Glob() gives wrong result when pattern not found

Bug #23769	Glob() gives wrong result when pattern not found
Submitted:	2003-05-23 04:11 UTC	Modified:	2003-06-03 06:04 UTC
From:	patric at liefdeis dot com	Assigned:	hholzgra (profile)
Status:	Closed	Package:	Filesystem function related
PHP Version:	4.3.2	OS:	FreeBSD
Private report:	No	CVE-ID:	None

[2003-05-23 04:11 UTC] patric at liefdeis dot com

When I do a glob() on my Linux machine with a pattern that doesn't have any hits, I receive an array with 0 entries. This is not the problem. When I do the same on my FreeBSD machine, I get 1 entry which is empty. This is a problem. The FreeBSD machine does run a CVS version of 4.3.0, where the Linux machine is running the 4.3.0 release version. I don't know if that has anything to do with it, or that the problem is OS based. Anywayz, tnx for looking in to it.

Patches

Pull Requests

History

AllCommentsChangesGit/SVN commitsRelated reports

[2003-05-28 06:10 UTC] patric at liefdeis dot com

I can't reproduce the bug in a simple form, nor can I upgrade the FreeBSD machine to a CVS version, so we will wait till the next official PHP version, then I will check again if the problem still exists. 

What I did notice was, that with PHP 4.3.1, Linux gives an empty array back, where FreeBSD gives nothing back. (print_r(glob("*.doesnotexist"));) Anyway, I will check back in after the next release.

[2003-05-31 05:02 UTC] patric at liefdeis dot com

Okay, I found the problem. Here is an example code:

<?
  echo "a";
  $f = glob("*.nothing");
  print_r($f);
  if (is_array($f)) {
    echo "c";
  }
  echo count($f)."b";
?>

When I look at this, I expect to get: 'a0b'. But I do get: 'a1b' on my FreeBSD 4.3 machine. On my Windows XP machine I get:

'aArray ( ) c0b'

That seems more okay. Also under Linux 2.4.18-3 and FreeBSD 4.8 it gives that same result. So I guess it is a typical FreeBSD 4.3 problem. Is it possible to fix this? Tnx anyway!

[2003-06-03 06:02 UTC] hholzgra@php.net

so what you actualy get is a boolean FALSE instead
of an empty array() ...

(you can see this if you change the print_r() call
in your test code to var_dump(), which is more verbose)

you can work around this by testing the result with
empty(), which returns TRUE for both empty arrays
and boolean FALSE values (amongst others) instad
of count(), which returns 1 if applied to any variable
that is not an array

any volunteer who can give me access to a FreeBSD 4.3 
system to identify which of the code pathes returning
FALSE it actualy runs into here so that we can 
hopefully fix this and get more consistent behaviour?

[2003-06-03 06:04 UTC] hholzgra@php.net

This bug has been fixed in CVS.

In case this was a PHP problem, snapshots of the sources are packaged
every three hours; this change will be in the next snapshot. You can
grab the snapshot at http://snaps.php.net/.
 
In case this was a documentation problem, the fix will show up soon at
http://www.php.net/manual/.

In case this was a PHP.net website problem, the change will show
up on the PHP.net site and on the mirror sites in short time.
 
Thank you for the report, and for helping us make PHP better.

found and fixed