I've had a look around, but can't find whether this is a known issue or I'm doing something stupid. Basically I'm trying to extract a series of tokens from a string. Classic use of the m/\G.../gc construct. I have a loop as such:

my $a = 'first second ...other non-\w char groups...'; my @s;

while (... not finished string ...) {
  if ($a =~ /\G(\w+) ?/gc) {
    push @s, $1;
  } elsif ... other stuff not important
}
[download]

Now most of the string happens to be \w+ type tokens, so I thought it might be a useful performance optimisation to try and grab a whole heap of matches at once as such:

while (... not finished string ...) {
  if (my @m = ($a =~ /\G(\w+) ?/gc)) {
    push @s, @m;
  } elsif ... other stuff not important
}
[download]

Now the matches work correctly, but it doesn't update the internal pos(),\G marker, so the next match occurs at the same place again (eg infinite loop). I also tried something like:

while (... not finished string ...) {
  if (my @m = ($a =~ /\G(?:(\w+) ?)+/c)) {
    push @s, @m;
  } elsif ... other stuff not important
}
[download]

But it only returns the last match and doesn't appear to update pos(),\G either. So a couple of questions. Is there a way to grab multiple matches with m/\G.../gc? Am I dreaming that any of this might be a performance win? (what about long strings where $' has already been referenced so there's lots of string copying to do after each match?) Rob