209 Commits

Author SHA1 Message Date
John McCall
da4458e98f Regularize support for naming conversion functions in using decls.
llvm-svn: 99979
2010-03-31 01:36:47 +00:00
John McCall
16df1e59f2 Propagate the "found declaration" (i.e. the using declaration instead of
the underlying/instantiated decl) through a lot of API, including "intermediate"
MemberExprs required for (e.g.) template instantiation.  This is necessary
because of the access semantics of member accesses to using declarations:
only the base class *containing the using decl* need be accessible from the
naming class.

This allows us to complete an access-controlled selfhost, if there are no
recent regressions.

llvm-svn: 99936
2010-03-30 21:47:33 +00:00
Douglas Gregor
838fcc318a Switch semantic analysis of the conditional operator from using
CheckReferenceInit to using the new initialization sequence code.

llvm-svn: 99647
2010-03-26 20:14:36 +00:00
Douglas Gregor
031296e316 Kill off two more uses of Sema::CheckReferenceInit in favor of the new
initialization code. Exposed a bug where we were not marking an
implicit conversion as an lvalue when we were forming a call to a
conversion function whose return type is a reference.

llvm-svn: 99459
2010-03-25 00:20:38 +00:00
John McCall
a0296f7987 Remember the "found declaration" for an overload candidate, which is the
entity (if applicable) which was actually looked up.  If a candidate was found
via a using declaration, this is the UsingShadowDecl;  otherwise, if
the candidate is template specialization, this is the template;  otherwise,
this is the function.

The point of this exercise is that "found declarations" are the entities
we do access control for, not their underlying declarations.  Broadly speaking,
this patch fixes access control for using declarations.

There is a *lot* of redundant code calling into the overload-resolution APIs;
we really ought to clean that up.

llvm-svn: 98945
2010-03-19 07:35:19 +00:00
John McCall
1064d7ef29 Perform access control for the implicit base and member destructor calls
required when emitting a destructor definition.

llvm-svn: 98609
2010-03-16 05:22:47 +00:00
John McCall
e40b58ec2f Implement -Wmissing-field-initializers. Patch by mikem!
llvm-svn: 98275
2010-03-11 19:32:38 +00:00
Douglas Gregor
1ce52caf0d Reference binding via user-defined conversion can compute a binding
that is not reference-related (because it requires another implicit
conversion to which we can find). Fixes PR6483.

llvm-svn: 97922
2010-03-07 23:17:44 +00:00
Tanya Lattner
5cb196e14b Fix 80 col violation.
llvm-svn: 97898
2010-03-07 04:47:12 +00:00
Tanya Lattner
5cbff48b01 Fix indentation, use string directly instead of StringRef.
llvm-svn: 97896
2010-03-07 04:40:06 +00:00
Tanya Lattner
351e4393f1 Fix some weird patch issue.
llvm-svn: 97894
2010-03-07 04:27:11 +00:00
Tanya Lattner
5029d56cc1 Implement missing-braces warning and add a test case.
llvm-svn: 97893
2010-03-07 04:17:15 +00:00
Douglas Gregor
cc3f325fa6 Reinstate r97674 with a fix for the assertion that was firing in <list>
llvm-svn: 97686
2010-03-03 23:55:11 +00:00
Douglas Gregor
979302e14a Revert r97674; it's causing failures
llvm-svn: 97677
2010-03-03 23:26:56 +00:00
Douglas Gregor
f2a42a66e2 Implement disambiguation of base class members via a
nested-name-specifier. For example, this allows member access in
diamond-shaped hierarchies like:

  struct Base {
    void Foo();
    int Member;
  };

  struct D1 : public Base {};
  struct D2 : public Base {};

  struct Derived : public D1, public D2 { }

  void Test(Derived d) {
    d.Member = 17; // error: ambiguous cast from Derived to Base
    d.D1::Member = 17; // error: okay, modify D1's Base's Member
  }

Fixes PR5820 and <rdar://problem/7535045>. Also, eliminate some
redundancy between Sema::PerformObjectMemberConversion() and
Sema::PerformObjectArgumentInitialization() -- the latter now calls
the former.

llvm-svn: 97674
2010-03-03 22:53:40 +00:00
Douglas Gregor
9bc6b7fc53 Use CXXTemporaryObjectExpr for explicitly-constructed temporaries. We
used to do this, but it got lost when we switched functional-style
cast syntax over to using the new initialization code. Fixes PR6457.

llvm-svn: 97568
2010-03-02 17:18:33 +00:00
Douglas Gregor
e6d276a559 Commit Eli's fix for implicit conversions to array type. Fixes PR6264.
llvm-svn: 97202
2010-02-26 01:17:27 +00:00
Ted Kremenek
013041eef0 Revert: "Change InitListExpr to allocate the array for holding references"
This was causing buildbot breakage.

This reverts commit d46e952cc8cb8d9eed8657d9a0b267910a0f745a.

llvm-svn: 96652
2010-02-19 01:50:18 +00:00
Ted Kremenek
445a603c78 Change InitListExpr to allocate the array for holding references
to initializer expressions in an array allocated using ASTContext.

