• J C Calvarese (33/33) Nov 05 2002 Hi.
• Burton Radons (4/9) Nov 05 2002 Yes, you'd need to recompile Phobos; structs have hidden variable data
• J C Calvarese (6/19) Nov 05 2002 Thanks. That explains it. I'll keep them in the main source. It
• Walter (7/11) Nov 23 2002 That's because D stores an initializer for the struct in the module that

J C Calvarese <jccalvarese yahoo.com> writes:

Hi.

Most of my experience has been in varous forms of BASIC, but I've also 
programmed in Pascal and Java.  I'm currently learning D.  I find the 
syntax a lot easier to understand than C++.

I've run into a problem.  I'm using Walter's GUI sample as a framework 
and adding more Win32 API controls.  I can get it to compile and run 
fine if I declare these structs (LVITEMA, LVCOLUMNA, SIZE, etc.) in the 
winsamp.d file, but the instant I move them to the dmd\src\windows.d 
file, my success ends.  The compiler seems to understand what I did, but 
the linker seems to be angry at me:

winsamp,,,user32.lib+gdi32.lib+comctl32.lib+user32+kernel32,winsamp.def/noi;
OPTLINK (R) for Win32  Release 7.50B1
Copyright (C) Digital Mars 1989 - 2001  All Rights Reserved

winsamp.obj(winsamp)
  Error 42: Symbol Undefined __init_windows_LVITEMA
winsamp.obj(winsamp)
  Error 42: Symbol Undefined __init_windows_LVCOLUMNA
winsamp.obj(winsamp)
  Error 42: Symbol Undefined __init_windows_SIZE
winsamp.obj(winsamp)
  Error 42: Symbol Undefined __init_windows_INITCOMMONCONTROLSEX
winsamp.obj(winsamp)
  Error 42: Symbol Undefined __init_windows_TOOLINFOA
winsamp.obj(winsamp)
  Error 42: Symbol Undefined __init_windows_SETTEXTEX
--- errorlevel 6


I'm probably trying to do something stupid.  Do I need to recompile 
phobos?  Do I need to develop initializer methods for these structures? 
  I don't have a clue why it works perfectly when I declare these 
structs in the main program, but it falls apart when I try to put them 
somewhere else.

Any help would be much appreciated.
Justin

Burton Radons <loth users.sourceforge.net> writes:

J C Calvarese wrote:
 I'm probably trying to do something stupid.  Do I need to recompile 
 phobos?  Do I need to develop initializer methods for these structures? 
  I don't have a clue why it works perfectly when I declare these structs 
 in the main program, but it falls apart when I try to put them somewhere 
 else.

Yes, you'd need to recompile Phobos; structs have hidden variable data 
for their initial content.  It's best to keep these structs in your 
program for now.

J C Calvarese <jccalvarese yahoo.com> writes:

Thanks.  That explains it.  I'll keep them in the main source.  It 
struck me as strange that I can move constants wherever I want (as long 
as I import it) and D doesn't complain, but the D system can't find my 
structs when I move them.

Justin

Burton Radons wrote:
 J C Calvarese wrote:
 
 I'm probably trying to do something stupid.  Do I need to recompile 
 phobos?  Do I need to develop initializer methods for these 
 structures?  I don't have a clue why it works perfectly when I declare 
 these structs in the main program, but it falls apart when I try to 
 put them somewhere else.

 
 
 Yes, you'd need to recompile Phobos; structs have hidden variable data 
 for their initial content.  It's best to keep these structs in your 
 program for now.
 

"Walter" <walter digitalmars.com> writes:

"J C Calvarese" <jccalvarese yahoo.com> wrote in message
news:3DC8928F.3070700 yahoo.com...
 Thanks.  That explains it.  I'll keep them in the main source.  It
 struck me as strange that I can move constants wherever I want (as long
 as I import it) and D doesn't complain, but the D system can't find my
 structs when I move them.

That's because D stores an initializer for the struct in the module that
declares it, then when you use a struct, it initializes it with that
initializer. The initializer isn't really necessary if the values are all 0,
and the compiler shouldn't demand one if so, but the current one does.

-Walter