Summary:
Certain implicitly generated coroutine statements, such as the calls to 'return_value()' or `return_void()` or `get_return_object_on_allocation_failure()`, cannot be built until the promise type is no longer dependent. This means they are not built until after the coroutine body statement has been transformed.
This patch fixes an issue where these statements would never be built for coroutine templates.
It also fixes a small issue where diagnostics about `get_return_object_on_allocation_failure()` were incorrectly suppressed.
Reviewers: rsmith, majnemer, GorNishanov, aaron.ballman
Reviewed By: GorNishanov
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D31487
llvm-svn: 299380
- also replace direct equality checks against the ConstantEvaluated enumerator with isConstantEvaluted(), in anticipation of adding finer granularity to the various ConstantEvaluated contexts and reinstating certain restrictions on where lambda expressions can occur in C++17.
- update the clang tablegen backend that uses these Enumerators, and add the relevant scope where needed.
llvm-svn: 299316
Our _MM_HINT_T0/T1 constant values are 3/2 which matches gcc, but not icc or Intel documentation. Interestingly gcc had this same bug on their implementation of the gather/scatter builtins at one point too.
Fixes PR32411.
llvm-svn: 299233
Reasoning behind this change was allowing the function to accept all values
from range [-128, 255] since all of them can be encoded in an 8bit wide
value.
This differs from the prior state where only range [-128, 127] was accepted,
where values were assumed to be signed, whereas now the actual
interpretation of the immediate is deferred to the consumer as required.
Patch by Stefan Maksimovic.
Differential Revision: https://reviews.llvm.org/D31082
llvm-svn: 299229
This patch serializes the state of #pragma pack. It preserves the state of the
pragma from a PCH/from modules in a file that uses that PCH/those modules.
rdar://21359084
Differential Revision: https://reviews.llvm.org/D31241
llvm-svn: 299226
GCC has the alloc_align attribute, which is similar to assume_aligned, except the attribute's parameter is the index of the integer parameter that needs aligning to.
Differential Revision: https://reviews.llvm.org/D29599
llvm-svn: 299117
an ObjC object pointer
When ARC is enabled in Objective-C++, comparisons between a pointer and
Objective-C object pointer typically result in errors like this:
"invalid operands to a binary expression". This error message can be quite
confusing as it doesn't provide a solution to the problem, unlike the non-C++
diagnostic: "implicit conversion of Objective-C pointer type 'id' to C pointer
type 'void *' requires a bridged cast" (it also provides fix-its). This commit
forces comparisons between pointers and Objective-C object pointers in ARC to
use the Objective-C semantic rules to ensure that a better diagnostic is
reported.
rdar://31103857
Differential Revision: https://reviews.llvm.org/D31177
llvm-svn: 299080
FPContractModeKind is the codegen option flag which is already ternary (off,
on, fast). This makes it universally the type for the contractable info
across the front-end:
* In FPOptions (i.e. in the Sema + in the expression nodes).
* In LangOpts::DefaultFPContractMode which is the option that initializes
FPOptions in the Sema.
Another way to look at this change is that before fp-contractable on/off were
the only states handled to the front-end:
* For "on", FMA folding was performed by the front-end
* For "fast", we simply forwarded the flag to TargetOptions to handle it in
LLVM
Now off/on/fast are all exposed because for fast we will generate
fast-math-flags during CodeGen.
This is toward moving fp-contraction=fast from an LLVM TargetOption to a
FastMathFlag in order to fix PR25721.
---
This is a recommit of r299027 with an adjustment to the test
CodeGenCUDA/fp-contract.cu. The test assumed that even
though -ffp-contract=on is passed FE-based folding of FMA won't happen.
This is obviously wrong since the user is asking for this explicitly with the
option. CUDA is different that -ffp-contract=fast is on by default.
The test used to "work" because contract=fast and contract=on were maintained
separately and we didn't fold in the FE because contract=fast was on due to
the target-default. This patch consolidates the contract=on/fast/off state
into a ternary state hence the change in behavior.
