On Wed, Jul 2, 2008 at 4:42 AM, chromatic <[EMAIL PROTECTED]> wrote:
> On Tuesday 01 July 2008 20:05:54 [EMAIL PROTECTED] wrote:
>
>> Modified:
>> branches/no_builtin_methods/src/builtin.c
>> branches/no_builtin_methods/src/ops/io.ops
>> branches/no_builtin_methods/src/ops/ops.num
>> branches/no_builtin_methods/t/pmc/builtin.t
>>
>> Log:
>> first pass, delete say builtin support and create say
>> opcodes instead.
>>
>>
>> Modified: branches/no_builtin_methods/src/ops/io.ops
>> ===========================================================================
>>=== --- branches/no_builtin_methods/src/ops/io.ops (original)
>> +++ branches/no_builtin_methods/src/ops/io.ops Tue Jul 1 20:05:53
>> 2008
>> @@ -187,19 +187,57 @@
>>
>> op print(in STR) :base_io {
>> STRING * const s = $1;
>> - if (s && string_length(interp, s)) {
>> + if (s && string_length(interp, s))
>> PIO_putps(interp, _PIO_STDOUT(interp), s);
>> - }
>> }
>
> I understand the checking for NULLness, but why check for length here and not
> everywhere else?
>
> -- c
>
This code was cargo culted from the corresponding print ops, with the
trailing "\n" tacked on for good measure.
--
Will "Coke" Coleda
