Commit Graph

359 Commits

Author SHA1 Message Date
Douglas Gregor a5cb6da0cd Handle substitutions into the "first qualifier in scope" of a
qualified member access expression (e.g., t->U::member) when that
first qualifier refers to a template parameters.

llvm-svn: 84612
2009-10-20 05:58:46 +00:00
Douglas Gregor c59e56190e Parse a simple-template-id following a '~' when calling a destructor, e.g.,
t->~T<A0, A1>()

Fixes PR5213.

llvm-svn: 84545
2009-10-19 22:04:39 +00:00
Douglas Gregor 4bbd1acf8b When type-checking a C++ "new" expression, don't type-check the actual
initialization if any of the constructor/initialization arguments are
type-dependent. Fixes PR5224.

llvm-svn: 84365
2009-10-17 21:40:42 +00:00
Douglas Gregor 12e49d3250 Make the remaining explicit-instantiation semantic action use
CheckSpecializationInstantiationRedecl to check for
redeclarations/instantiations. Also fixes a longstanding issue where
our explicit-instantiation location information wasn't as good as it
could have been.

llvm-svn: 84216
2009-10-15 22:53:21 +00:00
Douglas Gregor e47f5a76cc Additional semantic checking for explicit template instantiations,
focusing on the scope- and qualifier-related semantic requirements in
C++ [temp.explicit]p2.

llvm-svn: 84154
2009-10-14 23:41:34 +00:00
Douglas Gregor 568a071dd2 When mapping from an injected-class-name to its corresponding
template, make sure to get the template that corresponds to *this*
declaration of the class template or specialization, rather than the
canonical specialization. Fixes PR5187.

llvm-svn: 84119
2009-10-14 17:30:58 +00:00
Douglas Gregor 4aa2dc41dc Implement support for overloaded operator uses that result to a call
to a member operator template. We missed updating this call site when
adding support for function templates; bug exposed by a test for
PR5072.

llvm-svn: 84111
2009-10-14 16:50:13 +00:00
Douglas Gregor 15e5602e59 Improve diagnostics when the parser encounters a declarator with an
unknown type name, e.g.,

  foo::bar x;

when "bar" does not refer to a type in "foo". 

With this change, the parser now calls into the action to perform
diagnostics and can try to recover by substituting in an appropriate
type. For example, this allows us to easily diagnose some missing
"typename" specifiers, which we now do:

  test/SemaCXX/unknown-type-name.cpp:29:1: error: missing 'typename'
        prior to dependent type name 'A<T>::type'
  A<T>::type A<T>::f() { return type(); }
  ^~~~~~~~~~
  typename 

Fixes PR3990.

llvm-svn: 84053
2009-10-13 23:27:22 +00:00
Douglas Gregor e40876a50c Unify our diagnostic printing for errors of the form, "we didn't like
what we found when we looked into <blah>", where <blah> is a
DeclContext*. We can now format DeclContext*'s in nice ways, e.g.,
"namespace N", "the global namespace", "'class Foo'".

This is part of PR3990, but we're not quite there yet.

llvm-svn: 84028
2009-10-13 21:16:44 +00:00
Douglas Gregor 3a88c1d784 Improve the internal representation and semantic analysis of friend
function templates.

This commit ensures that friend function templates are constructed as
FunctionTemplateDecls rather than partial FunctionDecls (as they
previously were). It then implements template instantiation for friend
function templates, injecting the friend function template only when
no previous declaration exists at the time of instantiation. 