---
Differential Revision: https://reviews.llvm.org/D31167
llvm-svn: 299033
FPContractModeKind is the codegen option flag which is already ternary (off,
on, fast). This makes it universally the type for the contractable info
across the front-end:
* In FPOptions (i.e. in the Sema + in the expression nodes).
* In LangOpts::DefaultFPContractMode which is the option that initializes
FPOptions in the Sema.
Another way to look at this change is that before fp-contractable on/off were
the only states handled to the front-end:
* For "on", FMA folding was performed by the front-end
* For "fast", we simply forwarded the flag to TargetOptions to handle it in
LLVM
Now off/on/fast are all exposed because for fast we will generate
fast-math-flags during CodeGen.
This is toward moving fp-contraction=fast from an LLVM TargetOption to a
FastMathFlag in order to fix PR25721.
Differential Revision: https://reviews.llvm.org/D31167
llvm-svn: 299027
Summary: After examining the remaining uses of LangOptions.ObjCAutoRefCount, found a some additional places to also check for ObjCWeak not covered by previous test cases. Added a test file to verify all the code paths that were changed.
Reviewers: rsmith, doug.gregor, rjmccall
Reviewed By: rjmccall
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D31007
llvm-svn: 299015
Summary: clang should produce the same errors Objective-C classes that cannot be assigned to weak pointers under both -fobjc-arc and -fobjc-weak. Check for ObjCWeak along with ObjCAutoRefCount when analyzing pointer conversions. Add an -fobjc-weak pass to the existing arc-unavailable-for-weakref test cases to verify the behavior is the same.
Reviewers: rsmith, doug.gregor, rjmccall
Reviewed By: rjmccall
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D31006
llvm-svn: 299014
Summary: -Warc-repeated-use-of-weak should produce the same warnings with -fobjc-weak as it does with -objc-arc. Also check for ObjCWeak along with ObjCAutoRefCount when recording the use of an evaluated weak variable. Add a -fobjc-weak run to the existing arc-repeated-weak test case and adapt it slightly to work in both modes.
Reviewers: rsmith, doug.gregor, jordan_rose, rjmccall
Reviewed By: rjmccall
Subscribers: arphaman, rjmccall, cfe-commits
Differential Revision: https://reviews.llvm.org/D31005
llvm-svn: 299011
Summary: Similar to ARC, in ObjCWeak Objective-C object pointers qualified with a weak lifetime are not POD or trivial types. Update the type trait code to reflect this. Copy and adapt the arc-type-traits.mm test case to verify correctness.
Reviewers: rsmith, doug.gregor, rjmccall
Reviewed By: rjmccall
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D31004
llvm-svn: 299010
Summary: When adding an Objective-C retainable type member to a C++ class, also check the LangOpts.ObjCWeak flag and the lifetime qualifier so __weak qualified Objective-C pointer members cause the class to be a non-POD type with non-trivial special members, so the compiler always emits the necessary runtime calls for copying, moving, and destroying the weak member. Otherwise, Objective-C++ classes with weak Objective-C pointer members compiled with -fobjc-weak exhibit undefined behavior if the C++ class is classified as a POD type.
Reviewers: rsmith, benlangmuir, doug.gregor, rjmccall
Reviewed By: rjmccall
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D31003
llvm-svn: 299008
If there is an unresolved member access AST node, and the base is
implicit, do not access/use it for generating candidate overloads for
code completion results.
Fixes PR31093.
llvm-svn: 298903
FIXME: ActOnReturnStmt expects a scope that is inside of the function, due
to CheckJumpOutOfSEHFinally(*this, ReturnLoc, *CurScope->getFnParent());
S.getCurScope()->getFnParent() == nullptr at ActOnFinishFunctionBody when
CoroutineBodyStmt is built. Figure it out and fix it.
llvm-svn: 298893
Summary:
If promise_type has get_return_object_on_allocation_failure defined,
check if an allocation function returns nullptr, and if so,
return the result of get_return_object_on_allocation_failure().
