http://d.puremagic.com/issues/show_bug.cgi?id=3492

           Summary: Can't overload nested functions
           Product: D
           Version: 2.036
          Platform: Other
        OS/Version: Windows
            Status: NEW
          Keywords: spec
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody@puremagic.com
        ReportedBy: dsimcha@yahoo.com


--- Comment #0 from David Simcha <dsimcha@yahoo.com> 2009-11-09 20:35:03 PST ---
Nested functions don't overload like non-nested functions.  If this behavior is correct, it seems like a fairly arbitrary limitation, but it should be clarified in the spec.

void main() {
    static void foo(uint a, float b) {}
    static void foo(float a, uint b) {}
}

Error: declaration foo is already defined

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492



--- Comment #1 from github-bugzilla@puremagic.com 2012-01-23 01:42:44 PST ---
Commit pushed to https://github.com/D-Programming-Language/d-programming-language.org

https://github.com/D-Programming-Language/d-programming-language.org/commit/5d98834d711cebd4d218a0580856b84de6dcb10f fix Issue 3492 - Can't overload nested functions

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492


Walter Bright <bugzilla@digitalmars.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |bugzilla@digitalmars.com
         Resolution|                            |FIXED


-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492


Andrej Mitrovic <andrej.mitrovich@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |andrej.mitrovich@gmail.com


--- Comment #2 from Andrej Mitrovic <andrej.mitrovich@gmail.com> 2012-02-26 14:52:16 PST ---
Nested functions are a great feature, especially when experimenting with code and when I need to do some quick filtering over e.g. hashes but need to access some outer state. With nested functions I don't have to use global variables or use ref parameters (I'm talking about non-static functions).

I've just tried using overloaded functions but failed and found this in bugzilla. It's not a big issue, I can rename the functions.. but I'd like to know the rationale for this limitation.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492



--- Comment #3 from Walter Bright <bugzilla@digitalmars.com> 2012-02-26 22:04:30 PST ---
The reason is straightforward - there's no point to it. Nested functions tend to be right next to where they are used. They're not off in a separate file.

We shouldn't have features just to have features. They have to have an identifiable benefit. Overloaded functions introduce complexity both to the user and the compiler - and there is no clear benefit for nested functions.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492



--- Comment #4 from Walter Bright <bugzilla@digitalmars.com> 2012-02-26 22:13:25 PST ---
The reason is straightforward - there's no point to it. Nested functions tend to be right next to where they are used. They're not off in a separate file.

We shouldn't have features just to have features. They have to have an identifiable benefit. Overloaded functions introduce complexity both to the user and the compiler - and there is no clear benefit for nested functions.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492


Jonathan M Davis <jmdavisProg@gmx.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jmdavisProg@gmx.com


--- Comment #5 from Jonathan M Davis <jmdavisProg@gmx.com> 2012-02-26 22:43:51 PST ---
Really? I'm constantly annoyed by that you can't overload nested functions when writing unit tests. I frequently need to overload functions in unittest blocks and have to play games to get around the fact that I can't. The worst is when you use foreach and TypeTuple. You can't just do something like

foreach(T; TypeTuple!(string, wstring, dstring))
{
    void test(T actual, T expected, ...)
    {
    }

    test(...);
    test(...);
    test(...);
    ...
}

because then every subsequent definition of the nested function conflicts with the rest. You're force to use string mixins simply to give each overload of the nested function a unique name. It's an extra complication that really shouldn't be necessary IMHO.

If anything, I'd argue that the lack of ability to overload nested functions is an unnecassary inconsistency in the language.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492


timon.gehr@gmx.ch changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |timon.gehr@gmx.ch


--- Comment #6 from timon.gehr@gmx.ch 2012-02-27 05:16:59 PST ---
(In reply to comment #4)
> The reason is straightforward - there's no point to it. Nested functions tend to be right next to where they are used. They're not off in a separate file.
> 
> We shouldn't have features just to have features.

But we do have overloading in the language already. We shouldn't just deny access to existing features. Chances are that a compiler works better if there are no arbitrary inconsistencies (see all the bugs related to parsing storage classes of nested declarations for an example -- just define the language such that reusing parsing procedures is easy.)

> They have to have an identifiable benefit. Overloaded functions introduce complexity both to the user and the compiler - and there is no clear benefit for nested functions.

What introduces complexity to the user and the compiler is that the syntax (and, like in this case, even the semantics!) of function local declarations is substantially different from global declarations and declarations inside aggregates. It is not 'turtles all the way down', it is not consistent, and it is not useful. Those are arbitrary limitations. I'd suspect that DMD can even handle correctly most of the function local features that are currently invalid syntax.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=3492



--- Comment #7 from Walter Bright <bugzilla@digitalmars.com> 2012-02-27 10:48:35 PST ---
>because then every subsequent definition of the nested function conflicts with
the rest.

You can make test() a template function.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------