Re: Some odd ambiguity in this regex (updated)

[AnomalousMonk]

Further to hippo's post:

... something in the steps you are not showing us.

misterperl:   It's easy to contrive that scenario entirely separate from debug mode:

c:\@Work\Perl\monks>perl -wMstrict -le
"$_ = '1223w3433.45+34';
 ;;
 /\d\D*/g;
 printf qq{'$&' } while /\d\D*/g;
"
'2' '2' '3w' '3' '4' '3' '3.' '4' '5+' '3' '4'

m//g in scalar or void context "remembers" the point at which a previous match ended and continues matching from that point. See pos.

Update: See what happens if the side-effect-producing, void context
    /\d\D*/g;
statement is changed to
    () = /\d\D*/g;
(i.e., if list context is imposed).

[ikegami]

It's often the result of using the nonsensical if (/.../g)

$ perl -e'
   $_ = "abc";
   if (/\w/g) {
      CORE::say for /(\w)/g;
   }
'
b
c

[misterperl]

I dont think I "contrived" anything. I did "R" and the first line was :

$_= '1223w3433.45+34';

then I did the "x". If the debugger works as you suggest, then it's not working right. It should interpret the expression correctly; if something isn't reset (which doesnt make sense ,since this happens on line 2) than it needs to get reset..

Larry?

But, you're correct it IS a debugger thing because the interpretation is correct outside dbg...

[LanX]

Please apply this single line and show us the result:

  DB<5> x $_='12a3b'; /\d\D*/g
0  1
1  '2a'
2  '3b'
  DB<6>

> But, you're correct it IS a debugger thing

Very unlikely.

And you've already been quick to blame Perl in the past.

Edit

We need a possibility to reproduce your problem before searching a hypothetical bug.

So please reproduce, and tell us the versions of perl and the debugger.