Reviewers: rsmith, EricWF
Reviewed By: EricWF
Subscribers: mehdi_amini, cfe-commits
Differential Revision: https://reviews.llvm.org/D31399
llvm-svn: 298891
Summary:
This patch implements parsing of [[clang::suppress(rule, ...)]]
and [[gsl::suppress(rule, ...)]] attributes.
C++ Core Guidelines depend heavily on tool support for
rule enforcement. They also propose a way to suppress
warnings [1] which is by annotating any ancestor in AST
with the C++11 attribute [[gsl::suppress(rule1,...)]].
To have a mechanism to suppress non-C++ Core
Guidelines specific, an additional spelling of [[clang::suppress]]
is defined.
For example, to suppress the warning cppcoreguidelines-slicing,
one could do
```
[[clang::suppress("cppcoreguidelines-slicing")]]
void f() { ... code that does slicing ... }
```
or
```
void g() {
Derived b;
[[clang::suppress("cppcoreguidelines-slicing")]]
Base a{b};
[[clang::suppress("cppcoreguidelines-slicing")]] {
doSomething();
Base a2{b};
}
}
```
This parsing can then be used by clang-tidy, which includes multiple
C++ Core Guidelines rules, to suppress warnings (see
https://reviews.llvm.org/D24888).
For the exact naming of the rule in the attribute, there
are different possibilities, which will be defined in the
corresponding clang-tidy patch.
Currently, clang-tidy supports suppressing of warnings through "//
NOLINT" comments. There are some advantages that the attribute has:
- Suppressing specific warnings instead of all warnings
- Suppressing warnings in a block (namespace, function, compound
statement)
- Code formatting may split a statement into multiple lines,
thus a "// NOLINT" comment may be on the wrong line
I'm looking forward to your comments!
[1] https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md#inforce-enforcement
Reviewers: alexfh, aaron.ballman, rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D24886
llvm-svn: 298880
Sema holds the current FPOptions which is adjusted by 'pragma STDC
FP_CONTRACT'. This then gets propagated into expression nodes as they are
built.
This encapsulates FPOptions so that this propagation happens opaquely rather
than directly with the fp_contractable on/off bit. This allows controlled
transitioning of fp_contractable to a ternary value (off, on, fast). It will
also allow adding more fast-math flags later.
This is toward moving fp-contraction=fast from an LLVM TargetOption to a
FastMathFlag in order to fix PR25721.
Differential Revision: https://reviews.llvm.org/D31166
llvm-svn: 298877
for unused values.
This fixes a regression caused by r298676, where constructor calls to
classes with non-trivial dtor were marked as unused if the first
argument is an initializer list. This is inconsistent (as the test
shows) and also warns on a reasonbly common code pattern where people
just call constructors to create and immediately destroy an object.
llvm-svn: 298853
This change fixes a crash on initialization of a reference from ({}) during
template instantiation and incidentally improves diagnostics.
This reverts a prior attempt to handle this in r286721. Instead, we teach the
initialization code that initialization cannot be performed if a source type
is required and the initializer is an initializer list (which is not an
expression and does not have a type), and likewise for function-style cast
expressions.
llvm-svn: 298676
Correct class-template deprecation behavior
Based on the comment in the test, and my reading of the standard, a deprecated warning should be issued in the following case:
template<typename T> [[deprecated]] class Foo{}; Foo<int> f;
This was not the case, because the ClassTemplateSpecializationDecl creation did not also copy the deprecated attribute.
Note: I did NOT audit the complete set of attributes to see WHICH ones should be copied, so instead I simply copy ONLY the deprecated attribute.
Previous DiffRev: https://reviews.llvm.org/D27486, was reverted.
This patch fixes the issues brought up here by the reverter: https://reviews.llvm.org/rL298410
Differential Revision: https://reviews.llvm.org/D31245
llvm-svn: 298634
that became supported after r297019
The commit r297019 expanded the performSelector ObjC method family heuristic
to ensure that -Wobjc-unsafe-perform-selector covers all performSelector
variations. However, this made the -Warc-performSelector-leaks too noisy, as
that warning produces mostly false positives since the selector is unknown.
