It depends on how apache is configured. You may have the general log just log IP, time, and file. You don't want it to have Referer, User-Agent, etc except for the files gotten from FTP. I don't care if you use Linux with Mozilla coming from perlmonks to display a .gif, but I do care if all the people downloading a file have certain things in common. If I find out 50% of people downloading a program are using NT, but my program has a small bug with NT, then I'd probably fix it. Also, what if I have a specific version for NT, for Linux, for 95/98? I don't have to trust the user to download the right one, my script does it for me. | [reply] |
That would be a case to write in a separate logging function then. But for remote IP, filename, result code, and I believe size, the default Apache log has all of these.
Also, why would you write a HTTP CGI for FTP? Why not just send the file out via HTTP, this alleviates the overhead needed for FTP.
| [reply] |
Having a CGI to go to FTP could be used for load balancing or choosing the FTP server closest to you.
| [reply] |