r263854 - [TLS on Darwin] use CXX_FAST_TLS calling convention for tls_init.

Author: mren Date: Fri Mar 18 18:35:21 2016 New Revision: 263854 URL: http://llvm.org/viewvc/llvm-project?rev=263854=rev Log: [TLS on Darwin] use CXX_FAST_TLS calling convention for tls_init. This makes sure we don't generate a lot of code to spill/reload CSRs when calling tls_init from the

Re: [PATCH] D18191: [clang-tidy] Add check for function parameters that are const& to builtin types

sdowney updated this revision to Diff 50982. sdowney added a comment. Add more test cases, covering templates, template specializations, function definitions, member function templates. Check if parent function of the argument is a nullptr. Fix spelling mistake.

Re: [PATCH] D17482: [clang-tidy] Allow tests to verify changes made to header files

LegalizeAdulthood added a comment. In http://reviews.llvm.org/D17482#372677, @alexfh wrote: > Can you [...] say how many times this feature is going to be used, if it gets > added to the testing script? Since you can't verify changes against headers right now, noone has bothered to add test

[libcxx] r263746 - unord: Extract key to avoid preemptive mallocs in insert/emplace

Author: dexonsmith Date: Thu Mar 17 15:45:20 2016 New Revision: 263746 URL: http://llvm.org/viewvc/llvm-project?rev=263746=rev Log: unord: Extract key to avoid preemptive mallocs in insert/emplace unordered_set::emplace and unordered_map::emplace construct a node, then try to insert it. If

[PATCH] D18268: [Objective-c] Fix a crash in WeakObjectProfileTy::getBaseInfo

ahatanak created this revision. ahatanak added a reviewer: jordan_rose. ahatanak added a subscriber: cfe-commits. The crash occurs in WeakObjectProfileTy::getBaseInfo when getBase() is called on an ObjCPropertyRefExpr object whose receiver is an interface. This patch fixes the crash by checking

Re: [PATCH] D17951: Implement is_always_lock_free

bcraig added a subscriber: bcraig. Comment at: include/atomic:859 @@ +858,3 @@ +template <> _LIBCPP_CONSTEXPR bool __libcpp_always_lock_free = 2 == ATOMIC_CHAR32_T_LOCK_FREE; +template <> _LIBCPP_CONSTEXPR bool __libcpp_always_lock_free = 2 == ATOMIC_WCHAR_T_LOCK_FREE;

Re: [PATCH] D16579: Warn if friend function depends on template parameters.

rsmith added a subscriber: rsmith. rsmith added a comment. It seems useful to warn on this, but I'm concerned that there are valid code patterns that this will warn about where there is no reasonable way to rewrite the code that would suppress the warning. For instance: template class X {

[PATCH] D18217: [libcxx] Fix -Wdeprecated warnings

hintonda created this revision. hintonda added reviewers: rsmith, howard.hinnant, logan. hintonda added a subscriber: cfe-commits. Fix -Wdeprecated warning for throw() specs and implicit copy ctor's and assignmen oper's. http://reviews.llvm.org/D18217 Files: include/__bit_reference

Re: [PATCH] D18088: Add a new warning to notify users of mismatched SDK and deployment target

edward-san added a subscriber: edward-san. edward-san added a comment. Excuse me for intruding, but from what I see, this warning applies only to Apple SDK. Since SDK could mean everything, what about naming the warning accordingly, like '-Wincompatible-apple-sdk' or '-Wapple-incompatible-sdk'

Re: [PATCH] D18191: [clang-tidy] Add check for function parameters that are const& to builtin types

sdowney added a comment. In http://reviews.llvm.org/D18191#376471, @LegalizeAdulthood wrote: > There is utility in the definition of a function in saying that an argument > is `const int i` instead of `int i`. The const-ness declares the intent that > this local variable is not going to be

