On Thursday, 28 January 2016 at 13:36:46 UTC, Puming wrote:
I searched the forum and found that people use `const(char)[]`
or `in char[]` to accept both string and char[] arguments.
There's also the hyper-generic signatures Phobos uses like
void foo(S)(S s) if(isSomeString!S)
and the input range types and the isConvertibleToString, etc., if
you want to take different encodings, or string-like objects, or
string wrappers/generators, etc.
But for just both string and char[], yeah, const is the way to do
it.
What's the difference between `const(char)[]` and `in char[]`?
`in char[]` means the whole parameter is const and you are not
supposed to escape it. It expands to `const scope char[]`, with
the const on the outermost level and `scope` too which means a
reference to it should never escape this scope.
const(char)[] global;
void foo(const(char)[] arg) {
arg[0] = 'a'; // illegal, the contents are const
arg = "whole new thing"; // legal, you are allowed to rebind
global = arg; // legal, you can escape it too
}
contrast with:
const(char)[] global;
void foo(in char[] arg) {
arg[0] = 'a'; // illegal, the contents are const
arg = "whole new thing"; // illegal, arg is const too
global = arg; // illegal, you can't escape the reference either
}
Note that the `global = arg` will still compile - the compiler
doesn't properly check this right now - but the spec says it is
illegal and may be optimized or statically checked differently at
any time, so if you do it you are in undefined behavior
territory. Don't do it!
If they are not the same, then which is better? If they are,
then why both forms exists?
I like `in char[]` if you intend to follow all the rules.
Otherwise, use the const version, which cannot be aliased because
it isn't a type , it is a parameter storage class and only valid
in that context.
