- File::Copy's copy puts a lot effort in being portable.
- It's well tested.
- It doesn't launch a seperate process, so it's faster and uses less memory. Instead, it uses a system call (or Perl code if the OS doesn't have such a call).
- It can copy from/to file handles instead of just from/to files.
- Its arguments are not subject to shell interpretation. (The way in which you used system is buggy and unsafe.)
- It's makes checking for success much easier.
- It provides meaningful error messages in $!, on most systems.
On the other hand,
- Your expression is not portable.
- It's not well tested. In fact, there are bugs.
- It launchs *two* seperate processes, so it's slower and uses more memory.
- It can't copy from/to file handles, just from/to files.
- Its arguments are subject to shell interpretation (although that can be fixed). The way in which you used system is buggy and unsafe. (Bug 1)
- It thinks it succeeds even if the copy fails, as long as the shell was successfully loaded. (I think.) (Bug 2)
- $! does not contain what you think it does here. It definitely does not contain why the copy failed when it does. (Bug 3)
| [reply]
[d/l]
[select] |
For one thing, you don't have to spawn (I think that's the right term) a second process if you use Perl's copy instead of the system command. Also, I find it easier to verify that your copy in fact worked correctly (though this can be done with a system call as well I'm sure). | [reply] |