Regex bug or head made of solid bone?

alien_life_form has asked for the wisdom of the Perl Monks concerning the following question:

Greetings,

Consider the following snippet:


$orig='aaa([\+\-]{1})bbb';
$qre="\\Q([\\+\\-]{\\E\\d\\Q})\\E";
$bre='\(\[\\\\\+\\\-\]\{\d\}\)';

$str=$orig;
$str =~ s!\Q([\+\-]{\E\d\Q})\E!!;
print "Replace against literal quoted: $str\n";

$str=$orig;
$str =~ s!\(\[\\\+\\\-\]\{\d\}\)!!;
print "Replace against literal bslashed: $str\n";

$str=$orig;
$str =~ s!$qre!!;
print "Replace against variable quoted => $qre <=: $str\n";

$str=$orig;
$str =~ s!$bre!!;
print "Replace against variable bslashed => $bre <=: $str\n";
[download]

Running this (with either 5.6.1 or 5.8.0RC1:

c:\temp> D:\perl58\bin\perl.exe re.pl
 D:\perl58\bin\perl.exe re.pl
Replace against literal quoted: aaabbb
Replace against literal bslashed: aaabbb
Replace against variable quoted => \Q([\+\-]{\E\d\Q})\E <=: aaa([\+\-]
+{1})bbb
Replace against variable bslashed => \(\[\\\+\\-\]\{\d\}\) <=: aaabbb

c:\temp>perl re.pl
perl re.pl
Replace against literal quoted: aaabbb
Replace against literal bslashed: aaabbb
Replace against variable quoted => \Q([\+\-]{\E\d\Q})\E <=: aaa([\+\-]
+{1})bbb
Replace against variable bslashed => \(\[\\\+\\-\]\{\d\}\) <=: aaabbb

C:\temp\perl -v
This is perl, v5.6.1 built for MSWin32-x86-multi-thread
(with 1 registered patch, see perl -V for more detail)
[download]

Uh? Am I missing something?
Cheers,
alf

You can't have everything: where would you put it?

Comment on Regex bug or head made of solid bone? Select or Download Code

Replies are listed 'Best First'.
Re: Regex bug or head made of solid bone? by Abigail-II (Bishop) on Jun 06, 2002 at 18:32 UTC
I don't know what exactly your question is. You show us some code, its output, and then just wonder if you are missing something. I certainly am missing something, namely your expectation of what the code should do. But I take an educated guess. You are baffled by the behaviour of \Q and \E. One should realize that \Q and \E are dealt with in the interpolation stage, which happens before a regex is compiled. I refer to the section "Gory details of parsing quoted constructs" in the perlop manual. Abigail	[reply]
Re: Re: Regex bug or head made of solid bone? by alien_life_form (Pilgrim) on Jun 07, 2002 at 09:06 UTC
Greetings, After more grinding, my question actually became two questions. The first is actually about single quoting. The second one is addressed by your guess. (And I surely thought that my original post contained sections that qualify as 'some code' and 'output' but this may not be relevant at this point.) Cheers, alf You can't have everything: where would you put it?	[reply]
Re: Regex bug or head made of solid bone? by vladb (Vicar) on Jun 06, 2002 at 13:30 UTC
I've tried to play with this code: `$orig='aaa([\+\-]{1})bbb'; $res = '([\\+\\-]{'; $qres = '\\Q$res\\E'; $str=$orig; $str =~ s!\Q$res\E!!; print "Replace against variable quoted => $res <=: $str\n"; $str=$orig; $str =~ s!$qres!!; print "Replace against variable quoted => $qres <=: $str\n";` [download] Examining its output: `Replace against variable quoted => ([\+\-]{ <=: aaa1})bbb Replace against variable quoted => \Q$res\E <=: aaa([\+\-]{1})bbb` [download] I notice that placing special metacharacters inside your variable doesn't make them 'effective'. However, when they are outside of the variable, they get 'noticed' by the regular expression parser. So, could it be the answer to your inquiry? ;-) _____________________ $"=q;grep;;$,=q"grep";for(`find . -name ".saves~"`){s;$/;;;/(.-(\d+) +-.*)$/; $_=["ps -e -o pid \| "," $2 \| "," -v "," "];`@$_`?{print"+ $1"}:{print" +- $1"}&&`rm $1`; print$\;} [download]	[reply] [d/l] [select]
Re: Re: Regex bug or head made of solid bone? by alien_life_form (Pilgrim) on Jun 06, 2002 at 17:06 UTC
Greetings, interesting... I wonder whether this is by design or by accident? Looking back at my original example, another thing is appears to be non-obvious. The (backslashed) portion that reads: `\\\\\+\\\- # matches literal \+\- in a variable` [download] had originally been written as: `\\\+\\\- # matches literal \+\- in a literal regex` [download] why do I need an additional '\' before a quoted + and only within a variable is beyond me.... Same reason? Cheers, alf You can't have everything: where would you put it?	[reply] [d/l] [select]