www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 3367] New: Regression: assignment at declaration no longer supports opAssign or ctor overloads

reply d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=3367

           Summary: Regression: assignment at declaration no longer
                    supports opAssign or ctor overloads
           Product: D
           Version: 2.033
          Platform: x86
        OS/Version: Windows
            Status: NEW
          Keywords: rejects-valid
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody puremagic.com
        ReportedBy: sandford jhu.edu


--- Comment #0 from Rob Jacques <sandford jhu.edu> 2009-10-05 20:45:10 PDT ---
In DMD 2.032, when a struct was assigned to at it's declaration point, opAssign
overloads were called if other alternative were not available. In DMD 2.033,
neither opAssign or ctor overloads are called, only static opCall works

struct Foo {
    int x;
//    static opCall(int v) { // Un-comment this and the below works
//        Foo f;
//        f.x = v;
//        return f;
//    }
    this(int v){ x = v; }
    void opAssign(int v){
        x = v;
        return this;
    }
}

void main(char[][] args) {

    int y = 5;
    Foo f = y;  // fails
    f = y;      // okay
    f = Foo(y); // okay

}

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Oct 05 2009
next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=3367


Don <clugdbug yahoo.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |clugdbug yahoo.com.au
            Summary|Regression: assignment at   |Regression: struct
                   |declaration no longer       |initialization no longer
                   |supports opAssign or ctor   |supports ctor overloads
                   |overloads                   |


--- Comment #1 from Don <clugdbug yahoo.com.au> 2009-10-06 01:14:23 PDT ---
My fault. This is a consquence of the fix to bug 2702. Previously it used to
accept any old garbage inside a struct initializer; now it only accepts static
opCall. It should definitely allow constructors as well. 

Based on my reading of the spec, the fact that opAssign used to work seems to
have been a bug:
"assignment is defined as copying the contents of one object over another,
already initialized, type" (struct.html)

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Oct 06 2009
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=3367


Sobirari Muhomori <dfj1esp02 sneakemail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |regression


-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Oct 06 2009
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=3367


Don <clugdbug yahoo.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


--- Comment #2 from Don <clugdbug yahoo.com.au> 2009-10-29 12:58:44 PDT ---
PATCH: declaration.c, line 1094.
            {
            /* Look for opCall
             * See bugzilla 2702 for more discussion
             */
            Type *ti = ei->exp->type->toBasetype();
------- ADD THIS CODE:            
            // Look for ctor
            if (sd->ctor && 
                /* Initializing with the same type is done differently
                 */
                !(ti->ty == Tstruct && t->toDsymbol(sc) == ti->toDsymbol(sc)))
{
               // Rewrite as e1.call(arguments)
                Expression * eCall = new DotIdExp(loc, e1, Id::ctor);
                ei->exp = new CallExp(loc, eCall, ei->exp);
            } 
            else
-----------
            // Don't cast away invariant or mutability in initializer
            if (search_function(sd, Id::call) &&

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Oct 29 2009
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=3367


Leandro Lucarella <llucax gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |llucax gmail.com


--- Comment #3 from Leandro Lucarella <llucax gmail.com> 2009-11-04 06:39:53
PST ---
SVN commit: http://www.dsource.org/projects/dmd/changeset/235

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Nov 04 2009
prev sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=3367


Walter Bright <bugzilla digitalmars.com> changed:

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


--- Comment #4 from Walter Bright <bugzilla digitalmars.com> 2009-11-06
11:34:23 PST ---
Fixed dmd 2.036

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Nov 06 2009