Front page | perl.perl5.porters |
Postings from January 2010
* Dave Mitchell <davem@iabyn.com> [2010-01-04 00:25]: > Given that, should we add a warning to bytes.pm's pod to the > effect that you shouldn't be using this for new code unless you > really know what you're doing I’m in favour. I’d word it more strongly than that too. * Steffen Mueller <smueller@cpan.org> [2010-01-04 09:45]: > 1) If we add such a warning, we make certain assumptions* about > what the user was trying to do and ended up wanting to use > bytes.pm. Very nearly *everything* they would want to do with bytes.pm would be a mistake. Using bytes.pm in application code (as opposed to guts hackery) is almost by definition a bug. > 2) If we assume we have a good idea what she really wanted to > do, we should add pointers to the right way to do it next to > the warning. About the only useful thing I can think of is to point them to `utf8::downgrade`, which is sometimes us to cope with buggy XS code. > Without pointers to the right solution, such a warning wouldn't > be very helpful. The only correct thing you can do with bytes.pm is mess around with the byte buffer of an upgraded string while circumventing Perl’s own fixed-width/variable-width character conversion machinery. I would guess that of everyone in the world, maybe 5 guts-hacking CPAN authors will ever need to do that. And they probably know what they are doing anyway. Regards, -- Aristotle Pagaltzis // <http://plasmasturm.org/>Thread Previous | Thread Next