What do you mean when you say that it "doesn't seem to work"? Are you checking the return value from open? Have you tested the script from the command line? Are you sure the current working directory is what you think it is when the script is run from the web?

Anyway, it sounds like you need to use file locking. Opening for read/write won't prevent two instances of the script from doing conflicting writes. See flock and Fcntl.

I AM locking the files. But, when I open the file, it doesnt return an error value, but it doesnt open a file stream either. So I don't know what the heck is wrong...

As an aside, I'm assuming that you're using the "different instances" you refer to because you are having problems with open. If not (i.e. if you're going to be using multiple instances regardless of whether open behaves or not), have you looked at flock? (If you use flock, keep in mind that its locks are only advisory, so a completely different process can just clobber it.)

hdp.

I'm on perl, version 5.005_03 built for i386-freebsd...the code is as follows:

open(FH,"+<./file.dat") || die("Can't open ./file.dat for reading/writ
+ing: $!");
flock(FH,LOCK_EX);
seek(FH,0,0);
@dat = <FH>;
foreach (@dat) {chomp}
# this is where I process @dat, and when I feed text 
# manually into it, it works.  the problem doesn't
# lie in this segment of the code
seek(FH,0,0);
print FH join("\n",@dat);
flock(FH,LOCK_UN);
close(FH);
[download]

Now, i can open the file for reading, and i can open it for writing, and both work fine. But when I try reading AND writing, it doesnt spring an error, it simply doesnt read anything from or write anything to the file. Should I try +>instead of +<or what?

I've taken your code and run it unchanged on my machine except for the addition of this line (push @dat, $dat[rand(@dat)]) just before the second seek. It does what I'd expect it to, reading in the file and writing it out again with a random line from the file appended.

I suggest you show the rest of your code, even though you don't think the problem lies there.

hdp.

I'd concur with the other poster who encouraged you to look at the flock node and its sample code for doing file locking in Perl. That's a pretty standard approach for ensuring that multiple running scripts can't write to the same file at the same time. (At least if those scripts are all using flock, you won't have a problem. Read the fine print about advisory locking for more details.)