On Fri, Sep 10, 2010 at 03:09:02PM +0200, Manuel López-Ibáñez wrote: > On 10 September 2010 14:40, Richard Kenner <ken...@vlsi1.ultra.nyu.edu> wrote: > > > > But if this were done, then it would be trivial to have proprietary > > front ends, back ends, and optimizers. So RMS never allowed any such > > thing nor any scheme that resulted in having any file that could be > > used for such a purpose. > > As far as I know, you can currently plug GCC FEs to a proprietary LLVM > middle-end using llvm-gcc. It is not a possibility but a reality (read > the proceedings of the LLVM meetings). Just no one cares, because (a) > the result is not publicly distributed (only internally or for > research purposes) and (b) there is zero interest on > contributing/integrating any code back to GCC from both sides.
Huh? The dragon-egg gcc plugin to access llvm has been pubicly available since llvm 2.7... http://dragonegg.llvm.org/ http://dragonegg.llvm.org/#gettingrelease As to the second point, I believe that is entirely untrue (at least as far as patches required to build dragon-egg under FSF gcc is concerned). In the thread starting at http://gcc.gnu.org/ml/gcc/2010-04/msg00171.html, it is clear that Duncan Sands is interested in some level of coordination of dragon-egg development with FSF gcc. Jack > > There are two issues, whether is possible technically to plug clang to > GCC and whether is possible to create proprietary FEs to GCC. From > Richard comments, the answer to the former is probably yes with > perhaps less effort than Dragonegg. But that is not so important, > because the answer to the latter is also yes right now by building a > modified GCC (people have mentioned this in the lists). > > Cheers, > > Manuel.
