[LLVMdev] (possible) bytecode format change

Anton Korobeynikov asl at math.spbu.ru
Wed Apr 18 23:50:07 PDT 2007


> Is that change absolutely necessary?
Unfortunately, yes. We're having at least two PRs opened for aliases
including libstdc++ compilation in shared mode for x86/linux.

> I've just spent 2 days compiling benchmarks. So, now I'd need to
> ditch all that and start from scratch...
No. Bytecode will be breaking only if it have module-wide assembler. I
don't think you have too many such bytecodes.

> A question:
> Does that mean that function name overloading is now again allowed?
> [This breaks much more than 2 days of my work...]
No. Aliases are just "hacks" by its nature. The only reason of
introducing them to LLVM is GCC support. Alias "foo => bar" has meaning
"we're using name foo for calling this function in this module, but
linker well use name bar instead", so this is just quick way to rename
function at module level (e.g. via macros or via some configuration). I
don't see any another reason aliases are useful :)

With best regards, Anton Korobeynikov.

Faculty of Mathematics & Mechanics, Saint Petersburg State University.

More information about the llvm-dev mailing list