http://d.puremagic.com/issues/show_bug.cgi?id=9835

           Summary: DynamicArrayTypeOf and isDynamicArray work
                    inconsistently
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P2
         Component: Phobos
        AssignedTo: nobody@puremagic.com
        ReportedBy: tobias@pankrath.net


--- Comment #0 from Tobias Pankrath <tobias@pankrath.net> 2013-03-30 06:48:11 PDT ---
---
struct Wrapper {
    int[] data;
    alias data this;
};

pragma(msg, DynamicArrayTypeOf!Wrapper) // int[]
pragma(msg, isDynamicArray!Wrapper) // false
---

There are two issues here: The first one is that DynamicArrayTypeOf and isDynamicArray should work consistently that is, isDynamicArray!T should be true if and only if DynamicArrayTypeOf!T yields a type.

Second according to TDPL alias this makes Wrapper a subtype of int[] and isDynamicArray!Wrapper should be true or otherwise isDynamicArray is quite useless as a template constraint like in std.array.popFront.

-- 
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=9835



--- Comment #1 from Kenji Hara <k.hara.pg@gmail.com> 2013-03-30 07:12:21 PDT ---
Current behavior is completely intended, so it's not a bug.

In old days, I thought that std.traits.isSomething should consider alias this types. That was necessary for std.format and std.conv implementation. For the purpose I added std.traits.SomethingTypeOf, and used them as the implementation of isSomething.

https://github.com/D-Programming-Language/phobos/pull/488

But, the change of isSomething's behavior had been complained as a bug by some
peoples. They said "isDymanicArray!T should be true if and only if T is
actually a dynamic array type. If T is a struct with alias this, it should
return false".
I finally agreed with that and reverted the behavior.

https://github.com/D-Programming-Language/phobos/pull/976

Now, std.format and std.conv uses is(SomethingTypeOf!T).

Furthermore, SomethingTypeOf is now *undocumented* templates, so their specifications are "as-is". (But considering "alias this" type is sometimes useful, so they are not marked wit "package".)

If you really need SomethingTypeOf, please open an enhancement that they should be *documented*.

-- 
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=9835


Kenji Hara <k.hara.pg@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID


--- Comment #2 from Kenji Hara <k.hara.pg@gmail.com> 2013-03-30 07:13:59 PDT ---
I think that both isSomething and SomethingTypeOf should not be changed.

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