LDC and D2 plans (or lack of them)

15 views
Skip to first unread message

Mike Farnsworth

unread,
Nov 17, 2009, 2:03:30 PM11/17/09
to LDC - the LLVM D compiler
I saw on the D NG that Tomas said he (personally) has no plans to
support D2, due to issues with the front end. As D2 gets sorted out
and finalized (presumably by March-ish?) would it help to plan the
creation of a new frontend that could be used by projects such as ldc
and gdc?

I'm willing to contribute to the effort (I don't have gobs of free
time, but I could certainly spend some on it). It would give you guys
the freedom to be free of the jolting dmd frontend changes and current
limitations, and there are other benefits. Honestly, due to ldc's
potential for supporting many/more platforms and to generate code
well, it would likely become the de facto leading D compiler at that
point.

The question is, is it worth it to any of you to go to the effort? I
think it is, but my opinion is somewhat less relevant at this stage.

-Mike

Christian Kamm

unread,
Nov 18, 2009, 2:44:09 PM11/18/09
to ldc...@googlegroups.com
On Tuesday 17 November 2009 20:03 Mike Farnsworth wrote:
> I saw on the D NG that Tomas said he (personally) has no plans to
> support D2, due to issues with the front end.

While the frontend certainly is no pleasure to work with, I feel the main
reasons really are a lack of time and interest. After all we've usually
managed to adapt the D1 frontend where necessary.

> As D2 gets sorted out
> and finalized (presumably by March-ish?) would it help to plan the
> creation of a new frontend that could be used by projects such as ldc
> and gdc?

Creating a new frontend is a major undertaking, significantly larger in scope
than just adapting Walter's code. Previous attempts didn't get very far.
I think the most advanced D frontend at the moment is dil, found at
http://github.com/azizk/dil/.

Regards,
Christian

signature.asc
Reply all
Reply to author
Forward
0 new messages