llvm-svn: 39595

llvm-svn: 39594

llvm-svn: 39593

llvm-svn: 39592

llvm-svn: 39591

llvm-svn: 39590

Submitted by:
Reviewed by:
Touch up a couple comments (one was incorrect/out-of-date).

llvm-svn: 39589

llvm-svn: 39588

check canonical types in a few places.  Tighten up VerifyConstantArrayType
to diagnose more errors, now that we can evaluate i-c-e's.  Add some fixmes
about poor diagnostics.

We now correctly typecheck this example:

void s(void) {
  typedef int a[(int) +1.0];
  static a b;    // invalid, static VLA

  static int c[(int) +1.0];  // invalid, static VLA
}

void t(void) {
  typedef int a[(int)1.0];
  static a b;    // valid, constant size
}

void u() {
  static int X[-1];
  static int Y[0];
}

producing:

static-vla.c:3:12: error: variable length array declared outside of any function
  static a b;    // invalid, static VLA
           ^
static-vla.c:5:14: error: variable length array declared outside of any function
  static int c[(int) +1.0];  // invalid, static VLA
             ^ ~~~~~~~~~~
static-vla.c:15:14: error: array size is negative
  static int X[-1];
             ^ ~~
static-vla.c:16:14: warning: zero size arrays are an extension
  static int Y[0];
             ^ ~

llvm-svn: 39587

on complete types.

llvm-svn: 39586

  1. Compute and return the value of the i-c-e if the expression is one.
  2. Use this computation to correctly track whether subexprs are being
     evaluated, and use this to guide diagnostics appropriately.

This allows us to correctly handle all the cases in:

void bar() {
  int foo();
  switch (1) {
  case 1 ? 0 : foo():    // bad
  case 0 ? 0 : foo():    // bad

  case 0 ? 1/0 : 14 :  // ok
  case 1 ? 1/0 : 14 :  // bad
    ;
  }

  switch (1) {
  case 1 ? 2: (2, 3): // ok
  case 0 ? 2: (2, 3): // invalid comma.
    ;
  }
}

This code has numerous todo items.  Specifically, we need to:
  1. Pass in target info, so we know the size of the integers we are producing.
  2. Model type sizes and alignments correctly, so we can eval sizeof/alignof
  3. Handle promotions (need to talk to steve about this).
  4. Return an enum that can be used to better diagnose problems with i-c-e's.
     instead of just saying "this isn't valid" we should be able to say why.
  5. Various other miscellanea, like handling enums and character literals
     properly.

llvm-svn: 39585

isConstantExpr into just isIntegerConstantExpr in preparation for other
changes.

llvm-svn: 39584

llvm-svn: 39583

llvm-svn: 39582

Submitted by:
Reviewed by:
Fixed an "impurity" in the typechecking code (to more closely model the
C99 spec)...

- Added function Sema::DefaultFunctionArrayConversion(). This is called
by UsualUnaryConversions(). It is also called from several contexts that
don't call for integer promotions (logical negation for expressions, and
while/if/do/for conditions in statements). It is also used in
UsualAssignmentConversions (which is cleaner than the check I was using
before).

- Changed the name from UsualUnaryConversion->UsualUnaryConversions.

- Added comments to the header.

Since several contexts don't call for integer
promotions,

llvm-svn: 39581

Submitted by:
Reviewed by:
Apply UsualUnaryConversion() to statement conditions that expect scalars.
UsualUnaryConversion() converts function/arrays to pointers.

This fixes the following...

int func() {
  int A[10];
  while (A) {
  }
  if (A) ;
  for (; A; ) ;
}

llvm-svn: 39580

llvm-svn: 39579

llvm-svn: 39578

'char', which varies based on the target.

Instead of spreading target knowledge throughout the compiler, bifurcate char
into Char_S and Char_U, and have ASTContext create the right one based on the
target, when it starts up.

llvm-svn: 39577

ints to long long etc.  For int to longlong, we now get:

        %tmp = load i64* %F             ; <i64> [#uses=1]
        %tmp1 = load i32* %D            ; <i32> [#uses=1]
        %promote = sext i32 %tmp1 to i64                ; <i64> [#uses=1]
        %tmp2 = add i64 %tmp, %promote          ; <i64> [#uses=0]

llvm-svn: 39576

llvm-svn: 39575

ds.c:11:16: error: case label does not reduce to an integer constant
  case 0 ? 0 : foo():
       ~~~~~~~~^~~~~

llvm-svn: 39574

- Small error fix.

