David Blaikie [Sat, 5 Apr 2014 07:23:17 +0000 (07:23 +0000)]
DebugInfo: Avoid emitting constnants for every use
While the folding set would deduplicate the nodes themselves and LLVM
would handle not emitting the same global twice, it still meant creating
a long/redundant list of global variables.
Richard Trieu [Sat, 5 Apr 2014 05:17:01 +0000 (05:17 +0000)]
Add a new subgroup to -Wtautological-compare, -Wtautological-overlap-compare,
which warns on compound conditionals that always evaluate to the same value.
For instance, (x > 5 && x < 3) will always be false since no value for x can
satisfy both conditions.
This patch also changes the CFG to use these tautological values for better
branch analysis. The test for -Wunreachable-code shows how this change catches
additional dead code.
Jordan Rose [Sat, 5 Apr 2014 02:01:41 +0000 (02:01 +0000)]
[analyzer] Look through temporary destructors when finding a region to construct.
Fixes a false positive when temporary destructors are enabled where a temporary
is destroyed after a variable is constructed but before the VarDecl itself is
processed, which occurs when the variable is in the condition of an if or while.
Patch by Alex McCarthy, with an extra test from me.
Objective-C arc [Sema]. Allow bridge cast of a qualified-id expression
when bridged Objective-C type conforms to the protocols in CF types's
Objective-C class. // rdar://16393330
Try harder about not suggesting methods as corrections when they
obviously won't work. Specifically, don't suggest methods (static or
not) from unrelated classes when the expression is a method call
through a specific object.
David Blaikie [Fri, 4 Apr 2014 20:56:17 +0000 (20:56 +0000)]
DebugInfo: PR19298: function local const variables duplicated in the root scope
See the comment for CodeGenFunction::tryEmitAsConstant that describes
how in some contexts (lambdas) we must not emit references to the
variable, but instead use the constant directly - because of this we end
up emitting a constant for the variable, as well as emitting the
variable itself.
Should we just skip putting the variable on the stack at all and omit
the debug info for the constant? It's not clear to me - what if the
address of the local is taken?
This introduces the definitions needed for the Windows on ARM target. Add
target definitions for both the MSVC environment and the MSVC + Itanium C++ ABI
environment. The Visual Studio definitions correspond to the definitions
provided by Visual Studio 2012.
In preparation for being able to use simple Boolean logic expressions involving capabilities, the semantics for attributes now looks through the types of the constituent parts of a capability expression instead of at the aggregate expression type.
Visual Studio is the Integrated Development Environment. The toolchain is
generally referred to MSVC. Rename the target information to be more precise as
per the recommendation of Reid Kleckner.
CodeGen: Don't create branch weight metadata from empty profiles
If all of our weights are zero when calculating branch weights, it
means we haven't profiled the code in question. Avoid creating a
metadata node that says all branches are equally likely in this case.
The test also checks constructs that hit the other createBranchWeights
overload. These were already working.
Add clang-cl alias to allow users to disable c4005
If we ever want three or more aliases, at that point we should put MSVC
warning ids in DiagnosticGroups.td. We can use that to support #pragma
warning.
vector [Sema]. Check for proper use of 's' char prefix
(which indicates vector expression is a string of hex
values) instead of crashing in code gen. // rdar://16492792
By ignoring this pragma with a warning, we're essentially miscompiling
the user's program. WebKit / Blink use this pragma to disable dynamic
initialization and finalization of some static data, and running the
dtors crashes the program.
Error out for now, so that /fallback compiles the TU correctly with
MSVC. This pragma should be implemented some time this month, and we
can remove this hack.
Teach getTemplateInstantiationPattern to deal with generic lambdas.
No functionality change.
When determining the pattern for instantiating a generic lambda call operator specialization - we must not go drilling down for the 'prototype' (i.e. as written) pattern - rather we must use our partially transformed pattern (whose DeclRefExprs are wired correctly to any enclosing lambda's decls that should be mapped correctly in a local instantiation scope) that is the templated pattern of the specialization's primary template (even though the primary template might be instantiated from a 'prototype' member-template). Previously, the drilling down was haltted by marking the instantiated-from primary template as a member-specialization (incorrectly).