This plugs a memory leak when ASTContext uses a BumpPtrAllocator to
allocate memory for AST nodes.

In my mind this isn't an ideal solution; it would be nice to have
a general "vector"-like class that allocates memory using ASTContext,
but whose guts could be separated from the methods of InitListExpr
itself.  I haven't gone and taken this approach yet because it isn't
clear yet if we'll eventually want an alternate solution for recylcing
memory using by InitListExprs as we are constructing the ASTs.

llvm-svn: 96642
2010-02-19 00:42:33 +00:00
Chandler Carruth
dfe198b588 Silence a GCC warning about a possibly uninitialized variable. It's data flow
only flows so far it seems.

llvm-svn: 96085
2010-02-13 07:23:01 +00:00
Douglas Gregor
e6565625f4 Migrate the mish-mash of declaration checks in
Sema::ActOnUninitializedDecl over to InitializationSequence (with
default initialization), eliminating redundancy. More importantly, we
now check that a const definition in C++ has an initilizer, which was
an #if 0'd code for many, many months. A few other tweaks were needed
to get everything working again:

  - Fix all of the places in the testsuite where we defined const
    objects without initializers (now that we diagnose this issue)
  - Teach instantiation of static data members to find the previous
    declaration, so that we build proper redeclaration
    chains. Previously, we had the redeclaration chain but built it
    too late to be useful, because...
  - Teach instantiation of static data member definitions not to try
    to check an initializer if a previous declaration already had an
    initializer. This makes sure that we don't complain about static
    const data members with in-class initializers and out-of-line
    definitions.
  - Move all of the incomplete-type checking logic out of
    Sema::FinalizeDeclaratorGroup; it makes more sense in
    ActOnUnitializedDecl.

There may still be a few places where we can improve these
diagnostics. I'll address that as a separate commit.

llvm-svn: 95657
2010-02-09 07:26:29 +00:00
Douglas Gregor
34c0a90265 Be more careful when checking initializer lists that involve reference
types; we don't want to give an expression reference type. Fixes PR6177.

llvm-svn: 95635
2010-02-09 00:50:06 +00:00
John McCall
bc077cf589 Thread a source location into the template-argument deduction routines. There
may be some other places that could take advantage of this new information,
but I haven't really looked yet.

llvm-svn: 95600
2010-02-08 23:07:23 +00:00
Anders Carlsson
fcd764a168 Add support for threadsafe statics, and make them the default (matching gcc).
Daniel, I'd appreciate a review of the driver/cc1 parts.

llvm-svn: 95508
2010-02-06 23:23:06 +00:00
Douglas Gregor
0ab7af6c28 A dependent initializer with zero arguments should return a NULL
initializer (for no initialization) rather than a ParenListExpr with
zero arguments in it.

llvm-svn: 95382
2010-02-05 07:56:11 +00:00
Anders Carlsson
3b227bd629 Revert the new reference binding code; I came up with a way simpler solution for the reference binding bug that is preventing self-hosting.
llvm-svn: 95223
2010-02-03 16:38:03 +00:00
John McCall
c17ae44b09 Fix this comment.
llvm-svn: 95104
2010-02-02 19:02:38 +00:00
John McCall
760af170ff Access checking for implicit user-defined conversions.
llvm-svn: 94971
2010-02-01 03:16:54 +00:00
Anders Carlsson
ab0ddb57b1 Start creating CXXBindReferenceExpr nodes when binding complex types to references.
llvm-svn: 94964
2010-01-31 18:34:51 +00:00
Anders Carlsson
8abde4b447 Diagnose binding a non-const reference to a vector element.
llvm-svn: 94963
2010-01-31 17:18:49 +00:00
Douglas Gregor
7ae2d7758f Rework base and member initialization in constructors, with several
(necessarily simultaneous) changes:

  - CXXBaseOrMemberInitializer now contains only a single initializer
    rather than a set of initialiation arguments + a constructor. The
    single initializer covers all aspects of initialization, including
    constructor calls as necessary but also cleanup of temporaries
    created by the initializer (which we never handled
    before!).

  - Rework + simplify code generation for CXXBaseOrMemberInitializers,
    since we can now just emit the initializer as an initializer.

  - Switched base and member initialization over to the new
    initialization code (InitializationSequence), so that it

  - Improved diagnostics for the new initialization code when
    initializing bases and members, to match the diagnostics produced
    by the previous (special-purpose) code.

  - Simplify the representation of type-checked constructor initializers in
    templates; instead of keeping the fully-type-checked AST, which is
    rather hard to undo at template instantiation time, throw away the
    type-checked AST and store the raw expressions in the AST. This
    simplifies instantiation, but loses a little but of information in
    the AST.

  - When type-checking implicit base or member initializers within a
    dependent context, don't add the generated initializers into the
    AST, because they'll look like they were explicit.

  - Record in CXXConstructExpr when the constructor call is to
  initialize a base class, so that CodeGen does not have to infer it
  from context. This ensures that we call the right kind of
  constructor.

