[llvm-commits] [PATCHES] fix duplication and FIXMEs in JIT code

[nobled]

ping?

On Sun, Apr 24, 2011 at 6:20 PM, nobled <nobled at dreamwidth.org> wrote:
> Whoops, forgot to join the mailing list, so this got stuck in the
> moderation queue...
>
> ---------- Forwarded message ----------
> From: nobled <nobled at dreamwidth.org>
> Date: Sun, Apr 24, 2011 at 11:05 AM
> Subject: [PATCHES] fix duplication and FIXMEs in JIT code
> To: llvm-commits at cs.uiuc.edu
>
>
> (Just so you know, I do have commit access.)
>
> These patches address this comment from the MCJIT code:
>
>  // FIXME: This should be lifted out of here, it isn't something which should
>  // be part of the JIT policy, rather the burden for this selection should be
>  // pushed to clients.
> ====
> [PATCH 1/3] ExecutionEngine: fix JIT/MCJIT selectTarget() duplication
>
> (Merges the near-identical files lib/EE/JIT/TargetSelect.cpp and
> lib/EE/MCJIT/TargetSelect.cpp into lib/EE/TargetSelect.cpp.)
>
> This prepares for making JITCtor/MCJITCtor take a
> TargetMachine* directly from clients like EngineBuilder.
> ====
> [PATCH 2/3] ExecutionEngine: push TargetMachine creation into clients
>
> In particular, into EngineBuilder. This should only impact
> the private API between the EE and EB classes, not external
> clients, since JITCtor and MCJITCtor are both protected members.
> ====
> [PATCH 3/3] ExecutionEngine: move createJIT definition
>
> As an ExecutionEngine class function, its definition
> really belongs in ExecutionEngine.cpp instead of in JIT.cpp.
> ====
>
> As for this comment:
>
>   // Try to register the program as a source of symbols to resolve against.
>   //
>   // FIXME: Don't do this here.
>   sys::DynamicLibrary::LoadLibraryPermanently(0, NULL);
>
> where in the stack *should* that go, ideally?
>