www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 9281] New: Enum struct with op overloading doesnt works

reply d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=9281

           Summary: Enum struct with op overloading doesnt works
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: regression
          Priority: P2
         Component: DMD
        AssignedTo: nobody puremagic.com
        ReportedBy: kozzi11 gmail.com


--- Comment #0 from Daniel Kozak <kozzi11 gmail.com> 2013-01-08 01:42:33 PST ---
module main;

import std.algorithm;
import std.array;

immutable struct Column {
    string opAssign(V)(V tValue) {
        return tValue;
    }
}

immutable test1 = Column();
enum test2 = Column(); 

void main(string[] args)
{
    string where = test1 = "something"; // works ok
    std.stdio.writeln(where);
    where = test2 = "something else"; // works 2.060, dont compile on 2.061
    std.stdio.writeln(where);
}

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jan 08 2013
next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=9281



--- Comment #1 from Kenji Hara <k.hara.pg gmail.com> 2013-01-10 06:22:48 PST ---
Did this really work with 2.060? I cannot reproduce the "works 2.060".

Column.opAssign is an immutable member function, then we can call it from
immutable object test1, but cannot call from mutable object test2.

So, the error in test2 = "something else" is correct, as far as I know.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jan 10 2013
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=9281



--- Comment #2 from Daniel Kozak <kozzi11 gmail.com> 2013-01-10 07:42:45 PST ---
(In reply to comment #1)
 Did this really work with 2.060? I cannot reproduce the "works 2.060".
 
 Column.opAssign is an immutable member function, then we can call it from
 immutable object test1, but cannot call from mutable object test2.
 
 So, the error in test2 = "something else" is correct, as far as I know.

Yes my fault, I try to simplify too much. Here is more detailed description Code which works on 2.060 and don`t compile on 2.061 module main; import std.algorithm; import std.array; immutable struct Column { string opAssign(V)(V tValue) { return tValue; } } class Ob2 { enum : Column { COLUM_A = Column() } immutable COLUMNS = [ COLUM_A, ]; } immutable test1 = Column(); void main(string[] args) { string where = test1 = "something"; // works ok std.stdio.writeln(where); where = (Ob2.COLUM_A = "something else"); // works 2.060, dont compile on 2.061 std.stdio.writeln(where); } However I find out more interesing thing. This code is almost same however it doesn`t work on 2.061 neither 2.060: module main; import std.algorithm; import std.array; immutable struct Column { string opAssign(V)(V tValue) { return tValue; } } class Ob2 { enum : Column { COLUM_A = Column() } // comment this code make it not compilable in 2.060 /*immutable COLUMNS = [ COLUM_A, ];*/ } immutable test1 = Column(); void main(string[] args) { string where = test1 = "something"; // works ok std.stdio.writeln(where); where = (Ob2.COLUM_A = "something else"); // dont compile on 2.060 and 2.061 std.stdio.writeln(where); } -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jan 10 2013
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=9281



--- Comment #3 from Kenji Hara <k.hara.pg gmail.com> 2013-01-10 22:52:17 PST ---
(In reply to comment #2)
 Yes my fault, I try to simplify too much. Here is more detailed description
 
 Code which works on 2.060 and don`t compile on 2.061
 

The reason why (Ob2.COLUM_A = "something else") doesn't work in 2.061 is same as in comment #1. typeof(Ob2.COLUM_A) is a mutable Column, and cannot call immutable opAssign from that.
 However I find out more interesing thing. This code is almost same however it
 doesn`t work on 2.061 neither 2.060:
 

I think this _was_ an accepts-invalid bug in 2.060. If you define Ob2.COLUMNS in 2.060, Ob2.COLUM_A is _incorrectly_ typed as immutable(Column). This is definitely a bug (But I don't know what change is fixed the bug in 2.060). So, this is not a regression. -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jan 10 2013
prev sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=9281


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

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


--- Comment #4 from Kenji Hara <k.hara.pg gmail.com> 2013-01-10 23:16:33 PST ---
(In reply to comment #3)
 I think this _was_ an accepts-invalid bug in 2.060.
 If you define Ob2.COLUMNS in 2.060, Ob2.COLUM_A is _incorrectly_ typed as
 immutable(Column). This is definitely a bug (But I don't know what change is
 fixed the bug in 2.060).

I found a commit which the behavior is changed. https://github.com/D-Programming-Language/dmd/commit/e01eb59f842dfe7a5275d96c420691c4a64f57f4 The root cause of the bug 5779 was an optimizer bug. If the type of an optimized result is different from the type of a source expression, the source expression type had been accidentally modified. In this case, the declaration of COLUMNS invokes optimizer on the expression [ COLUM_A, ], and it accidentally modified the type of COLUM_A to immutable(Column). So the conclusion is: the original code had an accepts-invalid bug, and it was already fixed in 2.061. I'll mark this as "resolved-invalid" bug. -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Jan 10 2013