$ perl -MO=Deparse -e 'for ($x = 0; $x < 10; $x++) {}'
$x = 0;
while ($x < 10) {
    ();
}
continue {
    ++$x
}
[download]

which version of perl? I can't seem to reproduce it... Mine is 5.6.1 on a solaris

  me@myhost> perl -MO=Deparse -e 'for( my $x = 0; $x < 10; $x++ ) { pr
+int "foo\n" }'
  for (my $x = 0; $x < 10; ++$x) {
      print "foo\n";
  }
  -e syntax OK
  me@myhost>
[download]

5.6.0 on OS X. Same result on 5.005_03. Presumably either Deparse or the peephole optimizer were thumped in 5.6.1 giving the different result.

B::Deparse is your friend here.

Ah. Well. Coming from you I'll take your word for it. But frankly I tend not to trust B::Deparse too much. :-)

D:\Development>perl -MO=Deparse  -e "for ($x = 0; $x < 10; $x++) {}"
for ($x = 0; $x < 10; ++$x) {
    ();
}
-e syntax OK

D:\Development>perl -MO=Deparse,-x3  -e "for ($x = 0; $x < 10; $x++) {
+}"
$x = 0;
while ($x < 10) {
    ();
}
continue {
    ++$x
}
-e syntax OK
[download]

This is on 5.6.1 AS631. Btw from the documentation for the -x switch

---

Expand conventional syntax constructions into equivalent ones that expose their internal operation. LEVEL should be a digit, with higher values meaning more expansion. As with -q, this actually involves turning off special cases in B::Deparse's normal operations.

---

Which suggests to me that im right. (But which is also why I asked a Guru like yourself. I was hoping for a definitive answer that did not depend on Deparse.)

Nevertheless, thanks.

Yves / DeMerphq
---
Writing a good benchmark isnt as easy as it might look.

Older versions of deparse are accurate, as far as it goes. The problems in the node you referenced aren't Deparse's fault--it's not currently possible to turn an optree back into the source it came from. There's a not insignificant amount of stuff that's thrown away during the compile or peephole optimization phase, which makes things a bit tough.

Hey, at least I didn't recommend you rebuild perl with -DDEBUGGING and dig through the output of perl -Dt... :)