summaryrefslogtreecommitdiffstats
path: root/clang/lib/Sema
Commit message (Collapse)AuthorAgeFilesLines
...
* Revert r332470 (and corresponding tests in r332492).Richard Smith2018-05-184-146/+13
| | | | | | | This regressed our support for __attribute__((section)). See added test file for example of code broken by this. llvm-svn: 332760
* [CUDA] Allow "extern __shared__ Foo foo[]" within anon. namespaces.Justin Lebar2018-05-171-0/+5
| | | | | | | | | | | | | | | Summary: Previously this triggered a -Wundefined-internal warning. But it's not an undefined variable -- any variable of this form is a pointer to the base of GPU core's shared memory. Reviewers: tra Subscribers: sanjoy, rsmith Differential Revision: https://reviews.llvm.org/D46782 llvm-svn: 332621
* Fix __uuidof handling on non-type template parameter in C++17Nico Weber2018-05-171-1/+1
| | | | | | | | | | | | | | | Clang used to pass the base lvalue of a non-type template parameter to the template instantiation phase when the base part is __uuidof and it's running in C++17 mode. However, that drops its LValuePath, and unintentionally transforms &__uuidof(...) to __uuidof(...). This CL fixes that by passing whole expr. Fixes PR24986. https://reviews.llvm.org/D46820?id=146557 Patch from Taiju Tsuiki <tzik@chromium.org>! llvm-svn: 332614
* [Sema] Fix assertion when constructor is disabled with partially specialized ↵Volodymyr Sapsai2018-05-161-102/+112
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | template. The added test case was triggering assertion > Assertion failed: (!SpecializedTemplate.is<SpecializedPartialSpecialization*>() && "Already set to a class template partial specialization!"), function setInstantiationOf, file clang/include/clang/AST/DeclTemplate.h, line 1825. It was happening with ClassTemplateSpecializationDecl `enable_if_not_same<int, int>`. Because this template is specialized for equal types not to have a definition, it wasn't instantiated and its specialization kind remained TSK_Undeclared. And because it was implicit instantiation, we didn't mark the decl as invalid. So when we try to find the best matching partial specialization the second time, we hit the assertion as partial specialization is already set. Fix by reusing stored partial specialization when available, instead of looking for the best match every time. rdar://problem/39524996 Reviewers: rsmith, arphaman Reviewed By: rsmith Subscribers: cfe-commits Differential Revision: https://reviews.llvm.org/D46909 llvm-svn: 332509
* [Modules] Do not diagnose missing import in recovery mode if there isn't a ↵Bruno Cardoso Lopes2018-05-161-1/+1
| | | | | | | | | | | | | | | | | decl to lookup Clang often tries to create implicit module import for error recovery, which does a great job helping out with diagnostics. However, sometimes clang does not have enough information given that it's using an invalid context to move on. Be more strict in those cases to avoid crashes. We hit crash on invalids because of this but unfortunately there are no testcases and I couldn't manage to create one. The crashtrace however indicates pretty clear why it's happening. rdar://problem/39313933 llvm-svn: 332491
* Add support for __declspec(code_seg("segname"))Erich Keane2018-05-164-13/+146
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Add support for __declspec(code_seg("segname")) This patch is built on the existing support for #pragma code_seg. The code_seg declspec is allowed on functions and classes. The attribute enables the placement of code into separate named segments, including compiler-generated members and template instantiations. For more information, please see the following: https://msdn.microsoft.com/en-us/library/dn636922.aspx A new CodeSeg attribute is used instead of adding a new spelling to the existing Section attribute since they don’t apply to the same Subjects. Section attributes are also added for the code_seg declspec since they are used for #pragma code_seg. No CodeSeg attributes are added to the AST. The patch is written to match with the Microsoft compiler’s behavior even where that behavior is a little complicated (see https://reviews.llvm.org/D22931, the Microsoft feedback page is no longer available since MS has removed the page). That code is in getImplicitSectionAttrFromClass routine. Diagnostics messages are added to match with the Microsoft compiler for code-seg attribute mismatches on base and derived classes and virtual overrides. Differential Revision: https://reviews.llvm.org/D43352 llvm-svn: 332470
