It would return B. Based on OP's criteria as stated, any hyphenated "word" would normally be returned as multiple words, but 52 does not contain a letter so would not be returned.

1 members found this post helpful.

Probably. The \b escape sequence may mean different things in different languages/toolsets.

Well, yes, but the language we're talking about is regular expressions. And in regular expresions, \b means "word boundary".

\b is from PCRE (perl). Not yet standard in ERE (grep -E, sed -E, awk, ...)
It exists in a recent Linux glibc, but not all tools use it.
The older \< \> is always supported in Linux.

Compare with \s - the older [[:blank:]] works with all tools in Linux.

1 members found this post helpful.

No, \b is not "from" PCRE and it is not "PCRE (perl)" either.

PCRE is a C library started in 1997 that was originally inspired by Perl v5, but PCRE is not Perl, and nor is it where \b for word boundary came from.

I can't say for sure where \b originated from - other than definitely not from PCRE - Perl had \b in (at least) v4 in the early 90s (no idea about earlier versions), and the original v1 of Gawk released in the late 80s originally had \b for word boundary - with v3 it was switched to \y to allow compatibility with original AWK using \b for backspace.


Gawk had \< and \> in v2, but not in v1 - thus they were probably added early 90s - and it seems Howard Helman added them to Sed in 1991 - so it's possible \< and \> are newer constructs - despite many "modern" regex implementations for some reason not implementing them.

1 members found this post helpful.

AFAIR perl4 had \< \> only, and \b was new in perl5.

https://perldoc.perl.org/perlre
perl does not know \< and \>. But it has \A and \Z (which are not the same, just probably similar).

1 members found this post helpful.

That is perl5.
I like their subtle humor:
I think perl4, being from the pre-Web era, is completely extinguished.

I hope not - it is the one burned into my brain cell!

If that were true, it would be mentioned under https://perldoc.perl.org/5.6.2/perltrap#Perl4-to-Perl5-Traps]https://perldoc.perl.org/5.6.2/perltrap#Perl4-to-Perl5-Traps

(It isn't.)


They are not the same at all - they are equivalent to "^" and "$" when matching single lines, but for multiple lines "\A" and "\Z" only match once (for the first/last respectively). In the latter case, usually one wants lowercase "\z", which doesn't exclude a final newline.


Except, despite being a rewrite, Perl 5 mostly maintained compatibility (aside from the issues at the link above).

Anyway, who still uses Perl 5? All the cool kids have moved to Perl 7...

1 members found this post helpful.

Not really. :)