Hi,
I'm new to D and played a bit with templates and delegates.
Now i discovered some behaviore that i don't understand.
Can somebody explain me why i get two different outputs?


import std.stdio;


interface A(T){
	bool GetBool();
	T getT();
}

class C:A!(double){
	override bool GetBool(){
		return false;
	}
	override double getT(){
		return 1;
	}
}

void mwriteln(T)( A!T delegate() dg){
	writeln(dg().getT());
}

void main()
{
	auto c = new C();
	writeln(c.getT());
	mwriteln!double({return new C();});
}

On Mon, 31 Mar 2014 18:58:30 +0000, anonymous wrote:

> Hi,
> I'm new to D and played a bit with templates and delegates.
> Now i discovered some behaviore that i don't understand. Can somebody
> explain me why i get two different outputs?
> 
> 
> import std.stdio;
> 
> 
> interface A(T){
> 	bool GetBool();
> 	T getT();
> }
> 
> class C:A!(double){
> 	override bool GetBool(){
> 		return false;
> 	}
> 	override double getT(){
> 		return 1;
> 	}
> }
> 
> void mwriteln(T)( A!T delegate() dg){
> 	writeln(dg().getT());
> }
> 
> void main()
> {
> 	auto c = new C(); writeln(c.getT()); mwriteln!double({return new
> 	C();});
> }

Looks like a bug.  Here's a cleaned-up test case:

-----------------------------------------------------------
interface A(T)
{
	T getT();
}

class C : A!double
{
	double getT(){ return 1; }
}

void mwriteln(T)(A!T delegate() dg)
{
	import std.stdio;
	auto a = dg();
	writeln("Checkpoint 1");
	assert(a !is null);
	writeln(a);
	writeln("Checkpoint 2");
}

void main()
{
	import std.traits;
	static assert(isImplicitlyConvertible!(C, A!double));
	mwriteln!double({return new C();});
}
-----------------------------------------------------------

Backtrace:
-----------------------------------------------------------
Program received signal SIGSEGV, Segmentation fault.
#0  0x000000000042ac8c in _d_interface_cast ()
#1  0x0000000000427ec6 in
std.format.__T11formatValueTS3std5stdio4File17LockingTextWriterTC4test8__T1ATdZ1ATaZ.formatValue
() ()
#2  0x0000000000427e23 in
std.format.__T13formatGenericTS3std5stdio4File17LockingTextWriterTC4test8__T1ATdZ1ATaZ.formatGeneric
() ()
#3  0x0000000000427d29 in
std.format.__T14formattedWriteTS3std5stdio4File17LockingTextWriterTaTC4test8__T1ATdZ1AZ.formattedWrite
() ()
#4  0x0000000000427871 in
std.stdio.File.__T5writeTC4test8__T1ATdZ1ATaZ.write() ()
#5  0x00000000004277d5 in std.stdio.__T7writelnTC4test8__T1ATdZ1AZ.writeln
() ()
#6  0x000000000042771f in test.__T8mwritelnTdZ.mwriteln() ()
#7  0x00000000004276a9 in D main ()
#8  0x000000000042b61c in rt.dmain2._d_run_main() ()
#9  0x000000000042b576 in rt.dmain2._d_run_main() ()
#10 0x000000000042b5dc in rt.dmain2._d_run_main() ()
#11 0x000000000042b576 in rt.dmain2._d_run_main() ()
#12 0x000000000042b4f7 in _d_run_main ()
#13 0x0000000000429c5f in main ()
-----------------------------------------------------------

On Mon, 31 Mar 2014 14:58:30 -0400, anonymous <non@trash-mail.com> wrote:

> Hi,
> I'm new to D and played a bit with templates and delegates.
> Now i discovered some behaviore that i don't understand.
> Can somebody explain me why i get two different outputs?
>
>
> import std.stdio;
>
>
> interface A(T){
> 	bool GetBool();
> 	T getT();
> }
>
> class C:A!(double){
> 	override bool GetBool(){
> 		return false;
> 	}
> 	override double getT(){
> 		return 1;
> 	}
> }
>
> void mwriteln(T)( A!T delegate() dg){
> 	writeln(dg().getT());
> }
>
> void main()
> {
> 	auto c = new C();
> 	writeln(c.getT());
> 	mwriteln!double({return new C();});
> }

This is definitely a bug.

Reduced case:

import std.stdio;

interface A{
    void foo();
}

class C:A{
    override void foo(){
        writeln("here");
    }
}

void x( A delegate() dg){
    dg().foo();
}

void main()
{
    A c = new C;
    c.foo(); // prints "here"
    x({A a = new C; return a;}); // prints "here"
    x({return new C;}); // does not print
}

This is really an issue with delegate return type inferrence not working properly.

-Steve

Ok, thought i did something wrong or got some wrong idea how it should work.

Is this bug allready reported? or can somebody who has a deeper insight to this report it?

On Tuesday, 1 April 2014 at 05:51:46 UTC, anonymous wrote:
> Ok, thought i did something wrong or got some wrong idea how it should work.

On Tue, 01 Apr 2014 15:47:42 -0400, anonymous <non@trash-mail.com> wrote:

> Is this bug allready reported? or can somebody who has a deeper insight to this report it?

I don't know. I think you should report it. If it's already reported, someone will close it as a "duplicate"

-Steve

On Tuesday, 1 April 2014 at 19:55:05 UTC, Steven Schveighoffer wrote:
> On Tue, 01 Apr 2014 15:47:42 -0400, anonymous <non@trash-mail.com> wrote:
>
>> Is this bug allready reported? or can somebody who has a deeper insight to this report it?
>
> I don't know. I think you should report it. If it's already reported, someone will close it as a "duplicate"
>
> -Steve

I filed it.

https://d.puremagic.com/issues/show_bug.cgi?id=12508

Kenji Hara