This commit reverts the ObjC method family heuristics introduced in r297019.
This ensures that -Warc-performSelector-leaks isn't too noisy. The commit still
preserves the coverage of -Wobjc-unsafe-perform-selector.
rdar://31124629
llvm-svn: 298587
Summary:
This patch adopts the recent changes that renamed `set_exception(exception_pointer)` to `unhandled_exception()`.
Additionally `unhandled_exception()` is now required, and so an error is emitted when exceptions are enabled but the promise type does not provide the member.
When exceptions are disabled a warning is emitted instead of an error, The warning notes that the `unhandled_exception()` function is required when exceptions are enabled.
Reviewers: rsmith, GorNishanov, aaron.ballman, majnemer
Reviewed By: GorNishanov
Subscribers: mehdi_amini, cfe-commits
Differential Revision: https://reviews.llvm.org/D30859
llvm-svn: 298565
We don't know whether some other instantiation of the template might be able to
reach the annotation, so warning on it has a high chance of false positives.
Patch by Ahmed Asadi!
Differential Revision: https://reviews.llvm.org/D31069
llvm-svn: 298477
Based on the comment in the test, and my reading of the standard, a deprecated warning should be issued in the following case:
template<typename T> [[deprecated]] class Foo{}; Foo<int> f;
This was not the case, because the ClassTemplateSpecializationDecl creation did not also copy the deprecated attribute.
Note: I did NOT audit the complete set of attributes to see WHICH ones should be copied, so instead I simply copy ONLY the deprecated attribute.
Differential Revision: https://reviews.llvm.org/D27486
llvm-svn: 298410
Summary: I added a new rank to ImplicitConversionRank enum to resolve the function overload ambiguity with vector types. Rank of scalar types conversion is lower than vector splat. So, we can choose which function should we call. See test for more details.
Reviewers: Anastasia, cfe-commits
Reviewed By: Anastasia
Subscribers: bader, yaxunl
Differential Revision: https://reviews.llvm.org/D30816
llvm-svn: 298366
This commit adds support for a new attribute that will be used to
distinguish between extensible and inextensible enums. There are three
main purposes of this attribute:
1. Give better control over when enum-related warnings are issued.
For example, in the code below, clang will not issue a -Wassign-enum
warning if the enum is marked "open":
enum __attribute__((enum_extensibility(closed))) EnumClosed {
B0 = 1, B1 = 10
};
enum __attribute__((enum_extensibility(open))) EnumOpen {
C0 = 1, C1 = 10
};
enum EnumClosed ec = 100; // warning issued
enum EnumOpen eo = 100; // no warning
2. Enable code-completion and debugging tools to offer better
suggestions.
3. Make it easier for swift's clang importer to determine which swift
type an enum should be mapped to.
For more details, see the discussion I started on cfe-dev:
http://lists.llvm.org/pipermail/cfe-dev/2017-February/052748.html
rdar://problem/12764379
rdar://problem/23145650
Differential Revision: https://reviews.llvm.org/D30766
llvm-svn: 298332
Summary:
3.4.6 [basic.lookup.udir] paragraph 1:
In a using-directive or namespace-alias-definition, during the lookup for a namespace-name or for a name in a nested-name-specifier, only namespace names are considered.
Reviewers: rsmith, aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D30848
llvm-svn: 298126
This enhances the AST to keep track of locations of the names in those ObjC property attributes, and reports them for indexing.
Patch by Nathan Hawes!
https://reviews.llvm.org/D30907
llvm-svn: 297972
The instance method 'self' does not actually return an over-retained object,
so we shouldn't report an error when it's used with 'performSelector'.
rdar://31071620
llvm-svn: 297961
instance of a qualified Class object when that instance method comes from
a protocol that's implemented by NSObject
Instance methods from a root class like NSObject are also class methods because
the metaclass of root class derives from that root class. Therefore, we can
avoid the warning for instances of qualified Class objects that point to classes
that derive from NSObject. Note that we actually don't know if a Class instance
points to a class that derives from NSObject at compile-time, so we have to
make a reasonable assumption that the majority of instances will do so.
rdar://22812517
llvm-svn: 297862
This adds -Wbitfield-enum-conversion, which warns on implicit
conversions that happen on bitfield assignment that change the value of
some enumerators.
