<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On 4 April 2018 at 16:57, Akira Hatanaka via cfe-commits <span dir="ltr"><<a href="mailto:cfe-commits@lists.llvm.org" target="_blank">cfe-commits@lists.llvm.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="word-wrap:break-word"><br><div><span class="gmail-"><blockquote type="cite"><div>On Apr 4, 2018, at 4:47 PM, Richard Smith <<a href="mailto:richard@metafoo.co.uk" target="_blank">richard@metafoo.co.uk</a>> wrote:</div><br class="gmail-m_-6053479395256423563Apple-interchange-newline"><div><div dir="ltr" style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><div class="gmail_extra"><div class="gmail_quote">On 4 April 2018 at 16:24, Akira Hatanaka via cfe-commits<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><span dir="ltr"><<a href="mailto:cfe-commits@lists.llvm.org" target="_blank">cfe-commits@<wbr>lists.llvm.org</a>></span><span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="word-wrap:break-word"><span><div><br><blockquote type="cite"><div>On Apr 4, 2018, at 3:38 PM, Richard Smith <<a href="mailto:richard@metafoo.co.uk" target="_blank">richard@metafoo.co.uk</a>> wrote:</div><br class="gmail-m_-6053479395256423563m_3142434014750245271Apple-interchange-newline"><div><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">Hi Akira,</div><div class="gmail_quote"><br></div><div class="gmail_quote">This change has broken the C++ versions of these type traits for classes with volatile members. Such classes are required to claim to be trivial per C++ DR 2094 (<a href="http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_defects.html#2094" target="_blank">http://www.open-std.org/jtc1/<wbr>sc22/wg21/docs/cwg_defects.htm<wbr>l#2094</a>) but return false from isNonTrivialToPrimitiveCopy().</div><div class="gmail_quote"><br></div></div></div></div></blockquote><br></div></span><div>Oh that’s right. The function returns false when isNonTrivialToPrimitiveCopy() returns PCK_VolatileTrivial. That is wrong.</div></div></blockquote><div><br></div><div>Do you have a fix that's going to land soon? This has caused a regression for us, so I'd like to revert if it's not going to be fixed imminently.</div></div></div></div></div></blockquote><div><br></div></span><div>Yes, I plan to fix this soon, but feel free to revert it if you need to.</div></div></div></blockquote><div><br></div><div> Reverted in r329247 for now.</div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="word-wrap:break-word"><div><div><div class="gmail-h5"><blockquote type="cite"><div><div dir="ltr" style="font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:normal;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><div class="gmail_extra"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="word-wrap:break-word"><div><span><blockquote type="cite"><div><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">Also, exposing these __has_* traits more widely seems like a backwards step to me: these traits are deprecated, near-useless, and we're trying to remove them. No code should be using them under any circumstances; the __is_* traits should be used instead.</div><div class="gmail_quote"><br></div></div></div></div></blockquote><div><br></div></span><div>The __is_* traits (is_trivially_copy_constructib<wbr>le, etc.) are templates defined in libcxx, so it seems that we can’t use them when compiling in C mode. Is it OK to add their definitions to TokenKinds.def as non-C++ keywords?</div><div><div class="gmail-m_-6053479395256423563h5"><div><br></div><blockquote type="cite"><div><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On 27 March 2018 at 17:12, Akira Hatanaka via cfe-commits<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><span dir="ltr"><<a href="mailto:cfe-commits@lists.llvm.org" target="_blank">cfe-commits@<wbr>lists.llvm.org</a>></span><span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Author: ahatanak<br>Date: Tue Mar 27 17:12:08 2018<br>New Revision: 328680<br><br>URL:<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project?rev=328680&view=rev" rel="noreferrer" target="_blank">http://llvm.org/viewvc/<wbr>llvm-project?rev=328680&view=<wbr>rev</a><br>Log:<br>[ObjC] Make C++ triviality type traits available to non-trivial C<br>structs.