[Python-3000-checkins] r55865 - in python/branches/p3yk/Tools: i18n/msgfmt.py pybench/Arithmetic.py pybench/CommandLine.py pybench/Numbers.py pybench/systimes.py scripts/classfix.py scripts/fixcid.py scripts/ftpmirror.py scripts/linktree.py scrip

[Neal Norwitz]

On 6/10/07, georg.brandl <python-3000-checkins at python.org> wrote:
> Author: georg.brandl
> Date: Mon Jun 11 00:31:37 2007
> New Revision: 55865
>
> Log:
> Some octal literal fixes in Tools.
>
>
> Modified: python/branches/p3yk/Tools/pybench/Numbers.py
> ==============================================================================
> --- python/branches/p3yk/Tools/pybench/Numbers.py       (original)
> +++ python/branches/p3yk/Tools/pybench/Numbers.py       Mon Jun 11 00:31:37 2007
> ...
> -            1234567890L < 3456789012345L
> -            1234567890L > 3456789012345L
> -            1234567890L == 3456789012345L
> -            1234567890L > 3456789012345L
> -            1234567890L < 3456789012345L
> +            1234567890 < 3456789012345
> +            1234567890 > 3456789012345
> +            1234567890 == 3456789012345
> +            1234567890 > 3456789012345
> +            1234567890 < 3456789012345
> ...

This changes the semantics for this test on 64-bit boxes.  Well,
kinda.  It was comparing between ints and longs before this change.
At least, that was the original intent.  I'm not sure how much that
even made sense since ints and longs were unified.

I assume there are str/uni tests in pybench as well.  These won't make
sense once strings and unicode are merged.

n