For OpenCL, the private address space qualifier is 0 in AST. Before this change, 0 address space qualifier
is always mapped to target address space 0. As now target private address space is specified by
alloca address space in data layout, address space qualifier 0 needs to be mapped to alloca addr space specified by the data layout.

This change has no impact on targets whose alloca addr space is 0.

With contributions from Matt Arsenault, Tony Tye and Wen-Heng (Jack) Chung

Differential Revision: https://reviews.llvm.org/D31404


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299965 91177308-0d34-0410-b5e6-96231b3b80d8

rdar://31501863


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299962 91177308-0d34-0410-b5e6-96231b3b80d8

We need to address cases (breaking libc++) such as

template <class _Up> static int __test(...);

template<typename _Tp>
auto v = __test<_Tp>(0);





git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299956 91177308-0d34-0410-b5e6-96231b3b80d8

At present, clang-format mangles Java containing logical right shift operators
('>>>=' or '>>>'), splitting them in two, resulting in invalid code:

 public class Minimal {
   public void func(String args) {
     int i = 42;
-    i >>>= 1;
+    i >> >= 1;
     return i;
   }
 }

This adds both forms of logical right shift to the FormatTokenLexer, so
clang-format won't attempt to split them and insert bogus whitespace.

https://reviews.llvm.org/D31652
Patch from Richard Bradfield <bradfier@fstab.me>!


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299952 91177308-0d34-0410-b5e6-96231b3b80d8

Differential Revision: https://reviews.llvm.org/D25321


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299951 91177308-0d34-0410-b5e6-96231b3b80d8

as identifiers in Objective-C++

This commit improves the 'expected identifier' errors that are presented when a
C++ keyword is used as an identifier in Objective-C++ by mentioning that this is
a C++ keyword in the diagnostic message. It also improves the error recovery:
the parser will now treat the C++ keywords as identifiers to prevent unrelated
parsing errors.

rdar://20626062

Differential Revision: https://reviews.llvm.org/D26503


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299950 91177308-0d34-0410-b5e6-96231b3b80d8

Reviewed by Richard Smith (D29877)!



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299930 91177308-0d34-0410-b5e6-96231b3b80d8

Summary:
This fixes a few outstanding bugs:
  * incorrect breaking of NSString literals containing double-width characters;
  * inconsistent formatting of ObjC dictionary literals containing NSString
    literals;
  * AlwaysBreakBeforeMultilineStrings ignoring implicitly-concatenated NSString
    literals.

Reviewers: djasper

Reviewed By: djasper

Subscribers: klimek, cfe-commits

Differential Revision: https://reviews.llvm.org/D31706

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299927 91177308-0d34-0410-b5e6-96231b3b80d8

This is a compiler part of https://reviews.llvm.org/D29586. Enable LSan on arm Linux.

Differential Revision: https://reviews.llvm.org/D31760


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299921 91177308-0d34-0410-b5e6-96231b3b80d8

Reviewers: djasper

Reviewed By: djasper

Subscribers: Eugene.Zelenko, klimek, cfe-commits

Differential Revision: https://reviews.llvm.org/D31408

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299919 91177308-0d34-0410-b5e6-96231b3b80d8

docs: Use the term "whole-program devirtualization" instead of "virtual function call optimization".

The former term is probably more familiar to users. Also add references to
the command line flags used to enable the features described in the doc.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299902 91177308-0d34-0410-b5e6-96231b3b80d8

This re-lands r299875.