* [CodeComplete] Expose helpers to get RawComment of completion result.Ilya Biryukov2018-05-161-30/+62
| | | | | | | | | | | | | | Summary: Used in clangd, see D45999. Reviewers: sammccall, hokein, ioeric, arphaman Reviewed By: sammccall Subscribers: arphaman, cfe-commits Differential Revision: https://reviews.llvm.org/D46001 llvm-svn: 332457
* Don't produce a redundant "auto type is incompatible with C++98" on every ↵Richard Smith2018-05-151-1/+3
| | | | | | | | lambda with no explicit return type. We already warned about the lambda, and we don't have a source location for the imagined "auto" anyway. llvm-svn: 332401
* Address post-commit review comments after r328731. NFC.Akira Hatanaka2018-05-152-84/+78
| | | | | | | | | | | | | | | | | | | | | | - Define a function (canPassInRegisters) that determines whether a record can be passed in registers based on language rules and target-specific ABI rules. - Set flag RecordDecl::ParamDestroyedInCallee to true in MSVC mode and remove ASTContext::isParamDestroyedInCallee, which is no longer needed. - Use the same type (unsigned) for RecordDecl's bit-field members. For more background, see the following discussions that took place on cfe-commits. http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180326/223498.html http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180402/223688.html http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180409/224754.html http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180423/226494.html http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180507/227647.html llvm-svn: 332397
* Enable control flow pruning of float overflow warnings.Richard Trieu2018-05-141-1/+2
| | | | | | | | | Like other conversion warnings, allow float overflow warnings to be disabled in known dead paths of template instantiation. This often occurs when a template template type is a numeric type and the template will check the range of the numeric type before performing the conversion. llvm-svn: 332310
* [c++17] Fix assertion on synthesizing deduction guides after a fatal error.Volodymyr Sapsai2018-05-141-0/+2
| | | | | | | | | | | | | | | | | | | | | | | After a fatal error Sema::InstantiatingTemplate doesn't allow further instantiation and doesn't push a CodeSynthesisContext. When we tried to synthesize implicit deduction guides from constructors we hit the assertion > Assertion failed: (!CodeSynthesisContexts.empty() && "Cannot perform an instantiation without some context on the " "instantiation stack"), function SubstType, file clang/lib/Sema/SemaTemplateInstantiate.cpp, line 1580. Fix by avoiding deduction guide synthesis if InstantiatingTemplate is invalid. rdar://problem/39051732 Reviewers: rsmith Reviewed By: rsmith Subscribers: cfe-commits Differential Revision: https://reviews.llvm.org/D46446 llvm-svn: 332307
* Fix regression in r332076.Richard Smith2018-05-141-1/+11
| | | | | | | | | If the name after 'template' is an unresolved using declaration (not containing 'typename'), then we don't yet know if it's a valid template-name, so don't reject it prior to instantiation. Instead, treat it as naming a dependent member of the current instantiation. llvm-svn: 332291
* Remove the fixit for the diagnostics regarding capturing autoreleasing ↵George Karpenkov2018-05-141-24/+0
| | | | | | | | | | | | | | | variables in a block The fixit is actively harmful, as it encourages developers to ignore the warning and to write unsafe code. It is almost impossible to write safe code while capturing autoreleasing variables in the block, as in order to check that the block is never called in the autoreleasing pool the developer has to check the transitive closure of all potential callers of the block. Differential Revision: https://reviews.llvm.org/D46778 llvm-svn: 332288
* PR37450: Fix bug that disabled some type checks for variables with deduced ↵Richard Smith2018-05-142-7/+21
| | | | | | | | types. Also improve diagnostic for the case where a type is non-literal because it's a lambda. llvm-svn: 332286