This prompted Richard to remark (http://llvm-reviews.chandlerc.com/D1784?id=4687#inline-10272)
"It's a bit nasty to (essentially) set this bit incorrectly. Can you put the check into getTemplateInstantiationPattern instead?"
In my reckless youth, I chose to ignore that comment. With the passage of time, I have come to learn the value of bowing to the will of the angry Gods ;)
David Majnemer [Wed, 2 Apr 2014 23:17:29 +0000 (23:17 +0000)]
CodeGen: Emit some functions as weak_odr under -fms-compatibility
Summary:
MSVC always emits inline functions marked with the extern storage class
specifier. The result is something similar to the opposite of
__attribute__((gnu_inline)).
Richard Smith [Wed, 2 Apr 2014 18:28:36 +0000 (18:28 +0000)]
PR19305: Don't issue -Wunused-variable warnings on variable templates. It's not
meaningful to odr-use the VarDecl inside a variable template. (Separately, it'd
be nice to track referenced-ness for templates, and warn on unused ones, but
that's really a distinct issue...)
Move a test that generates and tests a warning-suppressing error out to its own
test file, so it doesn't have weird effects on the other tests in the same file.
David Blaikie [Wed, 2 Apr 2014 18:21:09 +0000 (18:21 +0000)]
DebugInfo: Include default template arguments in template type names
This was committed 4 years ago in 108916 with insufficient testing to
explain why the "getTypeAsWritten" case was appropriate. Experience says
that it isn't - the presence or absence of an explicit instantiation
declaration was causing this code to generate either i<int> or i<int,
int>.
That didn't seem to be a useful distinction, and omitting the template
arguments was destructive to debuggers being able to associate the two
types across translation units or across compilers (GCC, reasonably,
never omitted the arguments).
Summary:
Add support for named values in the parser.
This allows injection of arbitrary constants using a custom Sema object.
Completions are not supported right now.
Will be used by clang_query to support the 'let' command.
Usage example:
clang_query> let unique_ptr recordDecl(hasName("unique_ptr"))
clang_query> match varDecl(hasType(unique_ptr))
David Blaikie [Wed, 2 Apr 2014 05:58:29 +0000 (05:58 +0000)]
Render anonymous entities as '(anonymous <thing>)' (and lambdas as '(lambda at ... )')
For namespaces, this is consistent with mangling and GCC's debug info
behavior. For structs, GCC uses <anonymous struct> but we prefer
consistency between all anonymous entities but don't want to confuse
them with template arguments, etc, so we'll just go with parens in all
cases.
David Blaikie [Wed, 2 Apr 2014 05:48:29 +0000 (05:48 +0000)]
Add the location of Decls to ast dump.
While investigating some debug info issues, Eric and I came across a
particular template case where the location of a decl was quite
different from the range of the same decl. It might've been rather
helpful if the dumper had actually showed us this.
We don't want to encourage the code to emit a lexical block for
a function that needs one in order for the line table to change,
we need to grab the line information from the body of the pattern
that we were instantiated from, this code should do that.
Modify the test case to ensure that we're still looking in the
right place for all of the scopes and also that we haven't
created a lexical block where we didn't need one.
Updating the capability attribute diagnostics to be more capability-neutral. Instead of using terminology such as "lock", "unlock" and "locked", the new terminology is "acquire", "release" and "held". Additionally, the capability attribute's name argument is now reported as part of the diagnostic, instead of hard coding as "mutex."
Jordan Rose [Tue, 1 Apr 2014 16:39:59 +0000 (16:39 +0000)]
[analyzer] Remove incorrect workaround for unimplemented temporary destructors.
If we're trying to get the zero element region of something that's not a region,
we should be returning UnknownVal, which is what ProgramState::getLValue will
do for us.
David Majnemer [Tue, 1 Apr 2014 05:29:46 +0000 (05:29 +0000)]
MS ABI: Support mangling of return-types deducing to local types
The MS ABI forces us into catch-22 when it comes to functions which
return types which are local:
- A function is mangled with it's return type.
- A type is mangled with it's surrounding context.
Avoid this by mangling auto and decltype(autp) directly into the
function's return type. Using this mangling has the double advantage of
being compatible with the C++ standard without crashing the compiler.
N.B. For the curious, the MSVC mangling leads to collisions amongst
template functions and either crashes when faced with local types or is
otherwise incapable of returning them.
