ownership qualifier cast which won't work.
// rdar://10244607

llvm-svn: 143258

llvm-svn: 143257

Currently sorting by raw source location does work as intended but who knows
what may change in the future..

llvm-svn: 143256

if we are not interested in the decl.

llvm-svn: 143255

previous address only by the number of bytes consumed by the disassembly:

(lldb) x/4i 0x0000000100000ea9
0x100000ea9:  66 c7 45 fa 10 00     movw   $16, -6(%rbp)
0x100000eaf:  c7 45 f4 20 00 00 00  movl   $32, -12(%rbp)
0x100000eb6:  e8 47 00 00 00        callq  0x0000000100000f02       ; void f<nullptr_t>(nullptr_t)
0x100000ebb:  8b 45 fc              movl   -4(%rbp), %eax
(lldb) 
0x100000ebe:  48 83 c4 10  addq   $16, %rsp
0x100000ec2:  5d           popq   %rbp
0x100000ec3:  c3           ret    
0x100000ec4:  90           nop    
(lldb) 
0x100000ec5:  90  nop    
0x100000ec6:  90  nop    
0x100000ec7:  90  nop    
0x100000ec8:  90  nop    
(lldb) 
0x100000ec9:  90  nop    
0x100000eca:  90  nop    
0x100000ecb:  90  nop    
0x100000ecc:  90  nop    
(lldb) 
0x100000ecd:  90  nop    
0x100000ece:  90  nop    
0x100000ecf:  90  nop    
0x100000ed0:  55  pushq  %rbp

llvm-svn: 143254

allow it to complete types on behalf of any AST context
(including the "scratch" AST context associated with
the target), I scrapped its role as intermediary between
the Clang parser and ClangExpressionDeclMap, and instead
made ClangExpressionDeclMap inherit from ClangASTSource.

After this, I will migrate the functions that complete
types and perform namespace lookups from
ClangExpressionDeclMap to ClangASTSource.  Ultimately
ClangExpressionDeclMap's only responsiblity will be to
look up variables and ensure that they are materialized
and dematerialized correctly.

llvm-svn: 143253

commands to print the binary representaion of an integer.

llvm-svn: 143252

order. Also hooked up the new formats for instruction, hex float and address
to the new formats.

llvm-svn: 143251

binary operator isn't an rvalue if it's an assignment operator.

llvm-svn: 143250

when setting breakpoints, but only if no module is specified.  The Darwin 
platform uses this to not set breakpoints in dyld.

llvm-svn: 143249

define their own "nullptr" class in C++98 and build with -Wall.

llvm-svn: 143248

llvm-svn: 143247

llvm-svn: 143246

committee hasn't set a value for __STDC_VERSION__ yet, so this is a
placeholder. But at least it's > 199901L.

llvm-svn: 143245

the ensuing mayhem.

llvm-svn: 143244

will cause c-index-test to return as failed if a compiler error occurred.

llvm-svn: 143243

llvm-svn: 143242

the previous ASTDeserializationListener.

llvm-svn: 143241

in which case the ownership is redundant. Thanks to John for the suggestion.

llvm-svn: 143240

llvm-svn: 143239

Introduce a FILE_SORTED_DECLS [de]serialization record that contains
a file sorted array of file-level DeclIDs in a PCH/Module.
The rationale is to allow "targeted" deserialization of decls inside
a range of a source file.

Cocoa PCH increased by 0.8%
Difference of creation time for Cocoa PCH is below the noise level.

llvm-svn: 143238

llvm-svn: 143237

When '~imm' is encodable as a t2_so_imm but plain 'imm' is not. For example,
  mov r2, #-3
becomes
  mvn r2, #2

rdar://10349224

llvm-svn: 143235

constexpr function arguments outside of their function (passing or returning
them by reference) does not work correctly yet.

Calling constexpr function templates does not work yet, since the bodies are not
instantiated until the end of the translation unit.

llvm-svn: 143234

For example,

On ARM, "mov r3, #-3" is an alias for "mvn r3, #2", so we want to use a
matcher pattern that handles the bitwise negation when mapping to t2MVNi.

llvm-svn: 143233

llvm-svn: 143231

things like:

(lldb) x/32xb 0x1000

"x" is an alias to "memory read", so this will actually turn into:

(lldb) memory read --gdb-format=32xb 0x1000

This applies to all commands, so the GDB formats will work with "register read",
"frame variable", "target variable" and others. All commands that can accept
formats, counts and sizes have been modified to support the "--gdb-format"
option.

llvm-svn: 143230

llvm-svn: 143228

In case of template specialization, do not try to delay emitting debug info for concrete type in -flimit-debug-info mode. This fixes some of the failures from bs15503.exp tests in gdb testsuite.

llvm-svn: 143227

llvm-svn: 143225

llvm-svn: 143224

llvm-svn: 143223

the common case (-O0, no always_inline) fast.

llvm-svn: 143222

llvm-svn: 143220

an rvalue retainable object type with life-time qualifier has no
effect and wil be diagnosed as error. // rdar://10244607

llvm-svn: 143219

llvm-svn: 143218

llvm-svn: 143217

3.0. Fixses PR11252.

llvm-svn: 143216

[analyzer] ObjC message sends to nil receivers that return structs are now okay (compiler zeroes out the data).  Fixes <rdar://problem/9151319>.

llvm-svn: 143215

Spotted by my super-optimizer in 186.crafty and 450.soplex.  We really
need a proper infrastructure for handling generalizations of this kind
of thing (which occur a lot), however this case is so simple that I decided
to go ahead and implement it directly.

llvm-svn: 143214