* [AST] Print correct tag decl for tag specifierJoel E. Denny2018-05-141-6/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For example, given: void fn() { struct T *p0; struct T { int i; } *p1; } -ast-print produced: void fn() { struct T { int i; } *p0; struct T { int i; } *p1; } Compiling that fails with a redefinition error. Given: void fn() { struct T *p0; struct __attribute__((deprecated)) T *p1; } -ast-print dropped the attribute. Details: For a tag specifier (that is, struct/union/class/enum used as a type specifier in a declaration) that was also a tag declaration (that is, first occurrence of the tag) or tag redeclaration (that is, later occurrence that specifies attributes or a member list), clang printed the tag specifier as either (1) the full tag definition if one existed, or (2) the first tag declaration otherwise. Redefinition errors were sometimes introduced, as in the first example above. Even when that was impossible because no member list was ever specified, attributes were sometimes lost, thus changing semantics and diagnostics, as in the second example above. This patch fixes a major culprit for these problems. It does so by creating an ElaboratedType with a new OwnedDecl member wherever an occurrence of a tag type is a (re)declaration of that tag type. PrintingPolicy's IncludeTagDefinition used to trigger printing of the member list, attributes, etc. for a tag specifier by using a tag (re)declaration selected as described above. Now, it triggers the same thing except it uses the tag (re)declaration stored in the OwnedDecl. Of course, other tooling can now make use of the new OwnedDecl as well. Also, to be more faithful to the original source, this patch suppresses printing of attributes inherited from previous declarations. Reviewed by: rsmith, aaron.ballman Differential Revision: https://reviews.llvm.org/D45463 llvm-svn: 332281
* [AST] Fix -ast-print for _Bool when have diagnosticsJoel E. Denny2018-05-141-3/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For example, given: #define bool _Bool _Bool i; void fn() { 1; } -ast-print produced: tmp.c:3:13: warning: expression result unused void fn() { 1; } ^ bool i; void fn() { 1; } That fails to compile because bool is undefined. Details: Diagnostics print _Bool as bool when the latter is defined as the former. However, diagnostics were altering the printing policy for -ast-print as well. The printed source was then invalid because the preprocessor eats the bool definition. Problematic diagnostics included suppressed warnings (e.g., add -Wno-unused-value to the above example), including those that are suppressed by default. This patch fixes this bug and cleans up some related comments. Reviewed by: aaron.ballman, rsmith Differential Revision: https://reviews.llvm.org/D45093 llvm-svn: 332275
* [CodeComplete] Provide completion in decls even for incomplete typesIlya Biryukov2018-05-141-4/+2
| | | | | | | | | | | | | | | | | | | | Summary: This change fixes lack of completions in the following case ('^'designates completion points) : void f(^); struct Incomplete; Incomplete g(^); Reviewers: bkramer, aaron.ballman, sammccall Reviewed By: aaron.ballman Subscribers: cfe-commits Differential Revision: https://reviews.llvm.org/D46639 llvm-svn: 332244
* Added atomic_fetch_min, max, umin, umax intrinsics to clang.Elena Demikhovsky2018-05-131-2/+18
| | | | | | | | | These intrinsics work exactly as all other atomic_fetch_* intrinsics and allow to create *atomicrmw* with ordering. Updated the clang-extensions document. Differential Revision: https://reviews.llvm.org/D46386 llvm-svn: 332193
* [Hexagon] Implement checking arguments of builtin callsKrzysztof Parzyszek2018-05-111-0/+223
| | | | llvm-svn: 332105
* Improve diagnostics and error recovery for template name lookup.Richard Smith2018-05-114-80/+127
| | | | | | | | | | | | | For 'x::template y', consistently give a "no member named 'y' in 'x'" diagnostic if there is no such member, and give a 'template keyword not followed by a template' name error if there is such a member but it's not a template. In the latter case, add a note pointing at the non-template. Don't suggest inserting a 'template' keyword in 'X::Y<' if X is dependent if the lookup of X::Y was actually not a dependent lookup and found only non-templates. llvm-svn: 332076
* Don't propagate dllimport to base class template static data membersReid Kleckner2018-05-111-0/+17
| | | | | | MSVC doesn't, so we shouldn't. Fixes PR37232. llvm-svn: 332074
* Allow dllimport non-type template arguments in C++17Reid Kleckner2018-05-101-3/+4
