Looking at the syntax for defining a class, it seems that there's no way to have a class simply implement an interface...
Is that right? Or do you just put the interface declaration in place of the superclass declaration if there is no superclass?

-- 

-PIB

--
"C++ also supports the notion of *friends*: cooperative classes that
are permitted to see each other's private parts." - Grady Booch

In article <css9lq$1ek8$1@digitaldaemon.com>, Paul Bonser says...
>
>Looking at the syntax for defining a class, it seems that there's no way
>to have a class simply implement an interface...
>Is that right? Or do you just put the interface declaration in place of
>the superclass declaration if there is no superclass?
>

The documentation can be slightly misleading if one only looks at that syntax diagram, I'll agree.  Yes you can write a class that only implements an interface.  If no superclass is given in a class decleration, then DMD auto-magically makes it a child of std.object.Object and moves along.

-- Chris Sauls

"Paul Bonser" <misterpib@gmail.com> wrote in message news:css9lq$1ek8$1@digitaldaemon.com...
>
> Looking at the syntax for defining a class, it seems that there's no way
> to have a class simply implement an interface...
> Is that right?

No.

> Or do you just put the interface declaration in place of the superclass declaration if there is no superclass?

Yes. (If no superclass is specified, it is implicitly set to "Object"
anyway.)

Walter wrote:
> "Paul Bonser" <misterpib@gmail.com> wrote in message
> news:css9lq$1ek8$1@digitaldaemon.com...
> 
>>Looking at the syntax for defining a class, it seems that there's no way
>>to have a class simply implement an interface...
>>Is that right?
> 
> 
> No.
> 
> 
>>Or do you just put the interface declaration in place of
>>the superclass declaration if there is no superclass?
> 
> 
> Yes. (If no superclass is specified, it is implicitly set to "Object"
> anyway.)
> 
> 

Okay, that's what I was assuming.
So I may be asking all sorts of questions that are really picky like this because I am working on my own implementation of D right now (keeping the details to myself for now, hope to drop it as a pleasant surprise sometime soon).

In fact, I have another question.
In the enums section of the docs, the enum declaration is as follows:
EnumDeclaration:
	enum Identifier EnumBody
	enum EnumBody
	enum identifier : EnumBaseType EnumBody
	enum EnumBaseType : EnumBody

Why is the colon there after EnumBaseType in the last one? It seems like it's not neccesary, or am I missing the reasoning behind having it there?
Also, Identifier is not capitalized in the second to last line, but I'm pretty sure that's just a typo.

-- 
-PIB

--
"C++ also supports the notion of *friends*: cooperative classes that
are permitted to see each other's private parts." - Grady Booch

In article <cssigl$1o5c$1@digitaldaemon.com>, Paul Bonser says...
>In the enums section of the docs, the enum declaration is as follows:
>EnumDeclaration:
>	enum Identifier EnumBody
>	enum EnumBody
>	enum identifier : EnumBaseType EnumBody
>	enum EnumBaseType : EnumBody
>
>Why is the colon there after EnumBaseType in the last one? It seems like it's not neccesary, or am I missing the reasoning behind having it there? Also, Identifier is not capitalized in the second to last line, but I'm pretty sure that's just a typo.
>

I think the colon is a typo as well.  Having used that particular syntax a bit,
I believe the line should be
#
#  enum : EnumBaseType EnumBody
#

With the colon in the same place it would be were there an identifier.  Example:
#
#  enum : ubyte { FlagA = 0x01, FlagB = 0x02, FlagC = 0x04 };
#

-- Chris Sauls

That's right. I'll fix it. Thanks!