• akcom (6/6) Mar 22 2006 Just simply out of curiosity, why keep the back-end for dmd closed
• John Reimer (4/11) Mar 22 2006 You wouldn't find the dmd backend very useful for PowerPC unless you are...
• John Demme (11/18) Mar 22 2006 As JJR said, this won't help you, since the backend doesn't do PPC code.
• David Medlock (3/30) Mar 23 2006 Perhaps we could port it to : http://llvm.org/
• kris (6/10) Mar 23 2006 Good idea ~ that would quite possibly get D onto ARM and H8S (phones,

akcom <CppCoder gmail.com> writes:

Just simply out of curiosity, why keep the back-end for dmd closed 
source?  It would make life dev'ing on PowerPC linux so much nicer if I 
could just compile dmd myself instead of having to use an outdated 
version of gdc (no offense to the great job being done with gdc, it's 
just a bit dated at this point.

-Alex

John Reimer <terminal.node gmail.com> writes:

akcom wrote:
 Just simply out of curiosity, why keep the back-end for dmd closed 
 source?  It would make life dev'ing on PowerPC linux so much nicer if I 
 could just compile dmd myself instead of having to use an outdated 
 version of gdc (no offense to the great job being done with gdc, it's 
 just a bit dated at this point.
 
 -Alex


You wouldn't find the dmd backend very useful for PowerPC unless you are 
intent on using it as an x86 cross compiler.

-JJR

John Demme <me teqdruid.com> writes:

akcom wrote:

 Just simply out of curiosity, why keep the back-end for dmd closed
 source?  It would make life dev'ing on PowerPC linux so much nicer if I
 could just compile dmd myself instead of having to use an outdated
 version of gdc (no offense to the great job being done with gdc, it's
 just a bit dated at this point.
 
 -Alex


As JJR said, this won't help you, since the backend doesn't do PPC code.

Walter keeps it closed since he also uses it for dmc and dmc++.  His backend
contains some very powerful tools, such as the optimizer, which he has been
developing for a very long time.  I don't expect him (or ask of him) to
open it up.  It's great that the front-end is open, though, so we have GDC.

It would be nice, however, if a compiled version of the backend could be
supplied in a statis library so that one can compile the front-end and link
it with the backend.  This way, one can experiment with the front end.  I
don't know how much more this would reveal of the back-end, however.


~John Demme

David Medlock <noone nowhere.com> writes:

John Demme wrote:

 akcom wrote:
 
 
Just simply out of curiosity, why keep the back-end for dmd closed
source?  It would make life dev'ing on PowerPC linux so much nicer if I
could just compile dmd myself instead of having to use an outdated
version of gdc (no offense to the great job being done with gdc, it's
just a bit dated at this point.

-Alex

 
 
 
 As JJR said, this won't help you, since the backend doesn't do PPC code.
 
 Walter keeps it closed since he also uses it for dmc and dmc++.  His backend
 contains some very powerful tools, such as the optimizer, which he has been
 developing for a very long time.  I don't expect him (or ask of him) to
 open it up.  It's great that the front-end is open, though, so we have GDC.
 
 It would be nice, however, if a compiled version of the backend could be
 supplied in a statis library so that one can compile the front-end and link
 it with the backend.  This way, one can experiment with the front end.  I
 don't know how much more this would reveal of the back-end, however.
 
 
 ~John Demme

Perhaps we could port it to : http://llvm.org/

-DavidM

kris <foo bar.com> writes:

David Medlock wrote:
[snip]
 
 Perhaps we could port it to : http://llvm.org/
 
 -DavidM

Good idea ~ that would quite possibly get D onto ARM and H8S (phones, 
etc), since llvm can emit plain C code too ...

Doesn't GDC work with GCC 3.4?

- Kris