| | | | | | | | | | | Summary: Fixes PR35772. Reviewers: rsmith Differential Revision: https://reviews.llvm.org/D43320 llvm-svn: 332018
* [Clang] Implement function attribute no_stack_protector.Manoj Gupta2018-05-091-0/+4
| | | | | | | | | | | | | | | | | | | | | | Summary: This attribute tells clang to skip this function from stack protector when -stack-protector option is passed. GCC option for this is: __attribute__((__optimize__("no-stack-protector"))) and the equivalent clang syntax would be: __attribute__((no_stack_protector)) This is used in Linux kernel to selectively disable stack protector in certain functions. Reviewers: aaron.ballman, rsmith, rnk, probinson Reviewed By: aaron.ballman Subscribers: probinson, srhines, cfe-commits Differential Revision: https://reviews.llvm.org/D46300 llvm-svn: 331925
* [OpenCL] Add constant address space to __func__ in AST.Anastasia Stulova2018-05-091-13/+10
| | | | | | | | | | | | Added string literal helper function to obtain the type attributed by a constant address space. Also fixed predefind __func__ expr to use the helper to constract the string literal correctly. Differential Revision: https://reviews.llvm.org/D46049 llvm-svn: 331877
* [OpenCL] Restrict various keywords in OpenCL C++ modeSven van Haastregt2018-05-093-0/+26
| | | | | | | | | | | | | | | | | | | | | | | Restrict the following keywords in the OpenCL C++ language mode, according to Sections 2.2 & 2.9 of the OpenCL C++ 1.0 Specification. - dynamic_cast - typeid - register (already restricted in OpenCL C, update the diagnostic) - thread_local - exceptions (try/catch/throw) - access qualifiers read_only, write_only, read_write Support the `__global`, `__local`, `__constant`, `__private`, and `__generic` keywords in OpenCL C++. Leave the unprefixed address space qualifiers such as global available, i.e., do not mark them as reserved keywords in OpenCL C++. libclcxx provides explicit address space pointer classes such as `global_ptr` and `global<T>` that are implemented using the `__`-prefixed qualifiers. Differential Revision: https://reviews.llvm.org/D46022 llvm-svn: 331874
* Remove \brief commands from doxygen comments.Adrian Prantl2018-05-0939-1058/+1058
| | | | | | | | | | | | | | | | | | | This is similar to the LLVM change https://reviews.llvm.org/D46290. We've been running doxygen with the autobrief option for a couple of years now. This makes the \brief markers into our comments redundant. Since they are a visual distraction and we don't want to encourage more \brief markers in new code either, this patch removes them all. Patch produced by for i in $(git grep -l '\@brief'); do perl -pi -e 's/\@brief //g' $i & done for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done Differential Revision: https://reviews.llvm.org/D46320 llvm-svn: 331834
* Fix float->int conversion warnings when near barriers.Erich Keane2018-05-081-20/+12
| | | | | | | | | | | | As Eli brought up here: https://reviews.llvm.org/D46535 I'd previously messed up this fix by missing conversions that are just slightly outside the range. This patch fixes this by no longer ignoring the return value of convertToInteger. Additionally, one of the error messages wasn't very sensical (mentioning out of range value, when it really was not), so it was cleaned up as well. llvm-svn: 331812
* [C++2a] operator<=>: Fix incorrect use of Twine.Eric Fiselier2018-05-081-3/+3
| | | | llvm-svn: 331713
* [C++2a] Implement operator<=>: Address bugs and post-commit review comments ↵Eric Fiselier2018-05-082-6/+8
| | | | | | | | | | | | after r331677. This patch addresses some mostly trivial post-commit review comments received on r331677. Additionally, this patch fixes an assertion in `getNarrowingKind` caused by the use of an uninitialized value from `checkThreeWayNarrowingConversion`. llvm-svn: 331707
* [C++2a] Implement operator<=> CodeGen and ExprConstantEric Fiselier2018-05-074-67/+463
