Make this call an exception in ExprEngine::invalidateArguments:
'int pthread_setspecific(ptheread_key k, const void *)' stores
a value into thread local storage. The value can later be retrieved
with 'void *ptheread_getspecific(pthread_key)'. So even thought the
parameter is 'const void *', the region escapes through the
call.

(Here we just blacklist the call in the ExprEngine's default
logic. Another option would be to add a checker which evaluates
the call and triggers the call to invalidate regions.)

Teach the Malloc Checker, which treats all system calls as safe about
the API.

llvm-svn: 151220

llvm-svn: 151217

llvm-svn: 151216

llvm-svn: 151203

 - We should not evaluate strdup in the Malloc Checker, it's the job of
CString checker, so just update the RefState to reflect allocated
memory.

- Refactor to reduce LOC: remove some wrapper auxiliary functions, make
all functions return the state and add the transition in one place
(instead of in each auxiliary function).

llvm-svn: 151188

by -target and similar options. As discussed in PR 12026, the change
broke support for target-prefixed tools, i.e. calling x86_64--linux-ld
when compiling for x86_64--linux. Improve the test cases added
originally in r149083 to not require execution, just executable files.
Document the hack with appropiate FIXME comments.

llvm-svn: 151185

abi.

llvm-svn: 151176

are likely many other OPT_xxxx_EQ options that could/should be added here.
rdar://10704648

llvm-svn: 151174

llvm-svn: 151173

llvm-svn: 151172

llvm-svn: 151171

lambda closure type's function pointer conversion over user-defined
conversion via a lambda closure type's block pointer conversion,
always. This is a preference for more-standard code (since blocks
are an extension)  and a nod to efficiency, since function pointers
don't require any memory management. Fixes PR12063.

llvm-svn: 151170

Fix parsing and processing initializer lists in return statements and as direct member initializers.

llvm-svn: 151155

This adds the -Wformat-non-standard flag (off by default,
enabled by -pedantic), which warns about non-standard
things in format strings (such as the 'q' length modifier,
the 'S' conversion specifier, etc.)

llvm-svn: 151154

llvm-svn: 151152

llvm-svn: 151151

llvm-svn: 151150

I think there's a deeper problem here in the way TransformCXXConstructExpr works, but I won't tackle it now.

llvm-svn: 151146

In -fdelayed-template-parsing mode, reenter every scope when late parsing a templated function; (Not just the template parameter scope as previously). Also enter the scope stack in the correct order.

Otherwise this breaks some invariant during name lookup especially when dealing with shadowed declaration

Fix PR11931.

llvm-svn: 151140

the diagnostic for using a parenthesized direct-initializer in a condition.

llvm-svn: 151137

llvm-svn: 151133

block pointer that returns a block literal which captures (by copy)
the lambda closure itself. Some aspects of the block literal are left
unspecified, namely the capture variable (which doesn't actually
exist) and the body (which will be filled in by IRgen because it can't
be written as an AST).

Because we're switching to this model, this patch also eliminates
tracking the copy-initialization expression for the block capture of
the conversion function, since that information is now embedded in the
synthesized block literal. -1 side tables FTW.

llvm-svn: 151131

Improve diagnostics a bit for bad member initializers, and fix an obscure bug involving packs.  Fixes PR12049.

llvm-svn: 151130

llvm-svn: 151124

llvm-svn: 151122

, when we return a symbol reachable to the malloced one via pointer
arithmetic.

llvm-svn: 151121

llvm-svn: 151120

function call (or a comma expression with a function call on its right-hand
side), possibly parenthesized, then the return type is not required to be
complete and a temporary is not bound. Other subexpressions inside a decltype
expression do not get this treatment.

This is implemented by deferring the relevant checks for all calls immediately
within a decltype expression, then, when the expression is fully-parsed,
checking the relevant constraints and stripping off any top-level temporary
binding.

Deferring the completion of the return type exposed a bug in overload
resolution where completion of the argument types was not attempted, which
is also fixed by this change.

llvm-svn: 151117

llvm-svn: 151112

crash.  This can speedup the process of generating a delta reduced test case.
rdar://10905465

llvm-svn: 151109

any implementation in tu was not being translated.

llvm-svn: 151106

llvm-svn: 151103

that 'this' can be used in the brace-or-equal-initializer of a
non-static data member, and C++11 [expr.prim.lambda]p9, which says
that lambda expressions not in block scope can have no captures, side
fully with C++11 [expr.prim.general]p4 by allowing 'this' to be
captured within these initializers. This seems to be the intent of
non-static data member initializers.

llvm-svn: 151101

llvm-svn: 151100

rdar://problem/10904479

llvm-svn: 151089

expression after we've finished the function body of the corresponding
function call operator. Otherwise, ActOnFinishFunctionBody() will see
the (unfinished) evaluation context of the lambda expression
itself. Fixes PR12031.

llvm-svn: 151082

is non-null when diagnosing a broken attempt to write a for-range-statement.

llvm-svn: 151081

When calling a non variadic format function(vprintf, vscanf, NSLogv, …), warn if the format string argument is a parameter that is not itself declared as a format string with compatible format.

llvm-svn: 151080

arguments. There are two aspects to this:

  - Make sure that when marking the declarations referenced in a
  default argument, we don't try to mark local variables, both because
  it's a waste of time and because the semantics are wrong: we're not
  in a place where we could capture these variables again even if it
  did make sense.
  - When a lambda expression occurs in a default argument of a
  function template, make sure that the corresponding closure type is
  considered dependent, so that it will get properly instantiated. The
  second bit is a bit of a hack; to fix it properly, we may have to
  rearchitect our handling of default arguments, parsing them only
  after creating the function definition. However, I'd like to
  separate that work from the lambdas work.

llvm-svn: 151076

that's installed.

llvm-svn: 151058