Bug#749710: workaround

2014-10-13 Thread Johannes Schauer
Control: block 764909 by -1

Hi,

On Thu, 29 May 2014 17:16:01 +0200 Johannes Schauer j.scha...@email.de wrote:
 a workaround is to
 
 export PYTHONHASHSEED=0
 
 as described in [1] which will then lead to a deterministic result.

unfortunately, while this is enough to create stable output within an
architecture, it varies between 32 and 64 bit architectures because the output
of the hash() function varies between 32 and 64 bit architectures. The
following produces different output depending on whether it is run on a 32 or
64 bit architecture:

PYTHONHASHSEED=0 python3 -c 'print(hash(a))'

Thus it is blocking #764909 on 32 bit architectures because the test cases were
created for 64 bit.

cheers, josch


--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#749710: workaround

2014-05-29 Thread Johannes Schauer
Hi,

a workaround is to

export PYTHONHASHSEED=0

as described in [1] which will then lead to a deterministic result.

cheers, josch

[1] https://docs.python.org/3/using/cmdline.html#envvar-PYTHONHASHSEED


--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org