5b3638b6e7
Generate the correct results for the comma expression. Fixes PR3123.
...
llvm-svn: 60334
2008-12-01 06:44:05 +00:00
564730a857
Emit the correct diagnostic when a comma is in an ICE.
...
llvm-svn: 60316
2008-12-01 02:07:06 +00:00
b33d6c8611
Change the diagnostics that the evaluator reports to be of type NOTE.
...
llvm-svn: 60301
2008-11-30 18:37:00 +00:00
5862001157
Replace the isEvaluated bool with a ShortCircuit int, making it easier to handle recursion
...
llvm-svn: 60300
2008-11-30 18:26:25 +00:00
27b8c5c464
Pass the expression to the Error and Extension methods.
...
llvm-svn: 60299
2008-11-30 18:14:57 +00:00
7b6f0af83f
Add a new variant of Evaluate and reimplement the old Evaluate in terms of the new.
...
llvm-svn: 60298
2008-11-30 16:58:53 +00:00
f50de0c590
General cleanup, evaluate the RHS of a logical op even if the LHS will give us the result.
...
llvm-svn: 60297
2008-11-30 16:51:17 +00:00
bd1df8e866
EvalInfo now holds a reference to an EvalResult struct.
...
llvm-svn: 60296
2008-11-30 16:38:33 +00:00
615f34c719
remove debug-only assertion in the complex float evaluator as it makes some real apps crash
...
llvm-svn: 60069
2008-11-26 00:39:53 +00:00
a42ee44aab
The address of a variable is only constant if the variable has global storage.
...
llvm-svn: 59939
2008-11-24 04:41:22 +00:00
4c76e93507
Fix bug in the constant evaluator. Fixes PR3115.
...
llvm-svn: 59938
2008-11-24 04:21:33 +00:00
7c282e4f3f
Return false if we encounter a type we can't handle.
...
llvm-svn: 59889
2008-11-22 22:56:32 +00:00
475f4bce36
Case values must be evaluated
...
llvm-svn: 59884
2008-11-22 21:50:49 +00:00
59689ed764
Use Expr::Evaluate for case statements. Fixes PR2525
...
llvm-svn: 59881
2008-11-22 21:04:56 +00:00
0e33c688d5
fix folding of '*doubleArray'
...
llvm-svn: 59647
2008-11-19 17:44:31 +00:00
899c7050b6
Address some comments Eli had.
...
llvm-svn: 59430
2008-11-16 22:46:56 +00:00
527b5a6858
use HandleConversionToBool() to check if a given cond is foldable (per Eli's comment)
...
llvm-svn: 59429
2008-11-16 22:06:39 +00:00
9ddf7bedf9
More complex float evaluator support.
...
llvm-svn: 59428
2008-11-16 21:51:21 +00:00
67d7b9204c
rename Expr::tryEvaluate to Expr::Evaluate.
...
llvm-svn: 59426
2008-11-16 21:24:15 +00:00
537969c237
Add very limited support for evaluating complex floats.
...
llvm-svn: 59425
2008-11-16 20:27:53 +00:00
e236a48210
fix folding of comma if given a non-constant operand.
...
Eli please take a look, as I'm not sure if this gets the extension warning in the right place
llvm-svn: 59422
2008-11-16 20:09:07 +00:00
420426100c
make IntExprEvaluator fold the ?: operator
...
llvm-svn: 59421
2008-11-16 19:28:31 +00:00
9f9e4249cc
More work on the constant evaluator. Eli, it would be great if you could have a look at this.
...
llvm-svn: 59420
2008-11-16 19:01:22 +00:00
acc7981883
Add the ability to evaluate comparison operators with floating point numbers as operands.
...
llvm-svn: 59408
2008-11-16 07:17:21 +00:00
5a332ea01f
Fix for crash issues with comma operators with a void first operand, and
...
some more bullet-proofing/enhancements for tryEvaluate. This shouldn't
cause any behavior changes except for handling cases where we were
crashing before and being able to evaluate a few more cases in tryEvaluate.
This should settle the minor mess surrounding r59196.
llvm-svn: 59224
2008-11-13 06:09:17 +00:00
8553a98b8e
Backout of r59196, plus a new ICE test. Sorry if this is a
...
little rude; I figure it's cleaner to just back this out now so
it doesn't get forgotten or mixed up with other checkins.
The modification to isICE is simply wrong; I've added a test that the
change to isICE breaks.
I'm pretty sure the modification to tryEvaluate is also wrong.
At the very least, there's some serious miscommunication going on here,
as this is going in exactly the opposite direction of r59105. My
understanding is that tryEvaluate is not supposed to care about side
effects. That said, a lot of the clients to tryEvaluate are
expecting it to enforce a no-side-effects policy, so we probably need
another method that provides that guarantee.
llvm-svn: 59212
2008-11-13 02:13:11 +00:00
c7ec5fcf24
Fix bug in constant evaluation exposed by 176.gcc.
...
- Evaluation of , operator used bogus assumption that LHS could be
evaluated as an integral expression even though its type is
unspecified.
This change is making isICE very permissive of the LHS in non-evaluated
contexts because it is not clear what predicate we would use to reject
code here. The standard didn't offer me any guidance; opinions?
llvm-svn: 59196
2008-11-13 00:03:19 +00:00
a461e6d124
Disable some debug prints.
