[LLVMdev] Removing AvailableExternal values in GlobalDCE (was Re: RFC: ThinLTO Impementation Plan)

[Duncan P. N. Exon Smith]

> On 2015 Jun 8, at 13:35, Eric Christopher <echristo at gmail.com> wrote:
> 
> 
> 
> On Mon, Jun 8, 2015 at 1:33 PM Reid Kleckner <rnk at google.com> wrote:
> On Mon, Jun 8, 2015 at 1:08 PM, Eric Christopher <echristo at gmail.com> wrote:
> I'd rather not have it be an llvm option at all. Just construct a different set of passes...
> 
> This would also solve the problem of needing multiple sets of options to be passed to the builder. It'd be a bit of a change (i.e. having clang do the pass setup), but I think it'd be worth it to have clang be able to explicitly say which passes it wants.
> 
> It's also a major change and would probably need to be discussed more widely so for now I guess this is sorta ok.
> 
> I think for now the bool on PassManagerBuilder is a good way to make progress. Eventually, we may want to customize the LTO compilation phase optimization more fully (use lower inlining threshold, defer vectorization to link time), but we can cross that bridge when we get there.
> 
> Precisely. In case it wasn't clear I was saying this.

This SGTM too.