I don't see the relation between your link and that bit of the regex. However, that said, your skepticism of its utility seems justified. I tried to find a use for it, but I haven't been able to make \d{4,}? act any differently than \d{4}. It's either a useless construct, or a failure of my imagination in coming up with an appropriate test case.

Putting aside what YAPE::Regex::Explain says about it, when I looked at it originally, I thought "Yack! Perl is gonna bitch about that weird '?' character". I could think of a couple other interpretations, so I put together a bit of code to check 'em out:

my @tests = (               
    'First case',           
    '123456789second & third case',
);                          
                            
for my $t (@tests) {        
    print "\nchecking '$t'\n";                                        
+          
    if ($t=~/(\d{4,}?)(.*)/)   { print "A: $1, $2\n"; }
    if ($t=~/(\d{4,}?)(.*?)$/) { print "B: $1, $2\n"; }
}
[download]

The other interpretations I could think of were:

• An optional set of 4 or more digits, kind of like (?:\d{4,})?. If true, the first case would give us:

  A: , First case'
  [download]

• Exactly 4 digits, like \d{4}, giving us:

  checking '123456789second & third case'
  A: 1234, 56789second & third case
  B: 1234, 56789second & third case
  [download]

• 4 or more digits, with as few as possible, yielding:

  checking '123456789second & third case'
  A: 1234, 56789second & third case
  B: 123456789, second & third case
  [download]

On reading the ...explain() output, I thought that I could perhaps make the third case come about. But what I actually got was:

$ perl xxxyyyzzz.pl

checking 'First case'

checking '123456789second & third case'
A: 1234, 56789second & third case
B: 1234, 56789second & third case
[download]

So I'm thinking that my initial surprise was justified, even though it's syntactically correct.

...roboticus

When your only tool is a hammer, all problems look like your thumb.

It's different if there is some following match condition:

use strict;
use warnings;

for my $reg ('(x\d{3,}?)', '(x\d{3,}?x)', '(x\d{3,})', '(x\d{3})') {
    for my $str ('xx', 'x12x', 'x123456x', 'x12x x123x') {
        print "Matched using $reg: $1\n" if $str =~ $reg;
    }
}
[download]

Prints:

Matched using (x\d{3,}?): x123
Matched using (x\d{3,}?): x123
Matched using (x\d{3,}?x): x123456x
Matched using (x\d{3,}?x): x123x
Matched using (x\d{3,}): x123456
Matched using (x\d{3,}): x123
Matched using (x\d{3}): x123
Matched using (x\d{3}): x123
[download]

True laziness is hard work

GrandFather:

Thanks. I updated your code to make it a little more visually obvious to me, then added a couple cases. Now that I see what the difference is, I doubt that I'd ever use it. Not because it isn't useful, but rather because if I ever need it, I'm sure I'll have long forgotten it. But it's certainly educational.

For grins, here's what I came up with:

$ cat splok.pl
use strict;
use warnings;

my @regs = ('(x\d{3,}?)', '(x\d{3,}?x)', '(x\d{3,})',
             '(x\d{3})',   '(x\d{3,}x)',  '(x\d{3,}?x?)');
my @strs = ('xx', 'x12x', 'x123456x', 'x12x x123x', 'x123456y');

printf "%-12.12s ", $_ for " ", @regs;
print "\n";
for my $str (@strs) {
    printf "%-12.12s ", $str;
    for my $reg (@regs) {
        printf "%-12.12s ", ($str=~$reg) ? $1: '-nope-';
    }
    print "\n";
}

$ perl splok.pl
             (x\d{3,}?)   (x\d{3,}?x)  (x\d{3,})    (x\d{3})     (x\d{
+3,}x)   (x\d{3,}?x?)
xx           -nope-       -nope-       -nope-       -nope-       -nope
+-       -nope-
x12x         -nope-       -nope-       -nope-       -nope-       -nope
+-       -nope-
x123456x     x123         x123456x     x123456      x123         x1234
+56x     x123
x12x x123x   x123         x123x        x123         x123         x123x
+        x123x
x123456y     x123         -nope-       x123456      x123         -nope
+-       x123
[download]

(As you can tell, I like things laid out in grids. I organize lots of stuff with database tables and spreadsheets...)

...roboticus

When your only tool is a hammer, all problems look like your thumb.