Why delimiter on char? c-based

Question

I was writing some code in C# the other day and it got me thinking. If a char literal can only have 1 char (or a delimited character) then why do we need the closing delimiter?

For example, currently we need to write:

char character = 's';

And there is no circumstance where this can work (obviously):

char character = 'sa';

Then why do we have to put the closing ' rather than have it implied? Eg:

char character = 's;
char anotherCharacter = '\';

Readability would only be impacted because we are used to the current standard (not necessarily because this way is less readable).

Answer 1

In C#, Java and C++ the answer is simple:

Because that's the way it was implemented in C.

Why was it implemented in C that way?

Who knows? There may be some implementation quirk in the first C parser/compiler that made this choice obvious; I'm betting that K&R just didn't think about it. A single quote just looks weird.

Answer 2

1. It's a convention
2. Characters are not necessarily represented in one character, in this unicode world
3. Control / escape sequences can be longer than one character (e.g. '\0x1a' )

Answer 3

'abcd' is a multi-character literal; it's type is int, and its value is implementation defined. The closing ' is needed to mark the end of the literal.

Answer 4

The existing C syntax defines character-constant as:

character-constant:

' c-char-sequence '

L' c-char-sequence '

Your question boils down to why this couldn't be defined instead as:

character-constant:

' c-char

L' c-char

Well, it could - in the sense that this would still be a consistent parseable grammar, and you'd still be able to express single-character constants. What you couldn't do is express multiple-character constants (eg. 'ab') - these are legal, but have a value that is implementation-defined.

I suspect that the true reason is simply aesthetics. For example, there's also no reason in theory why parantheses are needed around the conditional expression of an if statement either.