Re: [PATCH] D17104: [VFS] Drop path traversal assertion

Thanks! On Wed, Mar 16, 2016 at 5:35 AM, NAKAMURA Takumi wrote: > chapuni added a subscriber: chapuni. > chapuni added a comment. > > Excuse me, I have reverted it in r263636. > > > > Comment at: lib/Frontend/ModuleDependencyCollector.cpp:65 > @@ +64,3 @@ >

Re: [PATCH] D17784: Check if LLVM_PREFIX is defined before using it.

chaoren updated this revision to Diff 50993. chaoren added a comment. Remove usage of LLVM_PREFIX altogether. http://reviews.llvm.org/D17784 Files: lib/Driver/ToolChains.cpp Index: lib/Driver/ToolChains.cpp === ---

Re: [PATCH] D16989: Change interpretation of function definition in friend declaration of template class.

sepavloff added a comment. Topic of friend function instantiation in C++ Standard discussion group

Re: r263732 - Remove defaulted move ops, the type is zero-cost copyable anyway, so there's no need for specific move ops

I just did some tests, and when a derived class is explicitly moved, e.g., with std::move(), the current behavior is (clang++/g++/vc++): 1) if base has implicit move, it gets called 2) if base doesn't have an implicit move (because a copy ctor was define, etc) copy ctor is called 3) if base class

r263766 - Remove usage of LLVM_PREFIX.

Author: chaoren Date: Thu Mar 17 19:05:37 2016 New Revision: 263766 URL: http://llvm.org/viewvc/llvm-project?rev=263766=rev Log: Remove usage of LLVM_PREFIX. Summary: LLVM_PREFIX could be undefined if CMAKE_INSTALL_PREFIX were set to empty. Reviewers: kparzysz, bkramer, chandlerc Subscribers:

Re: [PATCH] D18263: [clang-tools-extra] Add release notes documentation

alexfh accepted this revision. alexfh added a comment. This revision is now accepted and ready to land. Looks good with one comment. Thank you, Richard! Comment at: docs/ReleaseNotes.rst:29 @@ +28,3 @@ +For more information about Clang or LLVM, including information about +the

Re: [PATCH] D18280: [tsan] Allow -fsanitize=thread for iOS-style simulator targets

dcoughlin added a comment. There is a related patch for compiler-rt at http://reviews.llvm.org/D18277 http://reviews.llvm.org/D18280 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Re: [PATCH] D17950: Implement is_always_lock_free

On Thu, Mar 17, 2016 at 12:55 PM, Craig, Ben wrote: > I know that MIPS does that, and an out-of-tree implementation of hexagon > implements 1-byte cmpxchg in terms of the 4-byte version. The emulation > code isn't particularly small, and it seems reasonable to make it a

Re: [PATCH] D17950: Implement is_always_lock_free

jfb added a comment. In http://reviews.llvm.org/D17950#376349, @jyknight wrote: > This conflicts with http://reviews.llvm.org/D17933. Most of this change also > seems unnecessary. > > - I think the `is_always_lock_free` function should be defined based on the > existing

Re: [PATCH] D17950: Implement is_always_lock_free

> A 4 byte cmpxchg could be lock free, while a 1 byte cmpxchg may not be lock free. That's not possible: if you have a 4-byte cmpxchg instruction, you can use it to implement a 1-byte cmpxchg, too. Many targets do this already. (It would be better if that was available generically so that code

[clang-tools-extra] r263726 - Fix Clang-tidy modernize-deprecated-headers warnings; other minor fixes.

