From: Roman Neuhauser [mailto:neuhauser@sigpipe.cz] # telya@mail.ru / 2006-09-08 12:23:36 +0300:

...

Hi

http://www.boost.org/libs/regex/doc/syntax_perl.html http://www.boost.org/libs/regex/doc/syntax_perl.html says:

Perl Regular Expression Syntax

In Perl regular expressions, all characters match themselves except for the following special characters:

.[{()\*+?|^$

...

Q: what about "]}" symbols?

They're only special if there's a matching "[" or "{" earlier in the pattern:

> echo 'aaa]}]}]}bbb' | perl -pe 's/(]|})+//' aaabbb > echo 'aaa]}]}]}bbb' | perl -pe 's/[]}]+//' aaabbb > echo 'aaa}}}]]]bbb' | perl -pe 's/}+//;s/]+//' aaabbb

Moving this discussion from boost-users to boost-development... Thanks for explanation, Roman. I've checked this on boost v.1.33.1, but only unmatched "]" is allowed, construction of boost::regex("}") claims unmatched "}". Is it by design or already fixed in v.1.34? I had to handle special characters, so I just "copy/paste"ed from docs. And so, I didn't handle "}]". Do you think it's a good idea to expand description of special characters in docs to include explanation about "}]" characters? As I checked, both unmatched "]" (and probably "}" in 1.34) and unmatched but escaped "\}\]" are allowed and have the same effect. It was very useful for me, so maybe it's worth to mention in docs too. Thanks, Andriy.