On 1/30/15 5:06 AM, "Marc =?UTF-8?B?U2Now7x0eiI=?= <schue...@gmx.net>"
wrote:
On Friday, 30 January 2015 at 08:52:41 UTC, bearophile wrote:
Jeremy DeHaan:
I figured that it would be smart enough to
deduce the parameter type based on the type that it was trying to be
assigned to.
For that you need languages like Haskell/Rust. D type inference
doesn't work from the type something is assigned to.
But it could, right (for arguments and return expressions, too)? At
least I don't see any real obstacles. Just file an enhancement request
if you think it's worthwhile, or start a discussion on digitalmars.D.
Of course, it cannot work for `auto x = getString();` or
`foo(getString())` where there are several overloads of `foo()`. The
next step would then be overloading on the return type, which is related.
Walter made a conscious decision not to depend on return types for
overloads or any other deduction. It makes implementation of the
language simpler. C++ has tremendously complex rules for overloading
because of this.
Note, you can easily change this:
string x = getString();
to this:
auto x = getString!(char)();
which is still DRY.
One possible mechanism for working around this for when you already have
a variable is to change the return value into a parameter
void getStringParam(T)(ref T[] x) if ... { x = getString!(T)(); }
-Steve
