On Wednesday 14 May 2008 14:21:11 Ludovic Brenta wrote: > Wesley J. Landaker writes: > > Long term, upstream is (supposedly?) moving away from gcc dependence > > and going to be using LLVM. That will be a great day if/when it ever > > happens. ;) > > I follow LLVM development from a distance. Duncan Sands is busy > porting the GNAT front-end to it. If the compilation problems you > have are in the front-end, moving the back-end to LLVM is not going to > help. Still, we live interesting times indeed :)
Actually, once I work through a few quirks, the front-end doesn't seem to be the problem; I can get it to completely compile and build, but then I get horrible memory alignment failures when trying to run the resulting executables. Right now, to build GHDL you have to graft it into a specific version of a gcc tree, and it uses gcc's build system and backend to generate code. My hope is that if GHDL does move to LLVM, it will not be as hard to get compiled, won't depend on an exact gcc source tree, and the actual VHDL runtime will be a lot more platform independent. But anyway, that's a *LONG* term change, so I'm not holding my breath. ;) > > So anyway, worst case, perhaps I will just revert to gnat-4.1 and > > whittle down the supported architectures for a while until upstream > > has a better solution. > > I suppose this is an acceptable backup solution, especially > considering that upstream has not yet moved to gnat-4.2 either :) Well it's sad, but I'm going to be uploading a package reverting to use gnat-4.1 shortly. That will at least be a stop-gap until I can get it working with gnat-4.3, which I can see is probably not going to happen this month. -- Wesley J. Landaker <[EMAIL PROTECTED]> <xmpp:[EMAIL PROTECTED]> OpenPGP FP: 4135 2A3B 4726 ACC5 9094 0097 F0A9 8A4C 4CD6 E3D2
signature.asc
Description: This is a digitally signed message part.
