• bobef (2/2) Jan 06 2007 Hmmm... this was like... one year and 30 DMD versions ago. And in D 1.0 ...
• Sean Kelly (11/13) Jan 06 2007 I'm not sure this is an easy one to fix. In this case, the DLL contains...

bobef <a a.com> writes:

Hmmm... this was like... one year and 30 DMD versions ago. And in D 1.0 it is
still not fixed...
http://www.digitalmars.com/d/archives/digitalmars/D/34291.html

Sean Kelly <sean f4.ca> writes:

bobef wrote:
 Hmmm... this was like... one year and 30 DMD versions ago. And in D 1.0 it is
still not fixed...
 http://www.digitalmars.com/d/archives/digitalmars/D/34291.html

I'm not sure this is an easy one to fix.  In this case, the DLL contains 
the full D runtime and it expects to be initialized to operate 
properly--the Thread code is a part of the runtime.  The easiest fix 
would be to call Thread.thread_init() in the DLL before calling any 
other thread routines there.  Alternately, the DLL load routine would 
have to be altered to pass in a handle to the GC which should be used, 
and this would implicitly initialize the thread code, etc.  Personally, 
I think the current method of using D in DLLs kind of stinks anyway, so 
all of this may be worth investigating.


Sean