Values of enum type typically take on a very small range of values, so
they are frequently stored in bitfields. Unfortunately, there is no
convenient way to calculate the minimum number of bits necessary to
store all possible values at compile time, so users usually hard code a
bitwidth that works today and widen it as necessary to pass basic
testing and validation. This is very error-prone, and leads to stale
widths as enums grow. This warning aims to catch such bugs.
This would have found two real bugs in clang and two instances of
questionable code. See r297680 and r297654 for the full description of
the issues.
This warning is currently disabled by default while we investigate its
usefulness outside of LLVM.
The major cause of false positives with this warning is this kind of
enum:
enum E { W, X, Y, Z, SENTINEL_LAST };
The last enumerator is an invalid value used to validate inputs or size
an array. Depending on the prevalance of this style of enum across a
codebase, this warning may be more or less feasible to deploy. It also
has trouble on sentinel values such as ~0U.
Reviewers: rsmith, rtrieu, thakis
Reviewed By: thakis
Subscribers: hfinkel, voskresensky.vladimir, sashab, cfe-commits
Differential Revision: https://reviews.llvm.org/D30923
llvm-svn: 297761
The only valid values for scale immediate of scatter/gather builtins are 1, 2, 4, or 8. This patch enforces this in the frontend otherwise we generate invalid instruction encodings in the backend.
Differential Revision: https://reviews.llvm.org/D30875
llvm-svn: 297642
Given that we have already explicitly stated in the qualifier that the
expression is __unaligned, it makes little sense to diagnose that the address
of the packed member may not be aligned.
Differential Revision: https://reviews.llvm.org/D30884
llvm-svn: 297620
Summary:
Some coroutine diagnostics need to point to the location of the first coroutine keyword in the function, like when diagnosing a `return` inside a coroutine. Previously we did this by storing each *valid* coroutine statement in a list and select the first one to use in diagnostics. However if every coroutine statement is invalid we would have no location to point to.
This patch fixes the storage of the first coroutine statement location, ensuring that it gets stored even when the resulting AST node would be invalid.
This patch also removes the `CoroutineStmts` list in `FunctionScopeInfo` because it was unused.
Reviewers: rsmith, GorNishanov, aaron.ballman
Reviewed By: GorNishanov
Subscribers: mehdi_amini, cfe-commits
Differential Revision: https://reviews.llvm.org/D30776
llvm-svn: 297547
Summary:
Create only one OpaqueValue for await_ready/await_suspend/await_resume.
Store OpaqueValue used in the CoroutineSuspendExpr node, so that CodeGen does not have to hunt looking for it.
Reviewers: rsmith, EricWF, aaron.ballman
Reviewed By: EricWF
Subscribers: mehdi_amini, cfe-commits
Differential Revision: https://reviews.llvm.org/D30775
llvm-svn: 297541
Removes immediate range checks for these instructions, since they have GPR
rt as their input operand.
Patch by Stefan Maksimovic.
Differential Revision: https://reviews.llvm.org/D30693
llvm-svn: 297485
1. Reimplemented conditional operator so that it checks
compatibility of unqualified pointees of the 2nd and
the 3rd operands (C99, OpenCL v2.0 6.5.15).
Define QualTypes compatibility for OpenCL as following:
- corresponding types are compatible (C99 6.7.3)
- CVR-qualifiers are equal (C99 6.7.3)
- address spaces are equal (implementation defined)
2. Added generic address space to Itanium mangling.
Review: D30037
Patch by Dmitry Borisenkov!
llvm-svn: 297468
Summary:
This patch adds passing a coroutine_handle object to await_suspend calls.
It builds the coroutine_handle using coroutine_handle<PromiseType>::from_address(__builtin_coro_frame()).
