Airbnd suggests I do this:
!function() {
// ...
}();
Because:
This ensures that if a malformed module forgets to include a final semicolon there aren't errors in production when the scripts get concatenated.
The bang allows me to work around the language's grammar rules:
// Evaluated in Chromium 34 console.
function(){}(); // => SyntaxError: Unexpected token (
!function(){}(); // => true
And when concatenating other modules the bang seems to do the trick:
!function(){}();function(){}(); // => SyntaxError: Unexpected token (
!function(){}();!function(){}(); // => true
(function(){}());!function(){}(); // => true
However it doesn't seem to actually be "safe," because if someone else doesn't have a semi-colon at the end of his script:
!function(){}()!function(){}(); // => SyntaxError: Unexpected token !
(function(){}())!function(){}(); // => SyntaxError: Unexpected token !
It would appear that a leading semi-colon IIFE is better.
;(function() {
// ...
}());
!function(){}();(function(){}()); // => undefined
(function(){}());(function(){}()); // => undefined
!function(){}();;(function(){}()); // => undefined
(function(){}());;(function(){}()); // => undefined
Am I missing something? Is it actually acceptable to use bang "!" functions or are leading semi-colon ";" IIFEs truly superior because of the way they concatenate?
You always have IEFEs there. Whether you wrap them in parenthesis or prefix them with a !
is your choice, and doesn't make a difference. You need either one to force the function to be parsed as an expression. See javascript function leading bang ! syntax for details.
Whether you prefix that whole construct with a ;
to prevent errors from concatenation with poor-written scripts (What does the leading semicolon in JavaScript libraries do?) is totally unrelated. You can mix the patterns as you want:
!function(){…}() // not safe for arbitrary concatenation
(function(){…}()) // not safe for arbitrary concatenation either
;!function(){…}()
;(function(){…}())
However, there is one case of concatenation where ()
vs !
does make a difference: if two scripts are concatenated so that there is a newline in between, and the former does not end in a semicolon. This does allow automatic semicolon insertion to jump in - when the next line does begin with a bang!
1 + 2 // script A
!function(){…}(); // script B
// works!
1 + 2 // script A
(function(){…}()); // script B
// runtime error: "2 is not a function" (or whatever the previous line ends in)
We learn: Always end your script with a semicolon. Use intelligent concatenation. Start your script with a semicolon if you need to be safe against dumb concatenation.
The !
has actually nothing to with protecting against missing semicolons when concatenating files. It's often used to enforce that the following function definition is evaluated as expression, not as declaration.
A leading semicolon on the other hand ends any "open" expression statement that might come before it.
But if you follow the explanation, it seems like !
would suffice if it's the first character in the line. Apparently JS would add a semicolon to the previous due to ASI. However, that seems to be a more fragile approach, since you might not be in total control over how the modules are concatenated. Using a semicolon is definitely "safer".
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With