[PATCH] D81622: [Clang] Search computed sysroot for libc++ header paths

[Dan Albert via Phabricator via cfe-commits]

danalbert accepted this revision.
danalbert added a comment.

In D81622#2094409 <https://reviews.llvm.org/D81622#2094409>, @ldionne wrote:

> While this doesn't look wrong to me -- and the correctness of this depends entirely on where vendors decide to put their headers so it's hard for me to verify -- I'm wondering why not all toolchains use this mechanism. We seem to be adding an abstraction that's used only by some toolchains, but not all. I think it would be great to have a single canonical way of representing the system root.


Yeah, I agree that the inconsistencies between vendors in the driver are pretty odd. I think essentially the way we got into this situation is that it's not always clear what the right way to do things is, and if it's not even clear that there even is an existing mechanism we can easily end up creating our own per vendor :(

I'd be interested in seeing if there are things we can do to try to unify the behavior in this area. I suspect it could be difficult given that it necessarily involves all of the different vendors, but maybe we could come up with one recommended way of doing all this and any of us that don't fit that model can make the change at a pace convenient for us?

> Commenting out of curiosity, don't let this block you.

Thanks for making that clear :) We're going to go ahead and submit since this does fix a regression but I'm open to making changes here if (hopefully //when//) we find some cleanup that can be done.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D81622/new/

https://reviews.llvm.org/D81622