I introduced a bug in Clang code responsible for replacing K&R, no
prototype declarations with a real function definition with a prototype.
The bug was here:

       // Collect any return attributes from the call.
  -    if (oldAttrs.hasAttributes(llvm::AttributeList::ReturnIndex))
  -      newAttrs.push_back(llvm::AttributeList::get(newFn->getContext(),
  -                                                  oldAttrs.getRetAttributes()));
  +    newAttrs.push_back(oldAttrs.getRetAttributes());

Previously getRetAttributes() carried AttributeList::ReturnIndex in its
AttributeList. Now that we return the AttributeSetNode* directly, it no
longer carries that index, and we call this overload with a single node:
  AttributeList::get(LLVMContext&, ArrayRef<AttributeSetNode*>)

That aborted with an assertion on x86_32 targets. I added an explicit
triple to the test and added CHECKs to help find issues like this in the
future sooner.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299899 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299889 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299877 91177308-0d34-0410-b5e6-96231b3b80d8

If the type of the captured variable is a pointer(s) to variably
modified type, this type was not processed correctly. Need to drill into
the type, find the innermost variably modified array type and convert it
to canonical parameter type.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299868 91177308-0d34-0410-b5e6-96231b3b80d8

Add the 'z3' subdirectory to the list of possible path suffixes for
libz3 header search. The z3 headers are installed in /usr/include/z3
on Gentoo.

Differential Revision: https://reviews.llvm.org/D31756

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299813 91177308-0d34-0410-b5e6-96231b3b80d8

Previously __cfi_check was created in LTO optimization pipeline, which
means LLD has no way of knowing about the existence of this symbol
without rescanning the LTO output object. As a result, LLD fails to
export __cfi_check, even when given --export-dynamic-symbol flag.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299806 91177308-0d34-0410-b5e6-96231b3b80d8

crtbegin is not really a proper windows support thing.  This was
duplicated when the toolchain was initially built.  If the injection of
crtbegin is needed, it can be done via the `/include` directive.
Furthermore, since `-fPIC` doesnt make sense on PE/COFF, crtbegin and
crtbeginS dont really need to be different.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299800 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299785 91177308-0d34-0410-b5e6-96231b3b80d8

It's used by MS headers in VS 2017 without including intrin.h, so we
can't implement it in the header anymore.

Differential Revision: https://reviews.llvm.org/D31736

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299782 91177308-0d34-0410-b5e6-96231b3b80d8

MSDN (https://msdn.microsoft.com/en-us/library/h5w10wxs.aspx) indicates
that `__declspec(naked)` is only permitted on x86 and ARM targets.
Testing with cl does confirm this behaviour.  Provide a warning for use
of `__declspec(naked)` on x64.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299774 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299760 91177308-0d34-0410-b5e6-96231b3b80d8

Differential Revision: https://reviews.llvm.org/D31237


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299759 91177308-0d34-0410-b5e6-96231b3b80d8

Summary: When using the C preprocessor with assembly files, either with a
capital `S` file extension, or with `-xassembler-with-cpp`, the Unicode escape
sequence `\u` is ignored. The `\u` pattern can be used for expanding a macro
argument that starts with `u`.

Author: Salman Arif <salman.arif@arm.com>

Reviewers: rengolin, olista01

Reviewed By: olista01

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D31765


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299754 91177308-0d34-0410-b5e6-96231b3b80d8

Change constant address space from 4 to 2 for the new address space mapping in Clang.

Differential Revision: https://reviews.llvm.org/D31771


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299691 91177308-0d34-0410-b5e6-96231b3b80d8

When fixing a Clang-Tidy bug in D31406,
reuse of FileID enabled the missing highlightRange function.
Assertion in highlightRange failed because the end-of-range column
number was 2 + the last column of a line on Windows.
This fix is required to enable D31406.

Differential Revision: https://reviews.llvm.org/D31713



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299681 91177308-0d34-0410-b5e6-96231b3b80d8

These two attributes specify the same info in a different way.
AMGPU BE only checks the latter as a target specific attribute
as opposed to language specific reqd_work_group_size.

This change produces amdgpu_flat_work_group_size out of
reqd_work_group_size if specified.

Differential Revision: https://reviews.llvm.org/D31728

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299678 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299671 91177308-0d34-0410-b5e6-96231b3b80d8

Hopefully fix crashes by unshadowing the variable.


