About string and int spliced together.

[Adam D. Ruppe]

On Saturday, 20 June 2020 at 11:09:12 UTC, aberba wrote:
> So has D gotten so complicated under the hood that its not 
> possible to d:

In that proposal, i"" is a string BUILDER, not a string. So it 
returns an object you can get a string out of, or do other things 
with too.

C# and Javascript also work this way, you just don't notice it as 
much due to their different typing systems: the C# object 
implicitly converts to string on demand, inserting a hidden call 
to its toString method when you assign it like that, and the 
Javascript one is passed to a function that does the conversion 
with special built-in syntax. JS' foo`$bar` actually calls the 
function foo to convert the string builder object to whatever you 
want, and if you don't specify a function, the language inserts a 
hidden toString function for you automatically.

The DIP there works the same way as those, just without the 
hidden function calls. You need to choose what function you want 
to call yourself.

// so this works since you call the idup method yourself
// telling it to copy the result of the builder into a
// plain string
string sentence = i"I and ${expression}".idup;

Or you can call other functions and avoid copying for maximum 
efficiency and flexibility - no hidden functions to surprise the 
low-level D crowd while only needing a very simple function call 
if you don't care about that and just want to keep it simple.

Similar to how D's map, filter, etc. may need the extra call to 
.array in some use cases. It looks like extra work at first 
glace, but it actually enables better code once you get to know 
it.