www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Multiple functions, same signature

reply =?UTF-8?B?THXDrXM=?= Marques <luis luismarques.eu> writes:
I was surprised to find out today that this compiles:

void foo() {}
void foo() {}
void main() {}

Is it a bug, or just a weird design decision? "alphaglosined" on 
IRC seemed to think it was a regression. Please confirm, so that 
I can file a bug, or understand the design decision rationale.
Jul 11 2018
next sibling parent reply John Colvin <john.loughran.colvin gmail.com> writes:
On Wednesday, 11 July 2018 at 15:58:05 UTC, Luís Marques wrote:

 I was surprised to find out today that this compiles:

 void foo() {}
 void foo() {}
 void main() {}

 Is it a bug, or just a weird design decision? "alphaglosined" 
 on IRC seemed to think it was a regression. Please confirm, so 
 that I can file a bug, or understand the design decision 
 rationale.


Definitely a change, but it always compiled, it just used to fail 
to link

https://run.dlang.io/is/b0JxD9
Jul 11 2018
parent =?UTF-8?B?THXDrXM=?= Marques <luis luismarques.eu> writes:
On Wednesday, 11 July 2018 at 16:01:48 UTC, John Colvin wrote:

 On Wednesday, 11 July 2018 at 15:58:05 UTC, Luís Marques wrote:
 Definitely a change, but it always compiled, it just used to 
 fail to link


Do you know why the frontend doesn't complain about a 
redefinition, like C++ does?
Jul 11 2018
prev sibling next sibling parent rikki cattermole <rikki cattermole.co.nz> writes:
On 12/07/2018 3:58 AM, Luís Marques wrote:

 I was surprised to find out today that this compiles:
 
 void foo() {}
 void foo() {}
 void main() {}
 
 Is it a bug, or just a weird design decision? "alphaglosined" on IRC 
 seemed to think it was a regression. Please confirm, so that I can file 
 a bug, or understand the design decision rationale.


The reason I think that it is a regression is because of [0].

Either the change log didn't include some changes, environment/linker or 
its a regression IMHO.

```
Up to      2.071.2: Failure with output:
-----
onlineapp.o: In function `_D9onlineapp3fooFiZv':
/sandbox/onlineapp.d:2: multiple definition of `_D9onlineapp3fooFiZv'
onlineapp.o:onlineapp.d:(.text._D9onlineapp3fooFiZv+0x0): first defined here
collect2: error: ld returned 1 exit status
--- errorlevel 1
-----

2.072.2 to 2.074.1: Failure with output:
-----
onlineapp.o: In function `_D9onlineapp3fooFiZv':
/sandbox/onlineapp.d:2: multiple definition of `_D9onlineapp3fooFiZv'
onlineapp.o:onlineapp.d:(.text._D9onlineapp3fooFiZv+0x0): first defined here
collect2: error: ld returned 1 exit status
Error: linker exited with status 1
-----

Since      2.075.1: Success and no output
```


[0] https://run.dlang.io/is/AGuM6P
Jul 11 2018
prev sibling next sibling parent reply Jacob Carlborg <doob me.com> writes:
On 2018-07-11 17:58, Luís Marques wrote:

 I was surprised to find out today that this compiles:
 
 void foo() {}
 void foo() {}
 void main() {}
 
 Is it a bug, or just a weird design decision? "alphaglosined" on IRC 
 seemed to think it was a regression. Please confirm, so that I can file 
 a bug, or understand the design decision rationale.


You'll get an error if you call "foo".

-- 
/Jacob Carlborg
Jul 11 2018
parent =?UTF-8?B?THXDrXM=?= Marques <luis luismarques.eu> writes:
On Wednesday, 11 July 2018 at 16:21:26 UTC, Jacob Carlborg wrote:

 You'll get an error if you call "foo".


I understand that. Still seems like something that the frontend 
should detect, unless there's a good use case for multiple 
(re)definitions. If I had to guess, it's probably to support 
generic code which may have the same or different type modifiers 
applied, possibly resulting in a collision.

In my specific case, due to the functions being manipulated by 
generic code, it led to a harder to diagnose bug, since the 
functions weren't called directly (or at all, of course). 
Instead, the openmethods library builds a list of dispatch 
targets, but it didn't detect at compile-time that the targets 
were redundant, which led to a *runtime* error, and a harder to 
diagnose situation. openmethods.d could be changed to deal with 
that, but it seems like the frontend would be the ideal place to 
check it, unless D should really support redundant functions.
Jul 11 2018
prev sibling parent Seb <seb wilzba.ch> writes:
On Wednesday, 11 July 2018 at 15:58:05 UTC, Luís Marques wrote:

 I was surprised to find out today that this compiles:

 void foo() {}
 void foo() {}
 void main() {}

 Is it a bug, or just a weird design decision? "alphaglosined" 
 on IRC seemed to think it was a regression. Please confirm, so 
 that I can file a bug, or understand the design decision 
 rationale.


This will be deprecated soon: 
https://github.com/dlang/dmd/pull/8429
Jul 11 2018