(a revision of https://reviews.llvm.org/D26316 that for some reason refuses to apply via arc patch)
Reviewers: GorNishanov
Subscribers: mehdi_amini, cfe-commits, EricWF
Differential Revision: https://reviews.llvm.org/D30769
llvm-svn: 297356
Summary:
The changes contained in this patch are:
1. Defines a new AST node `CoawaitDependentExpr` for representing co_await expressions while the promise type is still dependent.
2. Correctly detect and transform the 'co_await' operand to `p.await_transform(<expr>)` when possible.
3. Change the initial/final suspend points to build during the initial parse, so they have the correct operator co_await lookup results.
4. Fix transformation of the CoroutineBodyStmt so that it doesn't re-build the final/initial suspends.
@rsmith: This change is a little big, but it's not trivial for me to split it up. Please let me know if you would prefer this submitted as multiple patches.
Reviewers: rsmith, GorNishanov
Reviewed By: rsmith
Subscribers: ABataev, rsmith, mehdi_amini, cfe-commits
Differential Revision: https://reviews.llvm.org/D26057
llvm-svn: 297093
Previously when a coroutine was building the implicit setup/destroy
constructs it would emit diagostics about failures on the first co_await/co_return/co_yield
it encountered. This was confusing because that construct may not itself be ill-formed.
This patch moves the diagnostics to the function start instead.
llvm-svn: 297089
that return record or vector types
The performSelector family of methods from Foundation use objc_msgSend to
dispatch the selector invocations to objects. However, method calls to methods
that return record types might have to use the objc_msgSend_stret as the return
value won't find into the register. This is also supported by this sentence from
performSelector documentation: "The method should not have a significant return
value and should take a single argument of type id, or no arguments". This
commit adds a new warning that warns when a selector which corresponds to a
method that returns a record type is passed into performSelector.
rdar://12056271
Differential Revision: https://reviews.llvm.org/D30174
llvm-svn: 297019
Summary:
Functions with the "xray_log_args" attribute will tell LLVM to emit a special
XRay sled for compiler-rt to copy any call arguments to your logging handler.
Reviewers: dberris
Reviewed By: dberris
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D29704
llvm-svn: 296999
and the nature of a declaration
This commit adds an external_source_symbol attribute to Clang. This attribute
specifies that a declaration originates from an external source and describes
the nature of that source. This attribute will be used to improve IDE features
like 'jump-to-definition' for mixed-language projects or project that use
auto-generated code.
rdar://30423368
Differential Revision: https://reviews.llvm.org/D29819
llvm-svn: 296649
Summary:
Don't warn about unused lambda captures that involve copying a
value of a type that cannot be trivially copied and destroyed.
Fixes PR31977
Reviewers: rsmith, aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D30327
llvm-svn: 296602
potential capture list.
Fix Sema::getCurLambda() to return the innermost lambda scope when there
is a block enclosed in the lambda. Previously, the method would return a
nullptr in such cases, which would prevent a variable captured by the
enclosed block to be added to the lambda scope's potential capture list.
rdar://problem/28412462
Differential Revision: https://reviews.llvm.org/D25556
llvm-svn: 296584
The exisiting warning for inconsistent overrides does not include the destructor
as it was noted in review that it was too noisy. Instead, add to a separate
warning group that is off by default for users who want consistent warnings
between methods and destructors.
llvm-svn: 296572
Essentially, as a base class constructor does not construct virtual bases, such
a constructor for an abstract class does not need the corresponding base class
construction to be valid, and likewise for destructors.
This creates an awkward situation: clang will sometimes generate references to
the complete object and deleting destructors for an abstract class (it puts
them in the construction vtable for a derived class). But we can't generate a
"correct" version of these because we can't generate references to base class
constructors any more (if they're template specializations, say, we might not
have instantiated them and can't assume any other TU will emit a copy).
Fortunately, we don't need to, since no correct program can ever invoke them,
so instead emit symbols that just trap.
We should stop emitting references to these symbols, but still need to emit
definitions for compatibility.
llvm-svn: 296275
The goal of this is to fix a bug in modules where we'd merge
FunctionDecls that differed in their pass_object_size attributes. Since
we can overload on the presence of pass_object_size attributes, this
behavior is incorrect.
