RE: [PHP] handling large files w/readfile

Curt Zirzow wrote:

* Thus wrote Richard Lynch: Sebastian wrote:

i'm working on a app which output files with readfile() and some

headers..

i read a comment in the manual that says if your outputting a file php will use the same amount of memory as the size of the file. so, if the file is 100MB php will use 100MB of memory.. is this true?

I don't know if it's STILL true (or ever was) that readfile() would suck the whole file into RAM before spitting it out... Seems real unlikely, but... Never was and still isn't. using either readfile or fpassthru is the best route.

All I know that I am hosting a GForge site, and if I leave the download.php code as is, I will send up with apache processes that are 200+Meg. (the size of my download files). http://gforge.org/plugins/scmcvs/cvsweb.php/gforge/www/frs/download.php?rev=1.6;content-type=text%2Fplain;cvsroot=cvsroot%2Fgforge (which uses readfile) I have tried fpassthru - same thing. I have even tried: $fp = fopen($sys_upload_dir.$group_name.'/'.$filename,'rb'); while (!feof($fp)) { $buff = fread($fp, 4096); print $buff; } fclose ($fp); and I get the same thing. The only thing that seems to work is: Header("Location: ".$html_pointer_to_fp); which lets apache do the downloading. I would do a apache_child_terminate, but the function does not seem to be available to me (see my previous question about this). Any thoughts, or suggestions, I am open to try. My next experiment is: ============================ var $buff; while (!feof($fp)) { $buff = fread($fp, 4096); print $buff; } unset($buff); fclose ($fp); ============================ Hopefully that will make sure that the var $buff is only created once, and that the memory is cleared after the function is done. -Robin