...
llvm-svn: 59174
2008-11-12 21:52:46 +00:00
9a156e5c1e
Some additions to tryEvaluate I've had sitting around for a while.
...
This pushes it a lot closer to being able to deal with most of the stuff
CodeGen's constant expression evaluator knows how to deal with. This
also fixes PR3003.
The test could possibly use some improvement, but this'll work for now.
Test 6 is inspired by PR3003; the other tests are mostly just designed
to exercise the new code. The reason for the funny structure of the
tests is that type fixing for arrays inside of structs is the only place
in Sema that calls tryEvaluate, at least for the moment.
llvm-svn: 59125
2008-11-12 09:44:48 +00:00
fac05aeae5
fix a crash analyzing constants in 176.gcc/expr.c with my next patch. It was
...
crashing because we errors are ignored in subexpressions that are not evaluated,
but we still evaluate the result of parents. This would cause an assertion
because the erroneous subexpr didn't have its result value set to the right type.
llvm-svn: 59110
2008-11-12 07:43:42 +00:00
aa3f951efc
Teach the aggressive constant folder to fold X && 0 -> 0 and X || 1 -> 1
...
llvm-svn: 59105
2008-11-12 07:04:29 +00:00
6f28289aca
Introduce a single AST node SizeOfAlignOfExpr for all sizeof and alignof expressions, both of values and types.
...
llvm-svn: 59057
2008-11-11 17:56:53 +00:00
d7be95d60a
PR2919: __builtin_types_compatible_p strips CRV qualifiers.
...
llvm-svn: 58079
2008-10-24 08:07:57 +00:00
c3d79cf6b3
Teach tryEvaluate that fabs, copysign, and unary +/- are constants for
...
floats.
- With testcase, which also has some other things GCC folds but we
don't commented out in it.
llvm-svn: 57624
2008-10-16 03:51:50 +00:00
1be9f88a2f
Fix test suite regression, getFloatTypeSemantics shouldn't be called
...
on non-float types.
llvm-svn: 57477
2008-10-14 05:41:12 +00:00
c48814bc98
Adjust calls to APFloat conversion for new interface.
...
llvm-svn: 57332
2008-10-09 23:02:32 +00:00
2c091b1ead
ExprConstant should not abort when it sees a pointer constant that isn't.
...
llvm-svn: 57162
2008-10-06 07:20:11 +00:00
cb13691a62
Add a Expr::isEvaluatable method, eliminate isBuiltinConstantExpr
...
which is checking for something that can be inconsistent with
what we can constant fold.
llvm-svn: 57159
2008-10-06 06:49:02 +00:00
86ee286b09
Move folding of __builtin_classify_type out of the CallExpr
...
interface into the constant folding interface.
llvm-svn: 57158
2008-10-06 06:40:35 +00:00
0b7282eafc
Move handling of __builtin_nan("") out of CGBuiltin.cpp into ExprConstant.cpp
...
llvm-svn: 57157
2008-10-06 06:31:58 +00:00
37346e081c
Teach FloatExprEvaluator to evaluate __builtin_huge_val and inf.
...
llvm-svn: 57154
2008-10-06 05:53:16 +00:00
4deaa4ea24
Add a comment that describes tryEvaluate. Make tryEvaluate fold
...
__builtin_constant_p properly, and add some scaffolding for
FloatExprEvaluator to eventually handle huge_val and inf.
llvm-svn: 57152
2008-10-06 05:28:25 +00:00
24c0154b77
Initial implementation of floats in Expr::tryEvaluate; this doesn't
...
implement some things, like unary operators and casts, but it's enough
to fix PR2703 as filed.
llvm-svn: 55155
2008-08-22 00:06:13 +00:00
3bab3d21f9
Add ExplicitCastExpr to replace the current CastExpr, and have ImplicitCastExpr and ExplicitCastExpr derive from a common base class (CastExpr):
...
Expr
-> CastExpr
-> ExplicitCastExpr
-> ImplicitCastExpr
llvm-svn: 54955
2008-08-18 23:01:59 +00:00
56fdb6ae69
More #include cleaning
...
- Kill unnecessary #includes in .cpp files. This is an automatic
sweep so some things removed are actually used, but happen to be
included by a previous header. I tried to get rid of the obvious
examples and this was the easiest way to trim the #includes in one
fell swoop.
- We now return to regularly scheduled development.
llvm-svn: 54632
2008-08-11 06:23:49 +00:00
d361a7df78
Remove the old evaluator code.
...
llvm-svn: 54615
2008-08-10 17:03:01 +00:00
5cf85cc977
Turn on the new constant expression evaluator; it isn't
...
anywhere near complete, but I don't know of any issues, and it can
handle enough to allow removing the I-C-E pointer cast hack.
llvm-svn: 54597
2008-08-09 22:21:56 +00:00
bd84059cdb
Minor tweak plus a couple of FIXMEs.
...
llvm-svn: 54119
2008-07-27 05:46:18 +00:00
477c4be1e9
expand casts to handle what isIntegerConstantExpr handles.
...
Casts from pointers remain a problem. :)
llvm-svn: 53501
2008-07-12 01:15:53 +00:00
7174bf3722
Add support for __builtin_type_compatible_p, enums, etc.
...
llvm-svn: 53500
2008-07-12 00:38:25 +00:00
Anders Carlsson