The document is correct on this issue, those character types may be used.

I've tested your code as:

if (ereg('<T>([\da-z]{1,})', '<T>try8')) { echo 'match'; }

Which, works just fine on win32 platforms, 4.1.1.  You may want to ensure there is no whitespace in the string you are testing.

Status -> Bogus

'Bug' withdrawn.

I mixed up perl & posix? :(  SORRY.

Regarding this reply I just want to make clear what was not working at my comp.

here is code which I used :

<?php
echo "*** POSIX<BR>\n";
if (eregi("<T>([\da-z\ ]{1,})", "<T>qw e1 asd!zxc", $out)) {
	echo "found=".count($out)."<BR>\n";
	foreach($out as $val) {
		echo "=$val<BR>\n";
	}
}
echo "<BR>*** PERL<BR>\n";
if (preg_match("/<T>([\d\ a-z]{1,})!([\w]+)/", "<T>qw e1 asd!zxc", $out)) {
	echo "found=".count($out)."<BR>\n";
	foreach($out as $val) {
		echo "=$val<BR>\n";
	}
}
?>

and output in my browses was:

*** POSIX
found=10
=qw e
=qw e
=
=
=
=
=
=
=
=

*** PERL
found=3
=qw e1 asd!zxc
=qw e1 asd
=zxc

See, there is difference even if match is found, but \d is not the same as in perl.
And I don't know what \d does in POSIX(probeable nothing)?
(Still checking documentation). :)

PS Just trying to help. :)