in reply to two regexes in one function call

It looks like second regex result is overwriting also the first result.

Indeed. 

$data = do { local $/ = undef; <DATA> };

print \$_, "\n" for
    scalar($data =~ s/\n/\n/g),
    scalar($data =~ s/(\w+)/\1/g);

__DATA__
Line1 Word Something
Line2 Other Word

[download]
SCALAR(0x239224)   \ Same var!
SCALAR(0x239224)   /

[download]

In ActivePerl 5.10.0 build 1004, s/// appears to place the return value in a global variable, and it appear to place that global variable (not a copy) on the stack.

By the way, it's wrong to use \1 in the replacement expression. Use $1.

Replies are listed 'Best First'.
Re^2: two regexes in one function call
by BrowserUk (Patriarch) on Jul 28, 2009 at 16:43 UTC

    I'm not sure that's the complete explaination as you get different results if you reverse the order of the statements: 

    #! perl -sw
use 5.010;

$data = do { local $/ = undef; <DATA> };

my $x = [
        scalar($data =~ s/(\w+)/$1/g),
        scalar($data =~ s/\n/\n/g),
];

say "@$x";

my $y = [
        scalar($data =~ s/\n/\n/g),
        scalar($data =~ s/(\w+)/$1/g),
];

say "@$y";

__DATA__
Line1 Word Something
Line2 Other Word

    [download]

    Gives:

    c:\test>783947.pl
6 2
6 6

    [download]

    (Plus tr isn't a regex!)

    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.
[reply]
[d/l]
[select]
      Minor change: 
      #! perl -sw
use 5.010;

$data = do { local $/ = undef; <DATA> };

my $x = [map \$_,
        scalar($data =~ s/(\w+)/$1/g),
        scalar($data =~ s/\n/\n/g),
];

say "@$x";

my $y = [map \$_,
        scalar($data =~ s/\n/\n/g),
        scalar($data =~ s/(\w+)/$1/g),
];

say "@$y";

__DATA__
Line1 Word Something
Line2 Other Word

      [download]
      SCALAR(0x2392c4) SCALAR(0x239224)
SCALAR(0x239314) SCALAR(0x239314)

      [download]

      The Ys use the same variable for the return value, but not the Xs.

      I don't actually know *why* the two calls return the same variable.

      Interestingly, the problem goes away if the captures are removed.

      Update: There has been a change in the relevant code since 5.10.0, but it appears to be an optimisation and not a fix.

      5.10.0:
      PUSHs(sv_2mortal(newSViv((I32)iters)));

      maint-5.10 as it stands right now:
      mPUSHi((I32)iters);

      Patch

      In both case, it looks like the operator always returns a new SV, so maybe there's some stack corruption??

[reply]
[d/l]
[select]

In Section Seekers of Perl Wisdom