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Summary: This patch tackles long hanging fruit for the builtin operator<=> expressions. It is currently needs some cleanup before landing, but I want to get some initial feedback. The main changes are: * Lookup, build, and store the required standard library types and expressions in `ASTContext`. By storing them in ASTContext we don't need to store (and duplicate) the required expressions in the BinaryOperator AST nodes. * Implement [expr.spaceship] checking, including diagnosing narrowing conversions. * Implement `ExprConstant` for builtin spaceship operators. * Implement builitin operator<=> support in `CodeGenAgg`. Initially I emitted the required comparisons using `ScalarExprEmitter::VisitBinaryOperator`, but this caused the operand expressions to be emitted once for every required cmp. * Implement [builtin.over] with modifications to support the intent of P0946R0. See the note on `BuiltinOperatorOverloadBuilder::addThreeWayArithmeticOverloads` for more information about the workaround. Reviewers: rsmith, aaron.ballman, majnemer, rnk, compnerd, rjmccall Reviewed By: rjmccall Subscribers: rjmccall, rsmith, aaron.ballman, junbuml, mgorny, cfe-commits Differential Revision: https://reviews.llvm.org/D45476 llvm-svn: 331677
* Correct warning on Float->Integer conversions.Erich Keane2018-05-071-0/+17
| | | | | | | | | | | | | | | | | | As identified and briefly discussed here: https://bugs.llvm.org/show_bug.cgi?id=37305 Converting a floating point number to an integer type when the integral part is out of the range of the integer type is undefined behavior in C. Additionally, CodeGen emits an undef in this situation. HOWEVER, we've been giving a warning that says that the value is changed. This patch corrects the warning to list that it is actually undefined behavior. Differential Revision: https://reviews.llvm.org/D46535 llvm-svn: 331673
* Fix explicit template parameter reporting for narrowing conversionsErich Keane2018-05-071-0/+4
| | | | | | | | | | | | | | | | | | I found that explicit template parameters that caused a narrowing integer conversion resulted in the incorrect parameter being mentioned in the note (see test attached). This is because the argument checking code doesn't check to see if it caused SFINAE errors when checking the arguments, so instead of giving up on the first error, it continues through the list. This makes the error reporting pick up the last template param every time. This patch checks these parameters on each argument and gives up if there is an error. The result is that only the required amount of arguments are checked, and that the 'Converted' array contains only the successful arguments before the first failure, as the calls seem to all expect. llvm-svn: 331651
* Disallow pointers to const in __sync_fetch_and_xxx.Aaron Ballman2018-05-051-0/+6
| | | | | | | | | | | | Diagnoses code like: void f(const int *ptr) { __sync_fetch_and_add(ptr, 1); } which matches the behavior of GCC and ICC. llvm-svn: 331598
* Allow writing calling convention attributes on function types.Aaron Ballman2018-05-031-4/+9
| | | | | | Calling convention attributes notionally appertain to the function type -- they modify the mangling of the function, change the behavior of assignment operations, etc. This commit allows the calling convention attributes to be written in the type position as well as the declaration position. llvm-svn: 331459
* Track the result of evaluating a computed noexcept specification on theRichard Smith2018-05-0313-226/+168
| | | | | | | | | | | | | | FunctionProtoType. We previously re-evaluated the expression each time we wanted to know whether the type is noexcept or not. We now evaluate the expression exactly once. This is not quite "no functional change": it fixes a crasher bug during AST deserialization where we would try to evaluate the noexcept specification in a situation where we have not deserialized sufficient portions of the AST to permit such evaluation. llvm-svn: 331428
* [Sema] Do not match function type with const T in template argument deductionLei Liu2018-05-031-0/+6
| | | | | | | | | | | From http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_active.html#1584, function type should not match cv-qualified type in template argument deduction. This also matches what GCC and EDG do in template argument deduction. Differential Revision: https://reviews.llvm.org/D45755 llvm-svn: 331424
* [ObjC] Supress the 'implementing unavailable method' warning whenAlex Lorenz2018-05-031-2/+10
| | | | | | | | | | | | | the method declaration is unavailable for an app extension platform Rationale: Classes are often shared between an app extension code and non-app extension code. There's no way to remove the implementation using preprocessor when building the app extension, so we should not warn here. rdar://38150617 llvm-svn: 331421
