Re: Memory leak in unicode substitution

"x" =~ / [\x{1234}] /x              for 0 .. 100_000;
"x" =~ /(?: \x{1234} | \x{1234} )/x for 0 .. 100_000;

"\x{4321}" =~ / \x{1234} /x         for 0 .. 100_000;
[download]

Curious, the bug doesn't bite if character is put in a class or dummy alternation. Most important, there's no bug if target string is utf8 itself. That's why, I think, it wasn't found sooner. Unicode in regexes most often means Unicode in texts.

Comment on Re: Memory leak in unicode substitution Download Code

Replies are listed 'Best First'.
Re^2: Memory leak in unicode substitution by holli (Abbot) on Aug 30, 2019 at 17:06 UTC
Good catch, can confirm. Surprisingly, simply adding `use utf8;` doesn't fix it though. Aren't all string literals to be Unicode when `utf8` is in effect? holli You can lead your users to water, but alas, you cannot drown them.	[reply] [d/l] [select]
Re^3: Memory leak in unicode substitution by choroba (Cardinal) on Aug 30, 2019 at 17:12 UTC
No. `use utf8;` just means UTF-8 is used in the source code (both in string literals or identifiers). But even `use feature qw( unicode_strings );` doesn't help here. `map{substr$_->[0],$_->[1]\|\|0,1}[\\|\|{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^ARGV,3]`	[reply] [d/l] [select]