With this behaviour I cannot run a script with different flags without having to modify it if it contains a shebang lines with flags.

Do Python and Ruby do this as well?

With this behaviour I cannot run a script with different flags without having to modify it if it contains a shebang lines with flags.

You could also argue the other way round, i.e. it provides for consistency independently of how exactly the interpreter is being invoked — in particular as there are some operating system that don't handle shebangs (one rather commonly used one is among them...).

And some options might be an integral part of a script, or at least important to its proper functioning — like -n, -p, -l, -C (which didn't work on the shebang line until recently, btw), etc. — so I would rather want them to remain in effect when I call the script as perl myscript.pl .

#!/home/mh/perl512/bin/perl -w

warn "Hubba\n";
# script a.pl, produces warning
[download]

When you run this as "perl -X a.pl" you still get the warning - even though -X should disable them.

Is that because -w (from shebang) is "stronger" than -X (from command line)?

UPDATE:

The above is crap - sorry.

I think the reason Morgon says it's crap is because -w has no effect on warn. If you change the test to actually produce a warning, then the -X on the command line is indeed "stronger" than the -w in the shebang, but that's because the -X is always stronger than the -w, even if they swap places, or are both together on the command line or both together in the shebang.




- Boldra