Chandler Carruth [Tue, 21 Feb 2012 21:51:40 +0000 (21:51 +0000)]
Clean up, add some documentation, and make this test return to checking
the linker toolchainness a bit more thoroughly. It used to work this
way, but hit buildbot issues. Hopefully subsequent fixes have addressed
those problems, but I'll be watching the bots.
Douglas Gregor [Tue, 21 Feb 2012 20:05:31 +0000 (20:05 +0000)]
Only pop the expression evaluation context corresponding to a lambda
expression after we've finished the function body of the corresponding
function call operator. Otherwise, ActOnFinishFunctionBody() will see
the (unfinished) evaluation context of the lambda expression
itself. Fixes PR12031.
Richard Smith [Tue, 21 Feb 2012 20:01:35 +0000 (20:01 +0000)]
Don't assume that a valid expression for the first part of a for-statement
is non-null when diagnosing a broken attempt to write a for-range-statement.
When calling a non variadic format function(vprintf, vscanf, NSLogv, …), warn if the format string argument is a parameter that is not itself declared as a format string with compatible format.
Douglas Gregor [Tue, 21 Feb 2012 19:11:17 +0000 (19:11 +0000)]
Improve our handling of lambda expressions that occur within default
arguments. There are two aspects to this:
- Make sure that when marking the declarations referenced in a
default argument, we don't try to mark local variables, both because
it's a waste of time and because the semantics are wrong: we're not
in a place where we could capture these variables again even if it
did make sense.
- When a lambda expression occurs in a default argument of a
function template, make sure that the corresponding closure type is
considered dependent, so that it will get properly instantiated. The
second bit is a bit of a hack; to fix it properly, we may have to
rearchitect our handling of default arguments, parsing them only
after creating the function definition. However, I'd like to
separate that work from the lambdas work.
Douglas Gregor [Tue, 21 Feb 2012 04:17:39 +0000 (04:17 +0000)]
Implement non-internal linkage for lambda closure types that need a
stable mangling, since these lambdas can end up in multiple
translation units. Sema is responsible for deciding when this is the
case, because it's already responsible for choosing the mangling
number.
Douglas Gregor [Tue, 21 Feb 2012 00:37:24 +0000 (00:37 +0000)]
Implement name mangling for lambda expressions that occur within the
default arguments of function parameters. This simple-sounding task is
complicated greatly by two issues:
(1) Default arguments aren't actually a real context, so we need to
maintain extra state within lambda expressions to track when a
lambda was actually in a default argument.
(2) At the time that we parse a default argument, the FunctionDecl
doesn't exist yet, so lambda closure types end up in the enclosing
context. It's not clear that we ever want to change that, so instead
we introduce the notion of the "effective" context of a declaration
for the purposes of name mangling.
[PCH] Recover gracefully if the ASTReader detects that a file is different
from the one stored in the PCH/AST, while trying to load a SLocEntry.
We verify that all files of the PCH did not change before loading it but this is not enough because:
- The AST may have been 1) kept around, 2) to do queries on it.
- We may have 1) verified the PCH and 2) started parsing.
Between 1) and 2) files may change and we are going to have crashes because the rest of clang
cannot deal with the ASTReader failing to read a SLocEntry.
Handle this by recovering gracefully in such a case, by initializing the SLocEntry
with the info from the PCH/AST as well as reporting failure by the ASTReader.
Richard Smith [Mon, 20 Feb 2012 23:28:05 +0000 (23:28 +0000)]
Fix a constexpr FIXME: When implicitly instantiating the primary template for an
explicit specialization of a function template, mark the instantiation as
constexpr if the specialization is, rather than requiring them to match.
Douglas Gregor [Mon, 20 Feb 2012 20:47:06 +0000 (20:47 +0000)]
Make sure that we set up the right declaration contexts when creating
and introducing the lambda closure type and its function call
operator. Previously, we assumed that the lambda closure type would
land directly in the current context, and not some parent context (as
occurs with linkage specifications). Thanks to Richard for the test case.
modern objc translator. Finish off first cut of the
modern meta-data translation by commenting out private ivar
declarations in user source. Also, added several tests.
Douglas Gregor [Mon, 20 Feb 2012 19:44:39 +0000 (19:44 +0000)]
Basic support for name mangling of C++11 lambda expressions. Because
name mangling in the Itanium C++ ABI for lambda expressions is so
dependent on context, we encode the number used to encode each lambda
as part of the lambda closure type, and maintain this value within
Sema.
Note that there are a several pieces still missing:
- We still get the linkage of lambda expressions wrong
- We aren't properly numbering or mangling lambda expressions that
occur in default function arguments or in data member initializers.
- We aren't (de-)serializing the lambda numbering tables
Eric Christopher [Mon, 20 Feb 2012 18:05:24 +0000 (18:05 +0000)]
Remove the type retaining from the clang frontend. This is now
handled by the caching and rauw. Also fix one cache that wasn't
being added to highlighted by this patch. Update all testcases
accordingly.
Douglas Gregor [Mon, 20 Feb 2012 16:12:14 +0000 (16:12 +0000)]
Deserialize the direct-initialization range of a "new" expression
properly. Previously, we deserialized it but failed to set the
corresponding member in CXXNewExpr. Fixes <rdar://problem/10893600>.
Chandler Carruth [Mon, 20 Feb 2012 07:35:45 +0000 (07:35 +0000)]
Add 3dNOW intrinsic header to x86intrin.h, conditioned on __3dNOW__ to
match the behavior of GCC. Also add a test for these intrinsics, which
apparently have *zero* tests. =[ Not surprisingly, Clang crashed when
compiling these.