Oh, and make sure that explicit specialization declarations are not
friends.

llvm-svn: 83970
2009-10-13 14:39:41 +00:00
Douglas Gregor 06db9f50a2 Diagnose the declaration of explicit specializations after an implicit
instantiation has already been required. To do so, keep track of the
point of instantiation for anything that can be instantiated.

llvm-svn: 83890
2009-10-12 20:18:28 +00:00
Douglas Gregor 7f34baeb4b When declaring a friend class template, we may end up finding an
injected-class-name (e.g., when we're referring to other
specializations of the current class template). Make sure that we see
the template rather than the injected-class-name. Fixes PR4768.

llvm-svn: 83672
2009-10-09 21:11:42 +00:00
Douglas Gregor c093c1de2b Make sure to set the template specialization kind of an explicit
template instantiation of a member function of a class template.
FIXME -= 2;

llvm-svn: 83520
2009-10-08 01:19:17 +00:00
Douglas Gregor d801b06232 Keep track of whether a member function instantiated from a member
function of a class template was implicitly instantiated, explicitly
instantiated (declaration or definition), or explicitly
specialized. The same MemberSpecializationInfo structure will be used
for static data members and member classes as well.

llvm-svn: 83509
2009-10-07 23:56:10 +00:00
Douglas Gregor 5c0405d484 Type checking for specializations of member functions of class
templates. Previously, these weren't handled as specializations at
all. The AST for representing these as specializations is still a work
in progress.

llvm-svn: 83498
2009-10-07 22:35:40 +00:00
Douglas Gregor 548886518d Refactor checking of the scope of explicit template specialization
declarations and explicit template instantiations, improving
diagnostics and making the code usable for function template
specializations (as well as class template specializations and partial
specializations). 

llvm-svn: 83436
2009-10-07 00:13:32 +00:00
Douglas Gregor f343fd8929 Make sure to free the explicit template arguments provided for an
explicit instantiation. Also, tighten up reference-count checking to
help catch these issues earlier. Fixes PR5069.

llvm-svn: 83225
2009-10-01 23:51:25 +00:00
Douglas Gregor 66950a32d9 When overload resolution fails for an overloaded operator, show the
overload candidates (but not the built-in ones). We still rely on the
underlying built-in semantic analysis to produce the initial
diagnostic, then print the candidates following that diagnostic. 

One side advantage of this approach is that we can perform more validation
of C++'s operator overloading with built-in candidates vs. the
semantic analysis for those built-in operators: when there are no
viable candidates, we know to expect an error from the built-in
operator handling code. Otherwise, we are not modeling the built-in
semantics properly within operator overloading. This is checked as:

      assert(Result.isInvalid() && 
             "C++ binary operator overloading is missing
             candidates!");
      if (Result.isInvalid())
        PrintOverloadCandidates(CandidateSet, /*OnlyViable=*/false);

The assert() catches cases where we're wrong in a +Asserts build. The
"if" makes sure that, if this happens in a production clang
(-Asserts), we still build the proper built-in operator and continue
on our merry way. This is effectively what happened before this
change, but we've added the assert() to catch more flies.

llvm-svn: 83175
2009-09-30 21:46:01 +00:00
Douglas Gregor a49cb4e494 Slightly improve the semantics of extern templates for member functions of class templates
llvm-svn: 83063
2009-09-29 14:38:03 +00:00
Douglas Gregor c5790dfeb4 Parse a C++ scope specifier followed by a "typename" annotation token as a type name within the declaration specifiers. Fixes PR5061.
llvm-svn: 82974
2009-09-28 07:26:33 +00:00
Douglas Gregor a29a3ffd18 Improve handling of friend function templates somewhat
llvm-svn: 82950
2009-09-28 00:08:27 +00:00
Douglas Gregor 2208a2912c Simplify the handling of non-dependent friend class template
specializations such as:

  friend class std::vector<int>;

by using the same code path as explicit specializations, customized to
reference an existing ClassTemplateSpecializationDecl (or build a new
"undeclared" one).

llvm-svn: 82875
2009-09-26 20:57:03 +00:00
Douglas Gregor 3dad842b35 Rework the Parse-Sema interaction for friends to better support friend
class templates. We now treat friend class templates much more like
normal class templates, except that they still get special name lookup
rules. Fixes PR5057 and eliminates a bunch of spurious diagnostics in
<iostream>.

llvm-svn: 82848
2009-09-26 06:47:28 +00:00
Douglas Gregor 0e876e0180 Use explicitly-specified template argument lists to help naming
explicit template specializations, when available.

