MinValue 

but

int.MinValue * 2 == 0 int.MinValue * 20 == 0 

All of this is in an unchecked context.

I would naively (and wrongly) assume that int.MinValue * (something other than 1 or 0) would yield a new bit pattern different than int.MinValue or 0.

Questions

1. Why does multiplying int.MinValue by these constants yield int.MinValue (2 cases) or 0 (2 cases)?
2. Does the behavior of int.MinValue indicate a flaw in the hash algorithm?

I am trying to generate a hash code from two integer inputs. The approach outlined in

Combining Java hashcodes into a "master" hashcode

seems to work well for many input values. However, when one of the input integers is int.MinValue, the behavior seems less than ideal. Specifically I observe

int.MinValue * 1013 == int.MinValue

int.MinValue * 1009 == int.MinValue

but

int.MinValue * 2 == 0

int.MinValue * 20 == 0

All of this is in an unchecked context.

I would naively (and wrongly) assume that int.MinValue * (something other than 1 or 0) would yield a new bit pattern different than int.MinValue or 0.

Questions

1. Why does multiplying int.MinValue by these constants yield int.MinValue (2 cases) or 0 (2 cases)?
2. Does the behavior of int.MinValue indicate a flaw in the hash algorithm?

I am trying to generate a hash code from two integer inputs. The approach outlined in

Combining Java hashcodes into a "master" hashcode

seems to work well for many input values. However, when one of the input integers is int.MinValue, the behavior seems less than ideal. Specifically I observe

int.MinValue * 1013 == int.MinValue int.MinValue * 1009 == int.MinValue 

but

int.MinValue * 2 == 0 int.MinValue * 20 == 0 

All of this is in an unchecked context.

I would naively (and wrongly) assume that int.MinValue * (something other than 1 or 0) would yield a new bit pattern different than int.MinValue or 0.

Questions

1. Why does multiplying int.MinValue by these constants yield int.MinValue (2 cases) or 0 (2 cases)?
2. Does the behavior of int.MinValue indicate a flaw in the hash algorithm?