I just came across this. Can't believe there's not been more people finding this:

char[] buf = new char[100];
buf.formattedWrite("This is a number: %s", 5); // Error cannot put a const(char)[] into a char[]


OK, challenge accepted!

buf[0 .. myName.length] = myName[]; // Hey phobos, look, this is easy!

But no, of course -- autodecoding. >:(

https://issues.dlang.org/show_bug.cgi?id=18790

-Steve

On Saturday, 21 April 2018 at 22:34:39 UTC, Steven Schveighoffer wrote:
> I just came across this. Can't believe there's not been more people finding this:
>
> char[] buf = new char[100];
> buf.formattedWrite("This is a number: %s", 5); // Error cannot put a const(char)[] into a char[]
>
>
> OK, challenge accepted!
>
> buf[0 .. myName.length] = myName[]; // Hey phobos, look, this is easy!
>
> But no, of course -- autodecoding. >:(
>
> https://issues.dlang.org/show_bug.cgi?id=18790
>
> -Steve

"Use .representation" would be the official answer for the second hack.
But this should absolutely work. It looks this looks the same thing as
https://issues.dlang.org/show_bug.cgi?id=18472#c10
---
enum s = "%1$s,%2$s".format("foo","bar");
---
phobos/std/range/primitives.d(405): Error: static assert:  "Cannot put a const(char)[] into a Appender!string."
phobos/std/format.d(1184):        instantiated from here: put!(Appender!string, const(char)[])
/phobos/std/format.d(473):        instantiated from here: writeUpToNextSpec!(Appender!string)
phobos/std/format.d(6168):        instantiated from here: formattedWrite!(Appender!string, char, string, string)
main.d(5):        instantiated from here: format!(char, string, string)

18472 broke compilation of DCompute under -betterC when upgrading LDC to a newer DMDFE, I "fixed" it by not compiling with -betterC but then Mike Franklin (thanks!) discovered the above error after fixing the -betterC side of things.

I'm going to be extremely unhappy if I can't compile dcompute at all with the latest LDC.
IMO this is a priority 1 regression.

On 4/21/18 7:30 PM, Nicholas Wilson wrote:

> "Use .representation" would be the official answer for the second hack.

I have since discovered that I can use byCodeUnit, but I'm still annoyed that this is required.

I *think* that this is the only problem:

https://github.com/dlang/phobos/blob/master/std/range/primitives.d#L365

Essentially, here is the block:

    //Optional optimization block for straight up array to array copy.
    else static if (isDynamicArray!R && !isNarrowString!R && isDynamicArray!E && is(typeof(r[] = e[])))
    {
        immutable len = e.length;
        r[0 .. len] = e[];
        r = r[len .. $];
    }

If we just take out the !isNarrowString!R, I think it will just work, and probably cause no problems. I have NO IDEA why we are purposely shooting ourselves in the foot here.

I'm going to at least try it (PR forthcoming).

There are a few bits inside put that are supposed to cater to narrow strings, I'm not sure why it's not working.

-Steve

On 4/21/18 7:56 PM, Steven Schveighoffer wrote:

> I'm going to at least try it (PR forthcoming).

https://github.com/dlang/phobos/pull/6471

-Steve