Front page | perl.perl5.porters |
Postings from December 2015
On 11/18/2015 08:37 PM, Ricardo Signes wrote: > * Karl Williamson <public@khwilliamson.com> [2015-11-16T21:57:35] >> I still don't think [operating bitwise on codepoints in character strings] is >> useful in general. Code points are not generally assigned where the >> relationship between them is meaningful in bit operations ways. > > Agreed. Although there are a few use cases that were pointed out, I think we > need to decide what the behavior of &. on strings is and be consistent. I > still think the correct thing to do is assert that the arguments to &. are > treated as, and must plausibly be, byte strings. > > It shouldn't matter whether the string is upgraded or not, as it's too easy to > end up with bytes in an upgraded string or codepoints in a downgraded string. > No Unicode Bug! > Ok. I'd like ideas on how to word the deprecation message that gets raised. I don't like what I've come up with: "It is deprecated to '%s' a string containing non-byte data", where %s is PL_op_name[PL_op->op_type]); I don't like it because it uses the passive voice, but more importantly, the name becomes something like 'bit_xor', which I don't think will necessarily make sense to the reader. So, any ideas?Thread Previous | Thread Next