Author: eugenezelenko Date: Thu Mar 17 12:02:25 2016 New Revision: 263726 URL: http://llvm.org/viewvc/llvm-project?rev=263726=rev Log: Fix Clang-tidy modernize-deprecated-headers warnings; other minor fixes. Differential revision: http://reviews.llvm.org/D18231 Modified:

Re: [PATCH] D17482: [clang-tidy] Allow tests to verify changes made to header files

LegalizeAdulthood added a comment. test/clang-tidy/misc-unused-parameters.cpp currently does a hack to verify headers by committing the manually fixed header and diffing it: // RUN: echo "static void staticFunctionHeader(int i) {}" > %T/header.h // RUN: echo "static void

Re: [PATCH] D18268: [Objective-c] Fix a crash in WeakObjectProfileTy::getBaseInfo

jordan_rose accepted this revision. jordan_rose added a comment. This revision is now accepted and ready to land. Ah, of course! Thanks for catching this, Akira. Can you add this case to the table in the doc comment for WeakObjectProfileTy? (That's how I convinced myself it was correct.)

Re: [PATCH] D17407: [Sema] PR25755 Fix crash when initializing out-of-order struct references

rsmith added inline comments. Comment at: lib/Sema/SemaInit.cpp:1732 @@ +1731,3 @@ + unsigned FieldIdx = 0; + std::vector SeenFields; + RecordDecl::field_iterator FieldStart = Field; Please use an `llvm::BitVector` here instead. Comment at:

r263718 - Use a simpler set of mock headers for the vfs+modules crash recovery tests.

Author: d0k Date: Thu Mar 17 11:19:51 2016 New Revision: 263718 URL: http://llvm.org/viewvc/llvm-project?rev=263718=rev Log: Use a simpler set of mock headers for the vfs+modules crash recovery tests. The System/ mock is large and too complex for this test. It can cause the tests to fail in

Re: [PATCH] D18024: Remove compile time PreserveName switch based on NDEBUG

joker.eph closed this revision. joker.eph added a comment. r263257 http://reviews.llvm.org/D18024 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Re: [PATCH] D17893: Sema: Add semantic analysis for the C++ ABI stability attributes and whitelist.

joker.eph added a comment. LGTM, but I'm not a clang expert, so if Richard or someone else could double-check? Comment at: lib/Sema/SemaDeclCXX.cpp:4999 @@ +4998,3 @@ +return; + } + Isn't this correct by the loop which starts with `DC =

Re: [PATCH] D15469: Expose cxx constructor and method properties through libclang and python bindings.

jbcoe added a comment. If I remove `= delete` from the declaration of the member function `foo` in test/Index/index-file.cpp line 36 then `foo` is reported in the output as [indexDeclaration]: kind: c++-instance-method | name: foo | USR: c:@S@B@F@foo# | lang: C++ | cursor: CXXMethod=foo:36:8

Re: [PATCH] D17990: [clang-tidy] minor improvements in modernise-deprecated-headers check

alexfh added inline comments. Comment at: clang-tidy/modernize/DeprecatedHeadersCheck.cpp:59 @@ +58,3 @@ +{"float.h", "cfloat"}, +{"inttypes.h", "cinttypes"}, +{"limits.h", "climits"}, According to

r263710 - clang-format: [JS] Fix incorrect spacing around contextual keywords.

Author: djasper Date: Thu Mar 17 07:17:59 2016 New Revision: 263710 URL: http://llvm.org/viewvc/llvm-project?rev=263710=rev Log: clang-format: [JS] Fix incorrect spacing around contextual keywords. Before: x.of (); After: x.of(); Modified: cfe/trunk/lib/Format/TokenAnnotator.cpp

Re: [PATCH] D18280: [tsan] Allow -fsanitize=thread for iOS-style simulator targets

kubabrecka added a comment. LGTM after adding a check for a simulator. http://reviews.llvm.org/D18280 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Re: [PATCH] D18261: Show members of DeclContexts (i.e., class members) in Visual Studio native visualizers

mspertus closed this revision. mspertus added a comment. Thanks, Aaron. Committed as revision 263794 http://reviews.llvm.org/D18261 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Re: [PATCH] D18138: Add -fnative-half-arguments-and-returns