We don't represent `N` in `pass_object_size(N)` as part of
ExtParameterInfo, since it's an error to overload solely on the value of
N. This means that we have a bug if we have two modules that declare
functions that differ only in their pass_object_size attrs, like so:
// In module A, from a.h
void foo(char *__attribute__((pass_object_size(0))));
// In module B, from b.h
void foo(char *__attribute__((pass_object_size(1))));
// In module C, in main.c
#include "a.h"
#include "b.h"
At the moment, we'll merge the foo decls, when we should instead emit a
diagnostic about an invalid overload. We seem to have similar (silent)
behavior if we overload only on the return type of `foo` instead; I'll
try to find a good place to put a FIXME (or I'll just file a bug) soon.
This patch also fixes a bug where we'd not output the proper extended
parameter info for declarations with pass_object_size attrs.
llvm-svn: 296076
Rather than attempting to compare whether the previous and current top of
context stack are "equal" (which fails for a number of reasons, such as the
context stack entries containing pointers to objects on the stack, or reaching
the same "top of stack" entry through two different paths), track the depth of
context stack at which we last emitted a note and invalidate it when we pop the
context stack to less than that depth.
This causes us to emit some missing "in instantiation of" notes and to stop
emitting redundant "in instantiation of" stacks matching the previous stack in
rare cases.
llvm-svn: 295921
A 'decltype(auto)' parameter can match any other kind of non-type template
parameter, so should be usable in place of any other parameter in a template
template argument. The standard is sadly extremely unclear on how this is
supposed to work, but this seems like the obviously-correct result.
It's less clear whether an 'auto' parameter should be able to match
'decltype(auto)', since the former cannot be used if the latter turns out to be
used for a reference type, but if we disallow that then consistency suggests we
should also disallow 'auto' matching 'T' for the same reason, defeating
intended use cases of the feature.
llvm-svn: 295866
checkNestingOfRegions uses CancelRegion to determine whether cancel and
cancellation point are valid in the given nesting. This leads to unuseful
diagnostics if CancelRegion is invalid. The given test case has produced:
region cannot be closely nested inside 'parallel' region
As a solution, introduce checkCancelRegion and call it first to get the
expected error:
one of 'for', 'parallel', 'sections' or 'taskgroup' is expected
Differential Revision: https://reviews.llvm.org/D30135
llvm-svn: 295808
We need to look through the PackExpansionType in the parameter type when
deducing, and we need to consider the possibility of deducing arguments for
packs that are not lexically mentioned in the pattern (but are nonetheless
deducible) when figuring out which packs are covered by a pack deduction scope.
llvm-svn: 295790
case where the class template has a parameter pack.
Checking of the template arguments expects an "as-written" template argument
list, which in particular does not have any parameter packs. So flatten the
packs into separate arguments before passing them in.
llvm-svn: 295710
template deduction guides for class template argument deduction.
Ensure that we have a local instantiation scope for tracking the instantiated
parameters. Additionally, unusually, we're substituting at depth 1 and leaving
depth 0 alone; make sure that we don't reduce template parameter depth by 2 for
inner parameters in the process. (This is probably also broken for alias
templates in the case where they're expanded within a dependent context, but
this patch doesn't fix that.)
llvm-svn: 295696
instantiation.
In preparation for converting the template stack to a more general context
stack (so we can include context notes for other kinds of context).
llvm-svn: 295686
Using the constructed name for the class properties with dot syntax may
yield an inappropriate selector (i.e. if it is specified via property
attributes). Prefer the declaration for the selector, falling back to
the constructed name otherwise.
Patch by David Herzka!
llvm-svn: 295683
These attributes effectively turn a non-defining declaration into a
definition, so the case when the declaration already has a body must
be diagnosed properly.
Differential Revision: https://reviews.llvm.org/D30032
llvm-svn: 295541
Reserve a spot for ODR hash in CXXRecordDecl and in its modules storage.
