Forgot to mention that it works perfectly with MySQL driver.

This bug has been fixed in CVS.

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/.
 
Thank you for the report, and for helping us make PHP better.

I couldn't get it to compile, so I corrected the mysql -> 
dblib, not sure whether this was the intention though. I 
don't have enough karma, so the patch is below:

RCS file: /repository/php-src/ext/pdo_dblib/dblib_stmt.c,v
retrieving revision 1.6.2.2.2.3
diff -u -r1.6.2.2.2.3 dblib_stmt.c
--- dblib_stmt.c        14 Jan 2007 16:57:50 -0000      
1.6.2.2.2.3
+++ dblib_stmt.c        14 Jan 2007 21:43:35 -0000
@@ -250,9 +250,9 @@
        return 1;
 }
 
-static int dblib_mysql_stmt_cursor_closer(pdo_stmt_t *stmt 
TSRMLS_DC)
+static int pdo_dblib_stmt_cursor_closer(pdo_stmt_t *stmt 
TSRMLS_DC)
 {
-       pdo_dblib_stmt *S = (pdo_mysql_stmt*)stmt-
>driver_data;
+       pdo_dblib_stmt *S = (pdo_dblib_stmt*)stmt-
>driver_data;
 
        if (S->rows) {
                free_rows(S TSRMLS_CC);
@@ -273,6 +273,6 @@
        NULL, /* get attr */
        NULL, /* meta */
        NULL, /* nextrow */
-       dblib_mysql_stmt_cursor_closer
+       pdo_dblib_stmt_cursor_closer
 };


However, it doesn't fix the problem. See, gdb breaks fine on 
pdo_dblib_stmt_dtor on each iteration and stepping it shows 
that it even attempts to free stuff. So I guess the error 
isn't here?

After running gdb with patch above it does close the cursor 
on each iteration, but it's not still affecting the 
increasing memory usage.

After playing around with GDB I noticed that the S->cols is 
being allocated on different memory block each time. S and S-
>rows stay in the same place, though. I suppose this is due 
the different methods used to allocate memory, but I just 
don't have enough clue to fix it.

And more issues :-)

That previous example was done with just one row in the 
resultset, but after I tried to return more rows, I found 
something disturbing:

Code:

$db = new PDO('dblib:host=hostname;dbname=database', 
'username', 'password');
$query = 'select top 10 * from table';
$stmt = $db->prepare($query);

for ($i = 0; $i < 10; $i++) {
    $stmt->execute();
    $stmt->fetchAll();
    $stmt->closeCursor();

    echo memory_get_usage() . "\n";
}

Output:

172240
240044
307824
375632
443464
511288
579064
646932
714776
782576

This is still a problem for me in 5.2.5 using the PDO DBLIB driver. Mysql and mssql_* functions work fine on the same query.

Please try using this CVS snapshot:

  http://snaps.php.net/php5.2-latest.tar.gz
 
For Windows:

  http://windows.php.net/snapshots/

No feedback was provided for this bug for over a week, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".

I have the same problem with version 5.3.10