I made a visit counter for multiple pages that stores the monthly visits in log.txt files that are formatted as follows: ID;COUNT
So whenever someone visits one of the pages the page redirects to a count.php
where the ID of the page that comes via GET in the URL will be looked up, if it exists in this months log.txt, the program proceeds to add 1 to the value after the semicolon.
So, say, someone visits page.html
, he gets redirected to count.php?id=0
, which contains among other stuff, the following lines.
$id = $_GET['id'];
$file = file('./logs/' . $date);
$arr = explode(";",$file[($id)]);
$arr[1] += 1;
$arr = implode(";",$arr) . "
";
$file[($id)] = $arr;
$fl = implode("",$file);
file_put_contents('./logs/' . $date,$fl);
after that, it looks for the ID in another file, where there's the actual target URL for that id, once it has that, the user gets redirected to that.
Most of the time it works as intended, but sometimes it messes up and my logs end up looking like that:
;19
;13
;25
;18
;10
;12
;16
;22
;10
;8
100;16
101;28
102;14
when it should be looking like that
0;13
1;27
2;48
3;990
4;331
.
.
.
101;323
102;392
103;87
104;63
One idea that I had was that, because it's quite possible for two people to visit those sites simultaneously, and I didn't implement any file locking, it may get confused with that. But wouldn't that only write old data and not mess with the order or the formatting