This revision was automatically updated to reflect the committed changes. Closed by commit rL263795: Add -fnative-half-arguments-and-returns (authored by pirama). Changed prior to commit: http://reviews.llvm.org/D18138?vs=50564=51036#toc Repository: rL LLVM http://reviews.llvm.org/D18138

[PATCH] D18243: [ASTMatchers] Existing matcher hasAnyArgument fixed

baloghadamsoftware created this revision. baloghadamsoftware added reviewers: klimek, sbenza. baloghadamsoftware added subscribers: cfe-commits, xazax.hun. Herald added a subscriber: klimek. A checker (will be uploaded after this patch) needs to check implicit casts. The checker needs matcher

r263635 - Directly get the canonical Type instead of going around through a CanQualType temporary, NFC.

Author: yrnkrn Date: Wed Mar 16 07:14:43 2016 New Revision: 263635 URL: http://llvm.org/viewvc/llvm-project?rev=263635=rev Log: Directly get the canonical Type instead of going around through a CanQualType temporary, NFC. Modified: cfe/trunk/lib/CodeGen/ItaniumCXXABI.cpp Modified:

Re: [PATCH] D18123: Fix implicit copy ctor and copy assignment operator warnings when -Wdeprecated passed.

The current behavior, albeit deprecated, is to implicitly define these. Therefore, it would be incorrect not to delete them if the implicit versions don't do the right thing. I'd be happy to add a FIXME, but I doubt they will ever be removed. At best, they'd be #ifdef'd away for some future

Re: [PATCH] D18199: CodeGen: Implement IR generation for the relative vtable ABI (PR26723).

pcc added inline comments. Comment at: lib/CodeGen/ItaniumCXXABI.cpp:1627 @@ +1626,3 @@ + +llvm::Value *CodeGenFunction::GetVirtualFunctionFromVTable(const CXXRecordDecl *RD, + llvm::Value *VTable, rnk

Re: [PATCH] D18127: Remove compile time PreserveName in favor of a runtime cc1 -discard-value-names option

joker.eph closed this revision. joker.eph added a comment. r263394 http://reviews.llvm.org/D18127 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

r263816 - [clang-cl] Allow use of -gline-tables-only

Author: rnk Date: Fri Mar 18 13:42:56 2016 New Revision: 263816 URL: http://llvm.org/viewvc/llvm-project?rev=263816=rev Log: [clang-cl] Allow use of -gline-tables-only Modified: cfe/trunk/include/clang/Driver/Options.td cfe/trunk/test/Driver/cl-options.c Modified:

Re: [PATCH] D18263: [clang-tools-extra] Add release notes documentation

LegalizeAdulthood updated this revision to Diff 51102. LegalizeAdulthood marked 2 inline comments as done. LegalizeAdulthood added a comment. Update from review comments. I do not have commit access. Patch by Richard Thomson http://reviews.llvm.org/D18263 Files: docs/ReleaseNotes.rst

Re: [PATCH] D18270: Typesafe visualization of PointerIntPairs in Visual Studio

mspertus closed this revision. mspertus added a comment. Committed as revision 263821. Thanks for the trust, Aaron! Note that my next revision (function and member prototype visualization) will be largish, and I would still appreciate a review before commit. http://reviews.llvm.org/D18270

Re: [PATCH] D18271: Avoid -Wshadow warnings about constructor parameters named after fields

dblaikie added a subscriber: dblaikie. dblaikie added a comment. It's not just modifications of shadowed variables that are a problem - one of the one's I'm concerned we should catch is: struct foo { std::unique_ptr p; foo(std::unique_ptr p) : p(std::move(p)) { f(*p); } };

Re: [PATCH] D17451: PR26448: [Sema] Fix determined type of ternary operator acting on two xvalue class types

erik.pilkington updated this revision to Diff 51045. erik.pilkington added a comment. Fix the wording of the quoted section of the standard. Would you mind committing this? Thanks for reviewing! http://reviews.llvm.org/D17451 Files: lib/Sema/SemaExprCXX.cpp

Re: [PATCH] D18123: Fix implicit copy ctor and copy assignment operator warnings when -Wdeprecated passed.