Fix the bug in CodeGen where we failed to bitcast the argument type to
x86mmx prior to calling the LLVM intrinsic. This fixes an assert on the
new 3dnow-builtins.c test.
This is one issue impacting the efforts to get Clang to emulate the
Microsoft intrinsics headers -- 3dnow intrinsics are implictitly made
available there.
Chandler Carruth [Sun, 19 Feb 2012 23:09:50 +0000 (23:09 +0000)]
Make the regular expressions in this test more narrow to ensure we're
actually matching the write substrings, and stop looking for a leading
'/' to try and finish fixing darwin and other hosts.
Rafael Espindola [Sun, 19 Feb 2012 01:38:32 +0000 (01:38 +0000)]
Implement a -gcc-toolchain command line option that does the same as
configure's --with-gcc-toolchain. The configure option is now just a default
value for the command line one.
Richard Smith [Sat, 18 Feb 2012 22:04:06 +0000 (22:04 +0000)]
Implement constant expression support for __real__ and __imag__ on lvalue
complex numbers. Treat complex numbers as arrays of the corresponding component
type, in order to make std::complex behave properly if implemented in terms of
_Complex T.
Apparently libstdc++'s std::complex is implemented this way, and we were
rejecting a member like this:
constexpr double real() { return __real__ val; }
because it was marked constexpr but unable to produce a constant expression.
Ted Kremenek [Sat, 18 Feb 2012 22:02:57 +0000 (22:02 +0000)]
Fix crash in analyzer diagnostic generation involving subexpressions of OpaqueValueExpr not appearing in the ParentMap. Fixes <rdar://problem/10797980>.
Ted Kremenek [Sat, 18 Feb 2012 20:53:30 +0000 (20:53 +0000)]
Adopt ExprEngine and checkers to ObjC property refactoring. Everything was working, but now diagnostics are aware of message expressions implied by uses of properties. Fixes <rdar://problem/9241180>.
Douglas Gregor [Sat, 18 Feb 2012 09:37:24 +0000 (09:37 +0000)]
Rewrite variable capture within lambda expressions and blocks,
eliminating a bunch of redundant code and properly modeling how the
captures of outside blocks/lambdas affect the types seen by inner
captures.
This new scheme makes two passes over the capturing scope stack. The
first pass goes up the stack (from innermost to outermost), assessing
whether the capture looks feasible and stopping when it either hits
the scope where the variable is declared or when it finds an existing
capture. The second pass then walks down the stack (from outermost to
innermost), capturing the variable at each step and updating the
captured type and the type that an expression referring to that
captured variable would see. It also checks type-specific
restrictions, such as the inability to capture an array within a
block. Note that only the first odr-use of each
variable needs to do the full walk; subsequent uses will find the
capture immediately, so multiple walks need not occur.
The same routine that builds the captures can also compute the type of
the captures without signaling errors and without actually performing
the capture. This functionality is used to determine the type of
declaration references as well as implementing the weird decltype((x))
rule within lambda expressions.
The capture code now explicitly takes sides in the debate over C++
core issue 1249, which concerns the type of captures within nested
lambdas. We opt to use the more permissive, more useful definition
implemented by GCC rather than the one implemented by EDG.
Douglas Gregor [Sat, 18 Feb 2012 05:51:20 +0000 (05:51 +0000)]
Unify our computation of the type of a captured reference to a
variable; it was previously duplicated, and one of the copies failed
to account for outer non-mutable lambda captures.
Richard Smith [Sat, 18 Feb 2012 04:58:18 +0000 (04:58 +0000)]
Fix a problem in the GCC testsuite, exposed by r150557. Compound literals
are represented as prvalues in C++; don't be fooled into thinking they're
global lvalues.
Richard Smith [Sat, 18 Feb 2012 04:13:32 +0000 (04:13 +0000)]
Diagnose uses of deleted destructors and inaccessible defaulted destructors.
We had two separate issues here: firstly, varions functions were assuming that
they did not need to perform semantic checks on trivial destructors (this is
not true in C++11, where a trivial destructor can nonetheless be private or
deleted), and a bunch of DiagnoseUseOfDecl calls were missing for uses of
destructors.
Richard Smith [Sat, 18 Feb 2012 02:02:13 +0000 (02:02 +0000)]
Initial refactoring of 'ShouldDeleteSpecialMember', in preparation for providing
decent diagnostics. Finish the work of combining all the 'ShouldDelete'
functions into one. In unifying the code, fix a minor bug where an anonymous
union with a deleted default constructor as a member of a union wasn't being
considered as making the outer union's default constructor deleted.
Eric Christopher [Sat, 18 Feb 2012 00:50:08 +0000 (00:50 +0000)]
Remove UpdateCompletedType from the debug info emission. We now
emit less than complete types on purpose on occasion and so
our caches aren't useful for this kind of lazy emitting.
Anna Zaks [Fri, 17 Feb 2012 22:35:31 +0000 (22:35 +0000)]
[analyzer] Fix another false positive in the Malloc Checker, by making
it aware of CString APIs that return the input parameter.
Malloc Checker needs to know how the 'strcpy' function is
evaluated. Introduce the dependency on CStringChecker for that.
CStringChecker knows all about these APIs.
Anna Zaks [Fri, 17 Feb 2012 22:35:26 +0000 (22:35 +0000)]
[analyzer] Generalize function name checking in CString checker.
(Ex: It was not treating __inline_strcpy as strcpy. Will add tests that
rely on this later on.)
projects / clang / log