* [ObjC] The absence of ownership qualifiers on an ambiguous property leadsAlex Lorenz2018-05-021-4/+14
| | | | | | | | | to synthesis of a valid property even when the selected protocol property has ownership qualifiers rdar://39024725 llvm-svn: 331409
* [OPENMP] Analyze the type of the mapped entity instead of its base.Alexey Bataev2018-05-021-1/+8
| | | | | | | If the mapped entity is a data member, we erroneously checked the type of its base rather than the type of the mapped entity itself. llvm-svn: 331385
* [OPENMP] Do not emit warning for implicitly declared target functions.Alexey Bataev2018-05-021-14/+16
| | | | | | | Since upcoming OpenMP 5.0 functions can be mapped implicitly as declare target and we should not emit warnings for such functions. llvm-svn: 331377
* [OPENMP] Enable c++ exceptions outside of the target constructs iff they areAlexey Bataev2018-05-022-2/+10
| | | | | | | | | | enabled for the host. If the compilation for the host enables C++ exceptions, but they are not supported by the device, we still need to allow the code with the exception handling constructs outside of the target regions. llvm-svn: 331372
* Implement P0482R2, support for char8_t type.Richard Smith2018-05-019-10/+70
| | | | | | | | | | | | This is not yet part of any C++ working draft, and so is controlled by the flag -fchar8_t rather than a -std= flag. (The GCC implementation is controlled by a flag with the same name.) This implementation is experimental, and will be removed or revised substantially to match the proposal as it makes its way through the C++ committee. llvm-svn: 331244
* [OPENMP] Do not emit warning about non-declared target function params.Alexey Bataev2018-04-301-1/+1
| | | | | | | | We should not emit warning that the parameters are not marked as declare target, these declaration are local and cannot be marked as declare target. llvm-svn: 331211
* PR37189 Fix incorrect end source location and spelling for a split '>>' token.Richard Smith2018-04-305-6/+8
| | | | | | | | | | | | | | | | | | When a '>>' token is split into two '>' tokens (in C++11 onwards), or (as an extension) when we do the same for other tokens starting with a '>', we can't just use a location pointing to the first '>' as the location of the split token, because that would result in our miscomputing the length and spelling for the token. As a consequence, for example, a refactoring replacing 'A<X>' with something else would sometimes replace one character too many, and similarly diagnostics highlighting a template-id source range would highlight one character too many. Fix this by creating an expansion range covering the first character of the '>>' token, whose spelling is '>'. For this to work, we generalize the expansion range of a macro FileID to be either a token range (the common case) or a character range (used in this new case). llvm-svn: 331155
* Test commit removing trailing whitespaceStuart Brady2018-04-271-1/+1
| | | | llvm-svn: 331053
* Parse A::template B as an identifier rather than as a template-id with noRichard Smith2018-04-273-11/+14
| | | | | | | | | | template arguments. This fixes some cases where we'd incorrectly accept "A::template B" when B is a kind of template that requires template arguments (in particular, a variable template or a concept). llvm-svn: 331013
* [X86] Make __builtin_ia32_readeflags_u32 and __builtin_ia32_writeeflags_u32 ↵Craig Topper2018-04-261-0/+17
| | | | | | | | | | | | only available on 32-bit targets. These builtins can't be handled by the backend on 64-bit targets. So error up front instead of throwing an isel error. Fixes PR37225 Differential Revision: https://reviews.llvm.org/D46132 llvm-svn: 330987
* Diagnose missing template arguments for a variable template even when there isRichard Smith2018-04-261-0/+8
| | | | | | | | | a preceding 'template' keyword. We only diagnose in the dependent case (wherein we used to crash). Another bug prevents the diagnostic from appearing in the non-template case. llvm-svn: 330894
* Revert addition of 'concept' to diagnostics in r330890.Richard Smith2018-04-261-2/+0
| | | | | | Matches revert in r330888 of r330794. llvm-svn: 330891
* Factor out common code for diagnosing missing template arguments.Richard Smith2018-04-264-22/+18
| | | | | | In passing, add 'concept' to the list of template kinds in diagnostics. llvm-svn: 330890
OpenPOWER on IntegriCloud