http://d.puremagic.com/issues/show_bug.cgi?id=7249

           Summary: 2.058 regression: "no size yet for forward reference"
                    when using this.init.tupleof
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: regression
          Priority: P2
         Component: DMD
        AssignedTo: nobody@puremagic.com
        ReportedBy: thecybershadow@gmail.com


--- Comment #0 from Vladimir Panteleev <thecybershadow@gmail.com> 2012-01-08 10:09:34 PST ---
struct S
{
    int x;
    alias typeof(this.init.tupleof[0]) T;
}

This worked in 2.057, but doesn't work in DMD git head.

-- 
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=7249



--- Comment #1 from Kenji Hara <k.hara.pg@gmail.com> 2012-01-08 15:05:22 PST ---
This causes by fixing bug 7190.

I think this is similar problem like follows:

struct S
{
    int x;
    alias typeof(this.init.sizeof) T;
    //  Error: struct test.S no size yet for forward reference
}

tupleof property requires complete semantic analysis to get the fields of a
type.
So it is meaningful forward reference, IMHO.

-- 
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=7249



--- Comment #2 from Vladimir Panteleev <thecybershadow@gmail.com> 2012-01-08 15:11:53 PST ---
In my case, the list of fields is passed as a template parameter:

struct S(string FIELDS)
{
    mixin(FIELDS);

    alias typeof(Fields.init.tupleof[0]) TypeOfFirstField;
}

There is a simple workaround available:

    struct Fields { mixin(FIELDS); }
    alias typeof(Fields.init.tupleof[0]) TypeOfFirstField;

So, if you think this is not a bug, I'm not against this change.

-- 
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=7249



--- Comment #3 from Vladimir Panteleev <thecybershadow@gmail.com> 2012-01-08 15:15:23 PST ---
(In reply to comment #2)
> In my case, the list of fields is passed as a template parameter:
> 
> struct S(string FIELDS)
> {
>     mixin(FIELDS);
> 
>     alias typeof(Fields.init.tupleof[0]) TypeOfFirstField;

Oops, last line should be:
      alias typeof(this.init.tupleof[0]) TypeOfFirstField;

-- 
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=7249



--- Comment #4 from Kenji Hara <k.hara.pg@gmail.com> 2012-01-08 15:25:00 PST ---
(In reply to comment #2)
> So, if you think this is not a bug, I'm not against this change.

Yes, I think this is not a bug.

Additionally, old behavior had a order dependent problem.

struct S
{
    alias typeof(this.init.tupleof[0]) T;
    // Error: array index [0] is outside array bounds [0 .. 0]
    // -> tupleof result was empty

    int x;

    alias typeof(this.init.tupleof[0]) T;
    // OK, T == int
}

Now they raise forward reference error as a same.
I think it's better consistency.

-- 
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=7249


Vladimir Panteleev <thecybershadow@gmail.com> changed:

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


--- Comment #5 from Vladimir Panteleev <thecybershadow@gmail.com> 2012-01-08 15:26:02 PST ---
OK, thanks for looking into this.

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