http://d.puremagic.com/issues/show_bug.cgi?id=9272

           Summary: opDispatch conflicts with UFCS on template functions
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody@puremagic.com
        ReportedBy: dransic@gmail.com


--- Comment #0 from Nicolas Sicard <dransic@gmail.com> 2013-01-05 04:57:36 PST ---
struct Foo {
    void opDispatch(string s)() {}
}

void bar(T)(Foo f) {}

unittest {
    Foo foo;
    bar!int(foo);  // OK
    foo.bar!int(); // Error: foo.opDispatch isn't a template
}

-- 
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=9272


Nils <nilsbossung@googlemail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |nilsbossung@googlemail.com
         Resolution|                            |INVALID


--- Comment #1 from Nils <nilsbossung@googlemail.com> 2013-01-05 06:33:55 PST ---
No bug here, I think. opDispatch takes precedence and just fails to compile.

The same thing happens with
---
struct Foo {
    void bar() {}
}
---

The error message could be improved though. foo.opDispatch is a template. It's foo.opDispatch!"bar" that is not.

-- 
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=9272



--- Comment #2 from Nicolas Sicard <dransic@gmail.com> 2013-01-05 09:05:00 PST ---
(In reply to comment #1)
> opDispatch takes precedence and just fails to compile.

Obviously. But is it really by design?

This works, where opDispatch does not take precedence:
---
struct Foo {
    void opDispatch(string s)() {}
}

void baz(Foo f) {}

unittest {
    Foo foo;
    foo.baz();     // OK
}
---

-- 
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=9272



--- Comment #3 from Kenji Hara <k.hara.pg@gmail.com> 2013-01-05 09:12:16 PST ---
(In reply to comment #2)
> (In reply to comment #1)
> > opDispatch takes precedence and just fails to compile.
> 
> Obviously. But is it really by design?
> 
> This works, where opDispatch does not take precedence:
> ---
> struct Foo {
>     void opDispatch(string s)() {}
> }
> 
> void baz(Foo f) {}
> 
> unittest {
>     Foo foo;
>     foo.baz();     // OK
> }
> ---

Just *template* opDispach is required for operator overloading. Therefore non-template function is simply ignored.

-- 
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=9272



--- Comment #4 from Nicolas Sicard <dransic@gmail.com> 2013-01-05 09:17:39 PST ---
(In reply to comment #3)
> (In reply to comment #2)
> > (In reply to comment #1)
> > > opDispatch takes precedence and just fails to compile.
> > 
> > Obviously. But is it really by design?
> > 
> > This works, where opDispatch does not take precedence:
> > ---
> > struct Foo {
> >     void opDispatch(string s)() {}
> > }
> > 
> > void baz(Foo f) {}
> > 
> > unittest {
> >     Foo foo;
> >     foo.baz();     // OK
> > }
> > ---
> 
> Just *template* opDispach is required for operator overloading. Therefore non-template function is simply ignored.

Yes. I understand why it doen't work at the moment. I just found it inconsistent. Maybe this should be an enhancement request then...

-- 
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=9272


Simen Kjaeraas <simen.kjaras@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |simen.kjaras@gmail.com


--- Comment #5 from Simen Kjaeraas <simen.kjaras@gmail.com> 2013-01-05 12:19:17 PST ---
(In reply to comment #2)
> This works, where opDispatch does not take precedence:
> ---
> struct Foo {
>     void opDispatch(string s)() {}
> }
> 
> void baz(Foo f) {}
> 
> unittest {
>     Foo foo;
>     foo.baz();     // OK
> }

Except, of course, opDispatch *does* take precedence here. Try this instead:

import std.stdio : writeln;

struct Foo {
    void opDispatch(string s)() { writeln("opDispatch: ", s); }
}

void baz(Foo f) { writeln("function: baz"); }

void main( ) {
    Foo foo;
    foo.baz();     // OK
}

I can assure you it prints opDispatch: baz.

-- 
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=9272



--- Comment #6 from Nicolas Sicard <dransic@gmail.com> 2013-01-06 04:49:57 PST ---
(In reply to comment #5)
> (In reply to comment #2)
> > This works, where opDispatch does not take precedence:
> > ---
> > struct Foo {
> >     void opDispatch(string s)() {}
> > }
> > 
> > void baz(Foo f) {}
> > 
> > unittest {
> >     Foo foo;
> >     foo.baz();     // OK
> > }
> 
> Except, of course, opDispatch *does* take precedence here. Try this instead:
> 
> import std.stdio : writeln;
> 
> struct Foo {
>     void opDispatch(string s)() { writeln("opDispatch: ", s); }
> }
> 
> void baz(Foo f) { writeln("function: baz"); }
> 
> void main( ) {
>     Foo foo;
>     foo.baz();     // OK
> }
> 
> I can assure you it prints opDispatch: baz.

Ah! I was trapped by the minimalism of my example. OK thanks.

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