llvm-svn: 82824
2009-09-25 23:53:26 +00:00
Douglas Gregor d90fd526d3 Declarators can now properly represent template-ids, e.g., for
template void f<int>(int);
                ~~~~~~
Previously, we silently dropped the template arguments. With this
change, we now use the template arguments (when available) as the
explicitly-specified template arguments used to aid template argument
deduction for explicit template instantiations.

llvm-svn: 82806
2009-09-25 21:45:23 +00:00
Douglas Gregor 450f0084a2 WIP implementation of explicit instantiation of function templates,
member functions of class template specializations, and static data
members. The mechanics are (mostly) present, but the semantic analysis
is very weak.

llvm-svn: 82789
2009-09-25 18:43:00 +00:00
Douglas Gregor 56751b5981 Fix checking for a null pointer constant when the expression itself is
value-dependent. Audit (and fixed) all calls to
Expr::isNullPointerConstant() to provide the correct behavior with
value-dependent expressions. Fixes PR5041 and a crash in libstdc++
<locale>.

In the same vein, properly compute value- and type-dependence for
ChooseExpr. Fixes PR4996.

llvm-svn: 82748
2009-09-25 04:25:58 +00:00
Douglas Gregor 5013a7e42d When entering the scope of a declarator, make sure that the scope is
complete (or, possibly causing template instantiation).

Test this via some explicit specializations of member functions.

llvm-svn: 82732
2009-09-24 23:39:01 +00:00
Douglas Gregor 3a923c2d37 WIP implementation of explicit function template specialization. This
first implementation recognizes when a function declaration is an
explicit function template specialization (based on the presence of a
template<> header), performs template argument deduction + ambiguity
resolution to determine which template is being specialized, and hooks

There are many caveats here:
  - We completely and totally drop any explicitly-specified template
  arguments on the floor
  - We don't diagnose any of the extra semantic things that we should
  diagnose. 
  - I haven't looked to see that we're getting the right linkage for
  explicit specializations

On a happy note, this silences a bunch of errors that show up in
libstdc++'s <iostream>, although Clang still can't get through the
entire header.

llvm-svn: 82728
2009-09-24 23:14:47 +00:00
Douglas Gregor 4f15f4dec1 Merge uninstantiated default arguments more carefully, and try not to
complain about specializations of member functions that are not
definitions. Fixes PR4995.

llvm-svn: 82159
2009-09-17 19:51:30 +00:00
Douglas Gregor 64621e6eb3 Teach Sema::FindInstantiatedDecl to find instantiated RecordDecls even
when we are not instantiating the corresponding "current
instantiation." This happens, e.g., when we are instantiating a
declaration reference that refers into the "current instantiation" but
occurs in a default function argument. The libstdc++ vector default
constructor now instantiates properly.

llvm-svn: 82069
2009-09-16 18:34:49 +00:00
Douglas Gregor 45068b3e29 Commit test case from PR4655, which tests the canonical-types fix in r81913
llvm-svn: 81916
2009-09-15 21:14:05 +00:00
Douglas Gregor 4fb9cde8ef When printing an overload candidate that is a function template specialization,
point at the template and print out its template arguments, e.g.,

ambiguous-ovl-print.cpp:5:8: note: candidate function template specialization
      [with T = int]
  void f(T*, long);

llvm-svn: 81907
2009-09-15 20:11:42 +00:00
Douglas Gregor 31fae89497 Add an assertion and a test case, in a fruitless attempt to track down an existing bug
llvm-svn: 81885
2009-09-15 18:26:13 +00:00
Douglas Gregor be999390eb Implement partial ordering of class template partial specializations
(C++ [temp.class.order]).

llvm-svn: 81866
2009-09-15 16:23:51 +00:00
Douglas Gregor 6010da024c Implement partial ordering of function templates when calling a
conversion function.

llvm-svn: 81807
2009-09-14 23:02:14 +00:00
Douglas Gregor 0ff7d92048 Implement partial ordering of function template specializations
(C++ [temp.func.order]). 