<br><br>r326307 and r327870 made changes that allowed using non-trivial C<br>structs with fields qualified with __strong or __weak. This commit makes<br>the following C++ triviality type traits available to non-trivial C<br>structs:<br><br>__has_trivial_assign<br>__has_trivial_move_assign<br>__has_trivial_copy<br>__has_trivial_move_constructor<br>__has_trivial_constructor<br>__has_trivial_destructor<br><br><a>rdar://problem/33599681</a><br><br>Differential Revision:<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="https://reviews.llvm.org/D44913" rel="noreferrer" target="_blank">https://reviews.<wbr>llvm.org/D44913</a><br><br>Added:<br> <span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>cfe/trunk/test/SemaObjC/non-<wbr>trivial-struct-traits.m<br>Modified:<br> <span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>cfe/trunk/include/clang/<wbr>Basic/TokenKinds.def<br> <span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>cfe/trunk/lib/Sema/<wbr>SemaExprCXX.cpp<br><br>Modified: cfe/trunk/include/clang/Basic/<wbr>TokenKinds.def<br>URL:<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/cfe/trunk/include/clang/Basic/TokenKinds.def?rev=328680&r1=328679&r2=328680&view=diff" rel="noreferrer" target="_blank">http://llvm.org/viewvc/<wbr>llvm-project/cfe/trunk/<wbr>include/clang/Basic/<wbr>TokenKinds.def?rev=328680&r1=<wbr>328679&r2=328680&view=diff</a><br>==============================<wbr>==============================<wbr>==================<br>--- cfe/trunk/include/clang/Basic/<wbr>TokenKinds.def (original)<br>+++ cfe/trunk/include/clang/Basic/<wbr>TokenKinds.def Tue Mar 27 17:12:08 2018<br>@@ -433,12 +433,12 @@ TYPE_TRAIT_1(__has_nothrow_ass<wbr>ign, HasNo<br> TYPE_TRAIT_1(__has_nothrow_mo<wbr>ve_assign, HasNothrowMoveAssign, KEYCXX)<br> TYPE_TRAIT_1(__has_nothrow_co<wbr>py, HasNothrowCopy, KEYCXX)<br> TYPE_TRAIT_1(__has_nothrow_co<wbr>nstructor, HasNothrowConstructor, KEYCXX)<br>-TYPE_TRAIT_1(__has_trivial_as<wbr>sign, HasTrivialAssign, KEYCXX)<br>-TYPE_TRAIT_1(__has_trivial_mo<wbr>ve_assign, HasTrivialMoveAssign, KEYCXX)<br>-TYPE_TRAIT_1(__has_trivial_co<wbr>py, HasTrivialCopy, KEYCXX)<br>-TYPE_TRAIT_1(__has_trivial_co<wbr>nstructor, HasTrivialDefaultConstructor, KEYCXX)<br>-TYPE_TRAIT_1(__has_trivial_mo<wbr>ve_constructor, HasTrivialMoveConstructor, KEYCXX)<br>-TYPE_TRAIT_1(__has_trivial_de<wbr>structor, HasTrivialDestructor, KEYCXX)<br>+TYPE_TRAIT_1(__has_trivial_as<wbr>sign, HasTrivialAssign, KEYALL)<br>+TYPE_TRAIT_1(__has_trivial_mo<wbr>ve_assign, HasTrivialMoveAssign, KEYALL)<br>+TYPE_TRAIT_1(__has_trivial_co<wbr>py, HasTrivialCopy, KEYALL)<br>+TYPE_TRAIT_1(__has_trivial_co<wbr>nstructor, HasTrivialDefaultConstructor, KEYALL)<br>+TYPE_TRAIT_1(__has_trivial_mo<wbr>ve_constructor, HasTrivialMoveConstructor, KEYALL)<br>+TYPE_TRAIT_1(__has_trivial_de<wbr>structor, HasTrivialDestructor, KEYALL)<br> TYPE_TRAIT_1(__has_virtual_de<wbr>structor, HasVirtualDestructor, KEYCXX)<br> TYPE_TRAIT_1(__is_abstract, IsAbstract, KEYCXX)<br> TYPE_TRAIT_1(__is_aggregate, IsAggregate, KEYCXX)<br><br>Modified: cfe/trunk/lib/Sema/SemaExprCXX<wbr>.cpp<br>URL:<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/SemaExprCXX.cpp?rev=328680&r1=328679&r2=328680&view=diff" rel="noreferrer" target="_blank">http://llvm.org/viewvc/<wbr>llvm-project/cfe/trunk/lib/<wbr>Sema/SemaExprCXX.cpp?rev=<wbr>328680&r1=328679&r2=328680&<wbr>view=diff</a><br>==============================<wbr>==============================<wbr>==================<br>--- cfe/trunk/lib/Sema/SemaExprCXX<wbr>.cpp (original)<br>+++ cfe/trunk/lib/Sema/SemaExprCXX<wbr>.cpp Tue Mar 27 17:12:08 2018<br>@@ -4513,6 +4513,8 @@ static bool EvaluateUnaryTypeTrait(Sema<br> // does not correctly compute triviality in the presence of multiple special<br> // members of the same kind. Revisit this once the g++ bug is fixed.