I'd guess that's an incorrect representation of what Grubba reported,
some time ago:
> I've now changed the implementation of the subtype for strings in
> Pike 7.7 to an integer range rather than a width in bits:
>
> > typeof("");
> (1) Result: string(zero)
> > typeof("foo");
> (2) Result: string(
great, thanks.
what about string(0)?
there is one reference like that in src/operators.c line 4304:
if(sp[-1].u.string->size_shift) {
bad_arg_error("`~", sp-1, 1, 1, "string(0)", sp-1,
"Expected 8-bit string.\n");
}
this seems rather odd...
greetings, martin.
string(0..255)
The specialization of the string type is going to allow a code point
range, much like how the int type with an integer range constrains
ints carried in the variable to the given range (type checker wise,
anyway).
while testing the documentation build i discovered that StringType was
defined as string(width), however the actual implementation appears to
require string(minwidth..maxwidth):
> string(8) foo;
Compiler Error: 1:parse error, unexpected `')'', expecting `TOK_DOT_DOT'
or `TOK_DOT_DOT_DOT'
Comp