llvm-svn: 81777
2009-09-14 18:39:43 +00:00
Anders Carlsson 40a90c8754 Fix a bug in getFriendObjectKind where we would clear out all mask bits except IDNS_TagFriend and IDNS_OrdinaryFriend but then check for IDNS_Tag and IDNS_Ordinary.
llvm-svn: 81730
2009-09-13 23:59:13 +00:00
Douglas Gregor 4aa04b155a Slight improvement for extern templates, so that an explicit
instantiation definition can follow an explicit instantiation
declaration. This is as far as I want to go with extern templates now,
but they will still need quite a bit more work to get all of the C++0x
semantics right.

llvm-svn: 81573
2009-09-11 21:19:12 +00:00
Douglas Gregor dd3f5f1fca Improve testing for extern temp templates, slightly. We are (properly) suppressing the implicit instantiation of members of extern templates
llvm-svn: 81567
2009-09-11 20:35:49 +00:00
Douglas Gregor c732aba9a9 Cleanup and test C++ default arguments. Improvements include:
- Diagnose attempts to add default arguments to templates (or member
    functions of templates) after the initial declaration (DR217).
  - Improve diagnostics when a default argument is redefined. Now, the
    note will always point at the place where the default argument was
    previously defined, rather than pointing to the most recent
    declaration of the function.

llvm-svn: 81548
2009-09-11 18:44:32 +00:00
John McCall 1806c2795b Track a class template specialization's point of instantiation separately
from its location.  Initialize appropriately.

When implicitly creating a declaration of a class template specialization
after encountering the first reference to it, use the pattern class's
location instead of the location of the first reference.

llvm-svn: 81515
2009-09-11 07:25:08 +00:00
Douglas Gregor e61ef62cc2 When re-entering a template scope, we may be entering a class template
partial specialization rather than a subclass of TemplateDecl. Fixes a
crash in libstdc++ 4.2's <map>.

llvm-svn: 81407
2009-09-10 00:12:48 +00:00
Douglas Gregor 308047d3a5 Initial stab at implement dependent member references to member
templates, e.g.,
  
  x.template get<T>

We can now parse these, represent them within an UnresolvedMemberExpr
expression, then instantiate that expression node in simple cases.

This allows us to stumble through parsing LLVM's Casting.h.

llvm-svn: 81300
2009-09-09 00:23:06 +00:00
Anders Carlsson faf1ced5ee Reapply 81096, now with a fix. Spot the bug:
for (unsigned i = numargs; i < NumArgs; ++i)
        Args[0] = 0;

;)

llvm-svn: 81123
2009-09-06 16:54:02 +00:00
Daniel Dunbar 94bcae46d5 Revert "Initialize default CXXConstructExpr arguments to 0. Fixes a crash when
destroying the CXXConstructExpr.", this is causing test failures across the
board.

llvm-svn: 81100
2009-09-06 01:31:23 +00:00
Anders Carlsson 8284d7caa3 Initialize default CXXConstructExpr arguments to 0. Fixes a crash when destroying the CXXConstructExpr.
llvm-svn: 81096
2009-09-05 22:51:06 +00:00
Anders Carlsson dc6d2c3cda Report errors for member functions correctly.
llvm-svn: 81063
2009-09-05 05:38:54 +00:00
Anders Carlsson 657bad441e Use a separate diagnostic for default function argument expressions.
llvm-svn: 81062
2009-09-05 05:14:19 +00:00
Douglas Gregor d94105a1c2 Don't generate any code for an explicit call to a trivial destructor.
Now that parsing, semantic analysis, and (I think) code generation of
pseudo-destructor expressions and explicit destructor calls works,
update the example-dynarray.cpp test to destroy the objects it
allocates and update the test to actually compile + link.
The code seems correct, but the Clang-compiled version dies with a
malloc error. Time to debug!

llvm-svn: 81025
2009-09-04 19:04:08 +00:00