There are also a few "opportunity" fixes here that were needed to not
regress, for example:

  - Diagnose default-initialization of a const-qualified class that
    does not have a user-declared default constructor. We had this
    diagnostic specifically for bases and members, but missed it for
    variables. That's fixed now.

  - When defining the implicit constructors, destructor, and
    copy-assignment operator, set the CurContext to that constructor
    when we're defining the body.

llvm-svn: 94952
2010-01-31 09:12:51 +00:00
Anders Carlsson
a18f0fbca2 Eliminate yet another old-school PerformCopyInitialization.
llvm-svn: 94874
2010-01-30 01:56:32 +00:00
Douglas Gregor
d1e08648c6 Fix reference-binding when we have a reference to const volatile type;
previously, we were allowing this to bind to a temporary. Now, we
don't; add test-cases and improve diagnostics.

llvm-svn: 94831
2010-01-29 19:39:15 +00:00
Douglas Gregor
65eb86e912 Fix reference binding of const lvalue references to bit-fields, which
requires a temporary. Previously, we were building an initialization
sequence that bound to the bit-field as if it were a real lvalue. Note
that we previously (and still) diagnose binding of non-const
references to bit-fields, as we should.

There's no real way to test that this code is correct, since reference
binding does not *currently* have any representation in the AST. This
fix should make it easier for that to happen, so I've verified this
fix with...

Added InitializationSequence::dump(), to print an initialization
sequence for debugging purposes.

llvm-svn: 94826
2010-01-29 19:14:02 +00:00
Anders Carlsson
a91be641a7 Simplify InitListChecker::CheckReferenceType
llvm-svn: 94793
2010-01-29 02:47:33 +00:00
John McCall
b89836b6db Pass access specifiers around in overload resolution.
llvm-svn: 94485
2010-01-26 01:37:31 +00:00
Eli Friedman
673f94a0c9 Code simplification; this should have no visible effects.
llvm-svn: 94427
2010-01-25 17:04:54 +00:00
Anders Carlsson
0bd52403d4 Use new initialization code when dealing with [dcl.init.aggr]p12. This fixes the bug where array elements and member initializers weren't copied correctly.
llvm-svn: 94340
2010-01-24 00:19:41 +00:00
Anders Carlsson
6cabf318c9 Change all InitializedEntity pointers over to be references.
llvm-svn: 94335
2010-01-23 23:23:01 +00:00
Anders Carlsson
3fa93b7e54 Start passing InitializedEntity to CheckDesignatedInitializer.
llvm-svn: 94334
2010-01-23 22:49:02 +00:00
Anders Carlsson
dbb25a38b9 More init work, adding more entity parameters.
llvm-svn: 94332
2010-01-23 20:47:59 +00:00
Anders Carlsson
73eb7cd351 Use the new init code for member subobjects.
llvm-svn: 94329
2010-01-23 20:20:40 +00:00
Anders Carlsson
0cf999b663 Switch some array initialization over to the new init code.
llvm-svn: 94327
2010-01-23 20:13:41 +00:00
Anders Carlsson
d0849254de Baby steps towards migrating the InitListChecker over to the new initialization code. Pass an InitializedEntity pointer through to most init checker functions. Right now, it's ignored everywhere except when initializing vectors in C++.
llvm-svn: 94325
2010-01-23 19:55:29 +00:00
Anders Carlsson
3cc795ae8d When an InitializedEntity is passed to CheckSingleInitializer, call the new Sema::PerformCopyInitialization overload.
llvm-svn: 94324
2010-01-23 19:22:30 +00:00
Anders Carlsson
26d0564fd1 Change CheckSingleInitializer to take/return OwningExprResults instead. No functionality change.
llvm-svn: 94316
2010-01-23 18:35:41 +00:00
Anders Carlsson
ed8d80d72b Separate EK_ArrayOrVectorElement into EK_ArrayElement and EK_VectorElement; arrays and vectors are pretty different beasts in C++. Doug, please review/comment.
llvm-svn: 94279
2010-01-23 04:34:47 +00:00
John McCall
ad371258aa Give UnresolvedSet the ability to store access specifiers for each declaration.
Change LookupResult to use UnresolvedSet.  Also extract UnresolvedSet into its
own header and make it templated over an inline capacity.

llvm-svn: 93959
2010-01-20 00:46:10 +00:00
Chandler Carruth
04bdce6407 Fix the CodeGen half of PR5911 by changing reference initialization to
correctly look through arrays to see cv-qualifiers. Also enhances the routine
for doing this to preserve more type sugaring for diagnostics.

llvm-svn: 93252
2010-01-12 20:32:25 +00:00
John McCall
ad907777e8 So I was sitting around, trying vainly to think of something to commit, and then
I said to myself, self, why don't you go add a couple of parameters to a method
and then fail to use them, and I thought that sounded like a pretty good idea,
so I did it.

llvm-svn: 93233
2010-01-12 07:18:19 +00:00