Transfer a created zip file to the client without store it on the server?

[IanYang]

I use pclzip.lib.php to compress files on the server into a zip file. For example, the following code will compress "text.txt" in the data folder and whole "images" folder into a zip file named "file.zip".

<?php
 
    include_once('pclzip.lib.php');
 
    $file = new PclZip('file.zip');
 
    $path = $file -> create('data/text.txt,images');
 
    if(!$path) die('Error : ' . $file -> errorInfo(true));
 
?>

But the file.zip will also be stored on the server. How to transfer it to the client instantly without store it on the server?

[logic_earth]

You are going have to store it some where. Having it dynamically created at some defined bitrate while the client's connection is in flux, is more trouble then its worth.

[IanYang]

I planned to transfer the created file using javascript at first:

<?php
 
    include_once('pclzip.lib.php');
 
    $file = new PclZip('file.zip');
 
    $path = $file -> create('data/text.txt,images');
 
    if(!$path) die('Error : ' . $file -> errorInfo(true));
 
    echo '<script type="text/javascript">window.location = "http://www.example.com/' . $file . '"</script>';
 
?>

But it may cause some problem. If one client execute this code, a file.zip is created and he is now downloading the file. But if another client execute this code before the previous client finishing his download, then the program will show some error indicate the zip file cannot be overwritten (because a zip file with the same file name is now downloaded by the previous client).

[sk89q]

Save it to a temporary filename. You don't have to redirect the user to the download by the way. You can just send the appropriate headers and then do readfile().

You may also want to not re-generate the ZIP file if someone recently requested the exact same one.

[IanYang]

Save it to a temporary filename. You don't have to redirect the user to the download by the way. You can just send the appropriate headers and then do readfile().

You may also want to not re-generate the ZIP file if someone recently requested the exact same one.

Thank you! The readfile() function works fine.

How can I make the server delete the zip file automatically right after all clients finished their downloads?

[crmalibu]

You could just unlink() the file after readfile() completes. If you have multiple users downloading the same file, this could cause problems. But if you just make a distinct filename per users, it should work pretty good.

You could also periodically delete zip files in that dir that have a fileatime() older than say, 15 minutes or whatever.

[IanYang]

You could just unlink() the file after readfile() completes. If you have multiple users downloading the same file, this could cause problems. But if you just make a distinct filename per users, it should work pretty good.

You could also periodically delete zip files in that dir that have a fileatime() older than say, 15 minutes or whatever.

Thank you. It seems make a distinct file name for each user is a better solution. I will handle that later.

I added unlink($file); after readfile($file); and the zip file was deleted successfully.

<?php
 
    include_once('pclzip.lib.php');
    $file = new PclZip('file.zip');
    $path = $file -> create('data/text.txt,images');
    if(!$path) die('Error : ' . $file -> errorInfo(true));
 
    header('Content-Description: File Transfer');
    header('Content-Type: application/octet-stream');
    header('Content-Disposition: attachment; filename=' . $file);
    header('Content-Transfer-Encoding: binary');
    header('Expires: 0');
    header('Cache-Control: must-revalidate, post-check=0, pre-check=0');
    header('Pragma: public');
    header('Content-Length: ' . filesize($file));
    ob_clean();
    flush();
    readfile($file);
    unlink($file);
    exit;
 
?>

But how did the server manage to do this? It deleted the file before I start to download the file. How come I'm still able to download the file?

[leelong]

But how did the server manage to do this? It deleted the file before I start to download the file. How come I'm still able to download the file?

well i am not sure, may be readfile($file) will buffer to memory and local copy of hdd been deleted?

pardon me if i am wrong

[crmalibu]

readfile() will return as soon as it writes the data to the output stream. php may buffer its own output stream, and the webserver may buffer as well. You can see flush() and ob_flush() if you want to try to minimize buffering before you unlink.

[IanYang]

I see. Thank you.

[sk89q]

By the way, you don't really need all those headers.

Just these three are sufficient (unless you do not want caching under any circumstances and the URLs are not unique):

header('Content-Type: application/octet-stream'); // Should be application/zip, really
header('Content-Disposition: attachment; filename=' . $file);
header('Content-Length: ' . filesize($file)); 

And you only need the Content-Disposition header if your URL ends in some ugly filename (like download.php?want=images,text).

[IanYang]

By the way, you don't really need all those headers.

Just these three are sufficient (unless you do not want caching under any circumstances and the URLs are not unique):

header('Content-Type: application/octet-stream'); // Should be application/zip, really
header('Content-Disposition: attachment; filename=' . $file);
header('Content-Length: ' . filesize($file)); 

And you only need the Content-Disposition header if your URL ends in some ugly filename (like download.php?want=images,text).

Thank you. That makes the code cleaner.