std.boxer problem

[Daniel Keep]

Tydr Schnubbis wrote:
> John Demme wrote:
>> Tydr Schnubbis wrote:
>>
>>> Alexander Panek wrote:
>>>> Tydr Schnubbis wrote:
>>>>> Why can't I cast from Object to Box?  That seems to defeat the purpose
>>>>> of boxing.  I get this error:
>>>>>
>>>>> serverlist.d(185): e2ir: cannot cast from object.Object to Box
>>>>> Internal error: e2ir.c 267
>>>>
>>>> Object o = new Object();
>>>> Box b = box(o);
>>>>
>>>> foo bar = unbox!(foo)(b);
>>>>
>>> No luck.
>>>
>>>
>>> import std.boxer;
>>>
>>> void main()
>>> {
>>>      Object o = new Object();
>>>      Box b = box(o);
>>>
>>>      int bar = unbox!(int)(b);
>>> }
>>>
>>> c:\prog\test\D>dmd boxing2.d
>>> c:\prog\dmd\bin\..\..\dm\bin\link.exe boxing2,,,user32+kernel32/noi;
>>> OPTLINK (R) for Win32  Release 7.50B1
>>> Copyright (C) Digital Mars 1989 - 2001  All Rights Reserved
>>>
>>> boxing2.obj(boxing2)
>>>   Error 42: Symbol Undefined _assert_3std5boxer
>>> --- errorlevel 1
>>>
>>>
>>> --------
>>> Maybe it's just a configuration problem?  I've got dmd 0.155.
>>
>> That one's a bug with templating code inside of libraries.  You need to
>> compile your code with -release.
> 
> Ah, so I guess it's currently unusable for real projects, then.  I
> solved my original problem without using boxing.  So I'm okay for now. :)

If I remember correctly, it's because we don't have a debug build of
Phobos.  You can build one yourself (ie: rebuild phobos with debug stuff
in).  Or, you can do what I did and "provide" the function that it's
missing:

// Burton Radons <burton-radons[at]smocky[dot]com>
// This entry allows -unittest for testing, and/or no need
// for the -release switch
private extern (C) void assert_3std5boxer() {}

I personally just stuck that in a file called
"dmdbug_std_boxer_assert.d", so that I can just remove the import when
it gets fixed :)

	-- Daniel

-- 

v1sw5+8Yhw5ln4+5pr6OFma8u6+7Lw4Tm6+7l6+7D
a2Xs3MSr2e4/6+7t4TNSMb6HTOp5en5g6RAHCP    http://hackerkey.com/