-Werror clean is great. If you could add -Wdeprecated, then we wouldn't need to delete them -- the warning is only issued with -Wdeprecated is passed. On Wed, Mar 16, 2016 at 1:49 PM, David Blaikie wrote: > > > On Wed, Mar 16, 2016 at 10:19 AM, don hinton

Re: [PATCH] D18231: [Clang-tools-extra] Fix Clang-tidy modernize-deprecated-headers warnings; other minor fixes

aaron.ballman accepted this revision. aaron.ballman added a comment. This revision is now accepted and ready to land. LGTM, thanks! Repository: rL LLVM http://reviews.llvm.org/D18231 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

Re: [PATCH] D18238: [clang-tidy] Fix clang-tidy crashes when using -fdelayed-template-parsing.

aaron.ballman requested changes to this revision. aaron.ballman added a comment. This revision now requires changes to proceed. Oops, I take that LGTM back. ;-) Almost LGTM -- this is missing test cases that exercise the behavior. http://reviews.llvm.org/D18238

Re: [modules] PR24954

On 16/03/16 03:22, Richard Smith wrote: Please restrict this to the case where the function template is a friend (check D->getFriendObjectKind()). Otherwise, this looks good to me. Thanks! Done and committed in r263634. Thanks! On Mon, Mar 7, 2016 at 2:23 AM, Vassil Vassilev

Re: [PATCH] D18275: [ASTMatchers] Add own version of VariadicFunction.

sbenza added a comment. Alex, this is what we discussed to make `hasAnyName` take a `vector` directly. http://reviews.llvm.org/D18275 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

Re: [PATCH] D18203: [OPENMP] Implementation of codegen for firstprivate clause of target directive

carlo.bertolli updated this revision to Diff 50829. carlo.bertolli marked an inline comment as done. carlo.bertolli added a comment. Modify diff to reflect comments. Repository: rL LLVM http://reviews.llvm.org/D18203 Files: include/clang/Basic/DiagnosticSemaKinds.td

Re: [PATCH] D17951: Implement is_always_lock_free

EricWF added inline comments. Comment at: include/atomic:859 @@ +858,3 @@ +template <> _LIBCPP_CONSTEXPR bool __libcpp_always_lock_free = 2 == ATOMIC_CHAR32_T_LOCK_FREE; +template <> _LIBCPP_CONSTEXPR bool __libcpp_always_lock_free = 2 == ATOMIC_WCHAR_T_LOCK_FREE; +template <>

Re: [PATCH] D18238: [clang-tidy] Fix clang-tidy crashes when using -fdelayed-template-parsing.

aaron.ballman accepted this revision. aaron.ballman added a comment. This revision is now accepted and ready to land. LGTM now (this time for sure), thank you! http://reviews.llvm.org/D18238 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

Re: [PATCH] D18123: Fix implicit copy ctor and copy assignment operator warnings when -Wdeprecated passed.

Done, however, I don't have commit access, so once it's approved, could you commit it for me? thanks... don On Wed, Mar 16, 2016 at 6:00 PM, David Blaikie wrote: > > > On Wed, Mar 16, 2016 at 11:38 AM, don hinton via cfe-commits < > cfe-commits@lists.llvm.org> wrote: > >>

Re: [PATCH] D18271: Avoid -Wshadow warnings about constructor parameters named after fields

+Lang, because he was asking me recently about this improvement & thinking of chipping in On Fri, Mar 18, 2016 at 10:56 AM, David Blaikie via cfe-commits < cfe-commits@lists.llvm.org> wrote: > dblaikie added a subscriber: dblaikie. > dblaikie added a comment. > > It's not just modifications of

Re: [PATCH] D18163: Add visualizers for more clang types. Create more C++-like visualizations for existing Clang types

aaron.ballman added inline comments. Comment at: clang.natvis:25 @@ +24,3 @@ +{*(clang::BuiltinType *)this} +{*(clang::PointerType *)this} +{*(clang::AttributedType *)this} If we're handling pointers, perhaps we should also do references similarly?

