ldionne abandoned this revision.
ldionne added a comment.
Herald added a project: clang-tools-extra.
Revisiting this, I don't think we need to move forward with this. Let's just
carry our downstream patch until we don't need it anymore (which should be soon
ish), and leave upstream Clang as it i
teemperor added a comment.
From the LLDB side we can just change our test setup to explicitly add an
include the libc++ in our build tree (and have -nostdinc to avoid the libc++ in
the SDK). So it wouldn't be a problem for us. I can make a review for that once
this gets accepted.
Repository:
dexonsmith requested changes to this revision.
dexonsmith added a comment.
This revision now requires changes to proceed.
@rsmith , @teemperor , @ldionne , thoughts on the above?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D88984/new/
https://revi
dexonsmith added reviewers: rsmith, teemperor.
dexonsmith added a comment.
Hmm, that's right, to support user-provided libc++ headers it's simpler if we
prefer the toolchain / next-to-clang.
The reason we need to invert it is transitional. The status quo is that we have
no headers in the SDKs (
teemperor added a comment.
On macOS, LLDB's tests pass `-isysroot /path/to/SDK` to Clang but expect that
Clang always prefers the the libc++ headers that are in the Clang build
directory. So if we change this behaviour then we should have a way to keep the
old behaviour around.
Repository:
rsmith added a comment.
This seems like it would break the opposite situation: a user places a custom
clang and the corresponding libc++ somewhere (perhaps in `$HOME`, or perhaps
they run Clang from the build tree). Right now we use the custom clang and its
corresponding libc++, but with this c
ldionne updated this revision to Diff 296772.
ldionne added a comment.
Split the patch in two
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D88984/new/
https://reviews.llvm.org/D88984
Files:
clang-tools-extra/test/clang-tidy/infrastructure/Inputs
dexonsmith added a comment.
Note that I don't see any issues besides splitting the patch and dealing with
the clang-format linter problems.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D88984/new/
https://reviews.llvm.org/D88984
_
dexonsmith requested changes to this revision.
dexonsmith added a comment.
This revision now requires changes to proceed.
> Furthermore, instead of always adding both search paths, we find
> the first search path that exists and only add that one.
Huh, I'm surprised this wasn't already the behavi
ldionne created this revision.
ldionne added reviewers: arphaman, dexonsmith.
Herald added subscribers: cfe-commits, jkorous.
Herald added a project: clang.
ldionne requested review of this revision.
Currently, Clang looks for libc++ headers alongside the installation
directory of Clang, and it al
10 matches
Mail list logo