Re: use feature 'unicode

Some Perl operators are currently buggy. If a string consists of "à",

Sometimes uc will uppercase it, sometimes it won't.
Sometimes /\w/ will match, sometimes it won't.
etc

But for "a" and for "ā",

uc will uppercase it.
/\w/ will match.
etc

What's special about "à"? It's in the iso-8859-1 character set and outside of the ASCII character set. That's obviously not a reason for the current misbehaviour.

This can't be fixed for backwards compatibility reasons, so a pragma was added.

Unicode::Semantics::up (aka utf8::upgrade) is a hack that addresses the same issue. However, it only affects one string, the effect is fleeting, and it forces the use of a less efficient storage format.

The pragma fixes all broken operators, without the side effects. And it's fixed mostly automatically; all you need is use 5.012;.

Comment on Re: use feature 'unicode_strings' Select or Download Code

Replies are listed 'Best First'.
Re^2: use feature 'unicode_strings' by Anonymous Monk on Mar 06, 2010 at 10:43 UTC
That was a pretty good explanation. What happens if I combine `use utf8;` and `use 5.012;`?	[reply]
Re^3: use feature 'unicode_strings' by ikegami (Patriarch) on Mar 06, 2010 at 21:09 UTC
You'll tell Perl your source in UTF-8 encoded, you want a version check for 5.12, and you want 5.12's backward-incompatible changes.	[reply]