Re: r263687 - Add an optional named argument (replacement = "xxx") to AvailabilityAttr.

Revert it in r263752. Thanks, Manman On Thu, Mar 17, 2016 at 2:17 PM, Manman Ren wrote: > Sorry, working on it. > > Manman > > On Thu, Mar 17, 2016 at 12:45 PM, Kostya Serebryany > wrote: > >> This change is causing ubsan bot to complain >>

Re: [PATCH] D17407: [Sema] PR25755 Fix crash when initializing out-of-order struct references

hintonda updated this revision to Diff 51099. hintonda added a comment. Addressed comments. http://reviews.llvm.org/D17407 Files: lib/Sema/SemaInit.cpp test/SemaCXX/cxx0x-initializer-constructor.cpp Index: test/SemaCXX/cxx0x-initializer-constructor.cpp

Re: [PATCH] D18088: Add a new warning to notify users of mismatched SDK and deployment target

bob.wilson added a comment. Yes, in its current usage this warning is only used for Apple's SDKs, but how does it help to put "apple" in the name of the diagnostic? Are you concerned about a name conflict with a similar diagnostic for non-Apple SDKs? http://reviews.llvm.org/D18088

Re: [PATCH] D17285: Fix printing of anonymous struct typedefs

This revision was automatically updated to reflect the committed changes. Closed by commit rL263836: Fix printing of anonymous struct typedefs. (authored by steven_watanabe). Changed prior to commit: http://reviews.llvm.org/D17285?vs=48042=51072#toc Repository: rL LLVM

Re: [PATCH] D18136: boost-use-to-string check

Prazek updated the summary for this revision. Prazek updated this revision to Diff 51057. Prazek added a comment. fixed many things thaks to Alexander http://reviews.llvm.org/D18136 Files: clang-tidy/boost/BoostTidyModule.cpp clang-tidy/boost/CMakeLists.txt

Re: [PATCH] D12147: Fix typos in lib/AST

rsmith added a comment. All the changes other than 'modelled' look fine :) Do you need someone to commit this for you? Comment at: lib/AST/ASTContext.cpp:1631 @@ -1630,3 +1630,3 @@ case BuiltinType::OCLSampler: - // Samplers are modeled as integers. + //

[PATCH] D18275: [ASTMatchers] Add own version of VariadicFunction.

sbenza created this revision. sbenza added a reviewer: alexfh. sbenza added a subscriber: cfe-commits. Herald added a subscriber: klimek. llvm::VariadicFunction is only being used by ASTMatchers. Having own copy here allows us to remove the other one from llvm/ADT. Also, we can extend the API to

Re: [PATCH] D18191: [clang-tidy] Add check for function parameters that are const& to builtin types

LegalizeAdulthood added a comment. In http://reviews.llvm.org/D18191#377686, @sdowney wrote: > In http://reviews.llvm.org/D18191#376471, @LegalizeAdulthood wrote: > > > There is utility in the definition of a function in saying that an argument > > is `const int i` instead of `int i`. The

Re: [PATCH] D15031: CFG: Add CFGElement for automatic variables that leave the scope

alexfh added a subscriber: alexfh. alexfh added a comment. Richard, can you review this? http://reviews.llvm.org/D15031 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[libcxx] r263798 - Missing ATOMIC_*_LOCK_FREE tests

Author: jfb Date: Fri Mar 18 12:48:58 2016 New Revision: 263798 URL: http://llvm.org/viewvc/llvm-project?rev=263798=rev Log: Missing ATOMIC_*_LOCK_FREE tests Forked from D17951, these tests should have been there but weren't. Modified: