what it does.

llvm-svn: 83740

Remove impdef of SRW. This fixes PR4779

llvm-svn: 83739

llvm-svn: 83738

llvm-svn: 83737

llvm-svn: 83736

llvm-svn: 83735

llvm-svn: 83734

llvm-svn: 83733

llvm-svn: 83732

variable inside the block.

llvm-svn: 83729

llvm-svn: 83728

llvm-svn: 83724

llvm-svn: 83722

llvm-svn: 83721

clang static analyser.  Decrease duplication in the text.

llvm-svn: 83720

llvm-svn: 83719

llvm-svn: 83717

llvm-svn: 83716

llvm-svn: 83715

Implement the core checking for compatible exception specifications in assignment and initialization.

The exception specification of the assignee must be the same or a subset of the target. In addition, exception specifications on arguments and return types must be equivalent, but this is not implemented yet.
This currently produces two diagnostics for every invalid assignment/initialization, due to the diagnostic produced outside PerformImplicitConversion, e.g. in CheckSingleInitializer. I don't know how to suppress this; in any case I think it is the wrong place for a diagnostic, since there are other diagnostics produced inside the function. So I'm leaving it as it is for the moment.

llvm-svn: 83710

llvm-svn: 83707

DemoteRegToStack.  This makes it more efficient (because it isn't
creating a ton of load/stores that are eventually removed by a later
mem2reg), and more slightly more effective (because those load/stores
don't get in the way of threading).

llvm-svn: 83706

works on unstructured CFGs.  This implements PR217, our oldest open PR.

llvm-svn: 83705

a Value* to a WeakVH was constructing a temporary WeakVH
(due to the implicit assignment operator).  This avoids
that cost.

llvm-svn: 83704

llvm-svn: 83703

Use, to complement the version that takes a use_iterator.

llvm-svn: 83702

llvm-svn: 83701

Split the various ambiguous result enumerators into their own enum.  Tests
for most of C++ [namespace.qual].

llvm-svn: 83700

into MachineInstrs. This is mostly just moving the code from
ScheduleDAGSDNodesEmit.cpp into a new class. This decouples MachineInstr
emitting from scheduling.

llvm-svn: 83699

since it won't do any folding. This will help avoid some inconvenient
casting.

llvm-svn: 83698

llvm-svn: 83697

MachineInstr::isInvariantLoad instead, which has the benefit of being
more complete.

llvm-svn: 83696

llvm-svn: 83695

llvm-svn: 83694

llvm-svn: 83693

Add CheckCallReturnType and start using it for regular call expressions. This will improve error messages. For 

struct B;

B f();

void g() {
f();
}

We now get

t.cpp:6:3: error: calling 'f' with incomplete return type 'struct B'
  f();
  ^~~
t.cpp:3:3: note: 'f' declared here
B f();
  ^
t.cpp:1:8: note: forward declaration of 'struct B'
struct B;
       ^

llvm-svn: 83692

it isn't needed in the ScheduleDAGSDNodes schedulers.

llvm-svn: 83691

fewer remat.

llvm-svn: 83690

when loading from an invariant memory location.

llvm-svn: 83688

is trivially rematerializable and integrate it into
TargetInstrInfo::isTriviallyReMaterializable. This way, all places that
need to know whether an instruction is rematerializable will get the
same answer.

This enables the useful parts of the aggressive-remat option by
default -- using AliasAnalysis to determine whether a memory location
is invariant, and removes the questionable parts -- rematting operations
with virtual register inputs that may not be live everywhere.

llvm-svn: 83687