Bob Wilson [Tue, 1 Apr 2014 01:38:16 +0000 (01:38 +0000)]
Disable this-return optimizations when targeting iOS 5 and earlier.
Clang implements the part of the ARM ABI saying that certain functions
(e.g., constructors and destructors) return "this", but Apple's version of
gcc and llvm-gcc did not. The libstdc++ dylib on iOS 5 was built with
llvm-gcc, which means that clang cannot safely assume that code from the C++
runtime will correctly follow the ABI. It is also possible to run into this
problem when linking with other libraries built with gcc or llvm-gcc. Even
though there is no way to reliably detect that situation, it is most likely
to come up when targeting older versions of iOS. Disabling the optimization
for any code targeting iOS 5 solves the libstdc++ problem and has a reasonably
good chance of fixing the issue for other older libraries as well.
<rdar://problem/16377159>
David Blaikie [Tue, 1 Apr 2014 00:27:28 +0000 (00:27 +0000)]
Fix up compression related test cases
Fallout from r205261, ensure it doesn't matter how we disable compressed
debug info, even if zlib is missing and that we warn when we don't have
zlib and don't warn when we do, all while silently suppressing these
tests on the systems they weren't intended for...
David Blaikie [Mon, 31 Mar 2014 23:13:30 +0000 (23:13 +0000)]
DebugInfo compression: Enable compression before any sections are created.
For those playing at home this produced some fairly subtle behavior. The
sections created in InitMCObjectFileInfo were created without compressed
debug info (a mistake, but not necessarily /broken). Since these
sections were almost always referenced by the existing MCSection object,
this almost worked.
This got weird when we got to handling the relocations for a section.
See ELFObjectWriter::WriteSection where we compute the true section for
a relocation section by simply stripping the ".rela" prefix and then
looking up that section - doing so hit the compression codepath, looked
up .zdebug_blah and found a newly constructed empty section... thus,
things got weird.
This is untestable without a cross-project test (let me know if people
would prefer that to no testing).
Aaron Ballman [Mon, 31 Mar 2014 18:18:43 +0000 (18:18 +0000)]
Unify __declspec attribute argument parsing with the common attribute argument parsing code.
This removes a diagnostic that is no longer required (the semantic engine now properly handles attribute syntax so __declspec and __attribute__ spellings no longer get mismatched). This caused several testcases to need updating for a slightly different wording.
Aaron Ballman [Mon, 31 Mar 2014 17:32:39 +0000 (17:32 +0000)]
Introduced an attribute syntax-neutral method for parsing attribute arguments that is currently being used by GNU and C++-style attributes. This allows C++11 attributes with argument lists to be handled properly, fixing the "deprecated", "type_visibility", and capability-related attributes with arguments.
David Majnemer [Mon, 31 Mar 2014 16:12:47 +0000 (16:12 +0000)]
MS ABI: Use the proper type for inalloca args
Summary:
The definition of a type later in a translation unit may change it's
type from {}* to (%struct.foo*)*. Earlier function definitions may use
the former while more recent definitions might use the later. This is
fine until they interact with one another (like one calling the other).
In these cases, a bitcast is needed because the inalloca must match the
function call but the store to the lvalue which initializes the argument
slot has to match the rvalue's type.
This technique is along the same lines with what the other,
non-inalloca, codepaths perform.
Tim Northover [Mon, 31 Mar 2014 15:47:09 +0000 (15:47 +0000)]
ARM64: enable aarch64-neon-intrinsics.c test
This adds support for the various NEON intrinsics used by
aarch64-neon-intrinsics.c (originally written for AArch64) and enables the
test.
My implementations are designed to be semantically correct, the actual code
quality looks like its a wash between the two backends, and is frequently
different (hence the large number of CHECK changes).
Aaron Ballman [Mon, 31 Mar 2014 13:14:44 +0000 (13:14 +0000)]
Reapplying r204952 a second time.
Clean up the __has_attribute implementation without modifying its behavior.
Replaces the tablegen-driven AttrSpellings.inc, which lived in the lexing layer with AttrHasAttributeImpl.inc, which lives in the basic layer. Updates the preprocessor to call through to this new functionality which can take additional information into account (such as scopes and syntaxes).
Expose the ability for parts of the compiler to ask whether an attribute is supported for a given spelling (including scope), syntax, triple and language options.
projects / clang / log