Default the hash value to 0 for all classes.
Differential Revision: https://reviews.llvm.org/D21675
llvm-svn: 295533
https://reviews.llvm.org/D29922
This patch adds two fields for use in the implementation of 'distribute parallel for':
The increment expression for the distribute loop. As the chunk assigned to a team is executed by multiple threads within the 'parallel for' region, the increment expression has to correspond to the value returned by the related runtime call (for_static_init).
The upper bound of the innermost loop ('for' in 'distribute parallel for') is not the globalUB expression normally used for pragma 'for' when found in isolation. It is instead the upper bound of the chunk assigned to the team ('distribute' loop). In this way, we prevent teams from executing chunks assigned to other teams.
The use of these two fields can be see in a related explanatory patch:
https://reviews.llvm.org/D29508
llvm-svn: 295497
A slightly weaker form of ODR checking than previous attempts, but hopefully
won't break the modules build bot. Future work will be needed to catch all
cases.
When objects are imported for modules, there is a chance that a name collision
will cause an ODR violation. Previously, only a small number of such
violations were detected. This patch provides a stronger check based on
AST nodes.
The information needed to uniquely identify an object is taken from the AST and
put into a one-dimensional byte stream. This stream is then hashed to give
a value to represent the object, which is stored with the other object data
in the module.
When modules are loaded, and Decl's are merged, the hash values of the two
Decl's are compared. Only Decl's with matched hash values will be merged.
Mismatch hashes will generate a module error, and if possible, point to the
first difference between the two objects.
The transform from AST to byte stream is a modified depth first algorithm.
Due to references between some AST nodes, a pure depth first algorithm could
generate loops. For Stmt nodes, a straight depth first processing occurs.
For Type and Decl nodes, they are replaced with an index number and only on
first visit will these nodes be processed. As an optimization, boolean
values are saved and stored together in reverse order at the end of the
byte stream to lower the ammount of data that needs to be hashed.
Compile time impact was measured at 1.5-2.0% during module building, and
negligible during builds without module building.
Differential Revision: https://reviews.llvm.org/D21675
llvm-svn: 295421
Removed ndrange_t as Clang builtin type and added
as a struct type in the OpenCL header.
Use type name to do the Sema checking in enqueue_kernel
and modify IR generation accordingly.
Review: D28058
Patch by Dmitry Borisenkov!
llvm-svn: 295311
Recommit r293585 that was reverted in r293611 with new fixes. The previous
issue was determined to be an overly aggressive AST visitor from forward
declared objects. The visitor will now only deeply visit certain Decl's and
only do a shallow information extraction from all other Decl's.
When objects are imported for modules, there is a chance that a name collision
will cause an ODR violation. Previously, only a small number of such
violations were detected. This patch provides a stronger check based on
AST nodes.
The information needed to uniquely identify an object is taken from the AST and
put into a one-dimensional byte stream. This stream is then hashed to give
a value to represent the object, which is stored with the other object data
in the module.
When modules are loaded, and Decl's are merged, the hash values of the two
Decl's are compared. Only Decl's with matched hash values will be merged.
Mismatch hashes will generate a module error, and if possible, point to the
first difference between the two objects.
The transform from AST to byte stream is a modified depth first algorithm.
Due to references between some AST nodes, a pure depth first algorithm could
generate loops. For Stmt nodes, a straight depth first processing occurs.
For Type and Decl nodes, they are replaced with an index number and only on
first visit will these nodes be processed. As an optimization, boolean
values are saved and stored together in reverse order at the end of the
byte stream to lower the ammount of data that needs to be hashed.
Compile time impact was measured at 1.5-2.0% during module building, and
negligible during builds without module building.
Differential Revision: https://reviews.llvm.org/D21675
llvm-svn: 295284
This appears to be the only template argument deduction context where we were
missing this check. Surprisingly, other implementations also appear to miss
the check in this case; it may turn out that important code is relying on
the widespread non-conformance here, in which case we'll need to reconsider.
llvm-svn: 295277
Eric Fiselier