On Tue, 2005-03-15 at 10:09, Autrijus Tang wrote:
> Currently Pugs numifies hexadecimal and octal strings as if they
> are literals; that means "0x123" and "0o456" all work as expected.
> Is that an acceptable treatment? What about "Inf" and "NaN" in
> numeric context?
This has long been a point of contention in Perl 5.
There are two camps when it comes to how to interpret strings as
numbers:
1. You must treat them the same for consistency.
2. You must not treat them the same because someone reading in
lines of a report like the following could be seriously shocked
by the behavior:
$ Book
5foundation
8rama
0xanth
The logic goes: why should "8rama" and "0xanth" be treated
differently when they are both numbers followed by letters?
There's no particularly right answer, IMHO, so why doesn't Perl 6 just
go all ways?
Have a Perl 5 atof mode, a hard-line mode that throws an exception if
handed anything other than a number that would be valid in Perl source
and a mode that silently makes assumptions, deletes trailing junk and
tries to act "perl6ish".
# Untried, likely wrong pseudo-code:
use atof::perl5;
assert { "0xanth" == 0 };
use atof::strict;
try { "0xanth" == 0; CATCH { assert { $@ && $@ ~ /atof/ } } }
assert { "0xa" == 0xa }
use atof::perl6; # Default?
assert { "0xanth" == 0xa }
Pardon me if I'm forgetting a document that already decided this. Of
course, partly this is better directed at p6l... I wasn't sure if I
should move it to that list or not.
--
Aaron Sherman <[EMAIL PROTECTED]>
Senior Systems Engineer and Toolsmith
"It's the sound of a satellite saying, 'get me down!'" -Shriekback
