yrouban wrote:

> Thanks you, @yrouban !
> 
> > By the way, does this change fix any issue or solve any problem other than 
> > the irrational design you explained?
> 
> I'm working on an RFC to rework the internals of the options interfaces in 
> order to make them thread-separated so we can reduce the global state surface 
> and _hopefully_ **eventually** let tooling (internal and external) run 
> concurrent LLVM-based tools without having to spawn processes. Of course, 
> there's no guarantee that the RFC gains traction, but as part of my 
> prototyping I decided to push out some of the more stylistic changes. Most of 
> the smaller changes (like this) are not strictly required in the larger 
> picture, but if people think it improves the quality of the interfaces I 
> think it's worth considering. 🚀

I proposed this idea in 2018 (https://reviews.llvm.org/D53424, 
https://lists.llvm.org/pipermail/llvm-dev/2018-October/127039.html).

https://github.com/llvm/llvm-project/pull/184581
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to