Original commit message:

A big part of the clone detection code is functionality for filtering clones and
clone groups based on different criteria. So far this filtering process was
hardcoded into the CloneDetector class, which made it hard to understand and,
ultimately, to extend.

This patch splits the CloneDetector's logic into a sequence of reusable
constraints that are used for filtering clone groups. These constraints
can be turned on and off and reodreder at will, and new constraints are easy
to implement if necessary.

Unit tests are added for the new constraint interface.

This is a refactoring patch - no functional change intended.

Patch by Raphael Isemann!

Differential Revision: https://reviews.llvm.org/D23418


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299653 91177308-0d34-0410-b5e6-96231b3b80d8

rdar://20441985


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299650 91177308-0d34-0410-b5e6-96231b3b80d8

object types should be preferred over conversions to other object pointers

This change ensures that Clang will select the correct overload for the
following code sample:

  void overload(Base *b);
  void overload(Derived *d);
  void test(Base<Base *> b) {
    overload(b); // Select overload(Base *), not overload(Derived *)
  }

rdar://20124827

Differential Revision: https://reviews.llvm.org/D31597


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299648 91177308-0d34-0410-b5e6-96231b3b80d8

lambda capture used by the created block

The commit r288866 introduced guaranteed copy elision to C++ 17. This
unfortunately broke the lambda to block conversion in C++17 (the compiler
crashes when performing IRGen). This commit fixes the conversion by avoiding
copy elision for the capture that captures the lambda that's used in the block
created by the lambda to block conversion process.

rdar://31385153

Differential Revision: https://reviews.llvm.org/D31669


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299646 91177308-0d34-0410-b5e6-96231b3b80d8

Attempt to satisfy llvm-clang-x86_64-expensive-checks-win by targeting
x86_64-apple-darwin10 for Sema/vector-ops.c. The underlying failure is
due to datatype differences between platforms.



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299643 91177308-0d34-0410-b5e6-96231b3b80d8

This improves some error messages which would otherwise refer to
ext_vector_type types in contexts where there are no such types.

Factored out from D25866 at reviewer's request.

Reviewers: bruno

Differential Revision: https://reviews.llvm.org/D31667



git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299641 91177308-0d34-0410-b5e6-96231b3b80d8

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299639 91177308-0d34-0410-b5e6-96231b3b80d8

- Replace documented return values (true/false) with what's actually
  returned
- Doxygenify the comment
- Reflow said comment to 80 cols

Not overly familiar with Doxygen, so nits are welcome. :)


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299603 91177308-0d34-0410-b5e6-96231b3b80d8

Two simplifications:
- We check `!Previous.empty()` above and only use `Previous` in const
  contexts after that check, so the `!Previous.empty()` check seems
  redundant.
- The null check looks pointless, as well: AFAICT, `LookupResults`
  should never contain null entries, and `OldDecl` should always be
  non-null if `Redeclaration` is true.


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299601 91177308-0d34-0410-b5e6-96231b3b80d8

The commit yesterday (r299473) to add the `-print-resource-dir`
option was supposed to emit a newline after the resource dir.

Differential Revision: https://reviews.llvm.org/D31447

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299597 91177308-0d34-0410-b5e6-96231b3b80d8

Move the test format into a standalone .py file and add it to the site
module search path. This allows us to run the test on Windows, and it
makes it compatible with the multiprocessing.Pool lit test execution
strategy.

I think this test was only passing everywhere else because
multiprocessing uses 'fork' to spawn workers, so the test format never
needs to be pickled.

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299577 91177308-0d34-0410-b5e6-96231b3b80d8

clang-format <<END
auto c1 = u8'a';
auto c2 = u'a';
END

Before:
  auto c1 = u8 'a';
  auto c2 = u'a';

Now:
  auto c1 = u8'a';
  auto c2 = u'a';

Patch from Denis Gladkikh <llvm@denis.gladkikh.email>!


git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@299574 91177308-0d34-0410-b5e6-96231b3b80d8