[m-rev.] Fix 64bit general/string_format_o testcase

[Julien Fischer]

On Sun, 7 Jan 2007, Keri Harris wrote:

> The general/string_format_o testcase contains a bug in
> string_format_o.exp2: 'string.format("%#.0o", 0, String)' should
> return  'String = "0"'. However the general/string_format_o testcase
> indicates otherwise:
>
> $ grep '%#\.0o' string_format_x.exp
>     %#.0o:''
>
> string.format relies on sprintf and the C standard for sprintf mentions:
>
> - #: For o conversion, it increases the precision, if and only if necessary,
> to force the ÿÿrst digit of the result to be a zero (if the value and
> precision are both 0, a single 0 is printed).

That particular expected output was a victim of the fact that quite a
lot of C library implementations get that particular conversion wrong.
(including I've just noticed the machine I'm writing this on :-( )

Thanks for the fix.

Julien.