llvm-svn: 39573

Reviewed by: Chris Lattner

- Add a reference decl type.

llvm-svn: 39572

Reviewed by: Chris Lattner

- Diagnostic messages for invalid reference usage.

llvm-svn: 39571

Reviewed by: Chris Lattner

- Method declaration.

llvm-svn: 39570

Reviewed by: Chris Lattner

- Disallow references to references, pointers to references, or arrays of
  references.

llvm-svn: 39569

Reviewed by: Chris Lattner

- If the LHS and/or RHS is a reference, then see if they're compatible.
  If so, the type is that of the LHS.

llvm-svn: 39568

Reviewed by: Chris Lattner

- Update the parsing of references. We allow "restrict" but not "const"
  or "volatile".

llvm-svn: 39567

Reviewed by: Chris Lattner

- Added a method to determine if two types, where at least one is a
  reference, are compatible. That is you can assign the RHS to the LHS.

llvm-svn: 39566

This allows us to compile:

  register short X;
  {
    int Y;
    return 1+X+Y;

into:

        %tmp = load i16* %X             ; <i16> [#uses=1]
        %promote = sext i16 %tmp to i32         ; <i32> [#uses=1]
        %tmp1 = add i32 1, %promote             ; <i32> [#uses=1]
        %tmp2 = load i32* %Y            ; <i32> [#uses=1]
        %tmp3 = add i32 %tmp1, %tmp2            ; <i32> [#uses=1]
        ret i32 %tmp3

Look at the amazing sign extension, so much happier than an assertion failure. :)

llvm-svn: 39565

This allows us to compile:

int func() {
  int A[10];
  if (!A) {

to:

define i32 @func() {
entry:
        %A = alloca [10 x i32]          ; <[10 x i32]*> [#uses=1]
        %arraydecay = getelementptr [10 x i32]* %A, i32 0, i32 0                ; <i32*> [#uses=1]
        %tobool = icmp ne i32* %arraydecay, null                ; <i1> [#uses=1]
        %lnot = xor i1 %tobool, true            ; <i1> [#uses=1]
        br i1 %lnot, label %ifthen, label %ifend

-Chris

llvm-svn: 39564

llvm-svn: 39563

typedef information associated with the input.

llvm-svn: 39562

compile:

void foo() {
  int A[10];
  int *P;

into:

entry:
        %A = alloca [10 x i32]          ; <[10 x i32]*> [#uses=0]
        %P = alloca i32*                ; <i32**> [#uses=0]

llvm-svn: 39561

For:

  register short X;
  if (!X) {

We now produce:

        %tmp = load i16* %X             ; <i16> [#uses=1]
        %tobool = icmp ne i16 %tmp, 0           ; <i1> [#uses=1]
        %lnot = xor i1 %tobool, true            ; <i1> [#uses=1]
        br i1 %lnot, label %ifthen, label %ifend

instead of:

        %tmp = load i16* %X             ; <i16> [#uses=1]
        %tobool = icmp ne i16 %tmp, 0           ; <i1> [#uses=1]
        %lnot = xor i1 %tobool, true            ; <i1> [#uses=1]
        %lnot.ext = zext i1 %lnot to i32                ; <i32> [#uses=1]
        %tobool1 = icmp ne i32 %lnot.ext, 0             ; <i1> [#uses=1]
        br i1 %tobool1, label %ifthen, label %ifend

llvm-svn: 39560

be shared.

Implement infrastructure for unary operator emission.

Implement basic logical not support.  We now compile:

  register short X;

  if (!X) {

into:

        %tmp = load i16* %X             ; <i16> [#uses=1]
        %tobool = icmp ne i16 %tmp, 0           ; <i1> [#uses=1]
        %lnot = xor i1 %tobool, true            ; <i1> [#uses=1]
        zext i1 %lnot to i32            ; <i32>:0 [#uses=1]
        %tobool1 = icmp ne i32 %0, 0            ; <i1> [#uses=1]
        br i1 %tobool1, label %ifthen, label %ifend

llvm-svn: 39559

llvm-svn: 39558

- Added C99 reference to why "auto" and "register" cannot be used as a
  storage class specifier for file scoped variable declarations.

llvm-svn: 39557

llvm-svn: 39556