<br> case UTT_HasTrivialDefaultConstruct<wbr>or:<br>+ if (T.isNonTrivialToPrimitiveDefa<wbr>ultInitialize())<br>+ return false;<br> //<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html" rel="noreferrer" target="_blank">http://gcc.gnu.org/<wbr>onlinedocs/gcc/Type-Traits.<wbr>html</a>:<br> // If __is_pod (type) is true then the trait is true, else if type is<br> // a cv class or union type (or array thereof) with a trivial default<br>@@ -4524,6 +4526,8 @@ static bool EvaluateUnaryTypeTrait(Sema<br> <span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>!RD-><wbr>hasNonTrivialDefaultConstructo<wbr>r();<br> return false;<br> case UTT_HasTrivialMoveConstructor:<br>+ if (T.isNonTrivialToPrimitiveDest<wbr>ructiveMove())<br>+ return false;<br> // This trait is implemented by MSVC 2012 and needed to parse the<br> // standard library headers. Specifically this is used as the logic<br> // behind std::is_trivially_move_constru<wbr>ctible (20.9.4.3).<br>@@ -4533,6 +4537,8 @@ static bool EvaluateUnaryTypeTrait(Sema<br> return RD->hasTrivialMoveConstructor(<wbr>) && !RD->hasNonTrivialMoveConstruc<wbr>tor();<br> return false;<br> case UTT_HasTrivialCopy:<br>+ if (T.isNonTrivialToPrimitiveCopy<wbr>())<br>+ return false;<br> //<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html" rel="noreferrer" target="_blank">http://gcc.gnu.org/<wbr>onlinedocs/gcc/Type-Traits.<wbr>html</a>:<br> // If __is_pod (type) is true or type is a reference type then<br> // the trait is true, else if type is a cv class or union type<br>@@ -4545,6 +4551,8 @@ static bool EvaluateUnaryTypeTrait(Sema<br> <span class="gmail-m_-6053479395256423563Apple-converted-space"> </span>!RD-><wbr>hasNonTrivialCopyConstructor()<wbr>;<br> return false;<br> case UTT_HasTrivialMoveAssign:<br>+ if (T.isNonTrivialToPrimitiveDest<wbr>ructiveMove())<br>+ return false;<br> // This trait is implemented by MSVC 2012 and needed to parse the<br> // standard library headers. Specifically it is used as the logic<br> // behind std::is_trivially_move_assigna<wbr>ble (20.9.4.3)<br>@@ -4554,6 +4562,8 @@ static bool EvaluateUnaryTypeTrait(Sema<br> return RD->hasTrivialMoveAssignment() && !RD->hasNonTrivialMoveAssignme<wbr>nt();<br> return false;<br> case UTT_HasTrivialAssign:<br>+ if (T.isNonTrivialToPrimitiveCopy<wbr>())<br>+ return false;<br> //<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html" rel="noreferrer" target="_blank">http://gcc.gnu.org/<wbr>onlinedocs/gcc/Type-Traits.<wbr>html</a>:<br> // If type is const qualified or is a reference type then the<br> // trait is false. Otherwise if __is_pod (type) is true then the<br>@@ -4624,6 +4634,8 @@ static bool EvaluateUnaryTypeTrait(Sema<br> return true;<br><br> case UTT_HasTrivialDestructor:<br>+ if (T.isDestructedType() == QualType::DK_nontrivial_c_stru<wbr>ct)<br>+ return false;<br> //<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://gcc.gnu.org/onlinedocs/gcc/Type-Traits.html" rel="noreferrer" target="_blank">http://gcc.gnu.org/<wbr>onlinedocs/gcc/Type-Traits.<wbr>html</a><br> // If __is_pod (type) is true or type is a reference type<br> // then the trait is true, else if type is a cv class or union<br><br>Added: cfe/trunk/test/SemaObjC/non-tr<wbr>ivial-struct-traits.m<br>URL:<span class="gmail-m_-6053479395256423563Apple-converted-space"> </span><a href="http://llvm.org/viewvc/llvm-project/cfe/trunk/test/SemaObjC/non-trivial-struct-traits.m?rev=328680&view=auto" rel="noreferrer" target="_blank">http://llvm.org/viewvc/<wbr>llvm-project/cfe/trunk/test/<wbr>SemaObjC/non-trivial-struct-<wbr>traits.m?rev=328680&view=auto</a><br>==============================<wbr>==============================<wbr>==================<br>--- cfe/trunk/test/SemaObjC/non-tr<wbr>ivial-struct-traits.m (added)<br>+++ cfe/trunk/test/SemaObjC/non-tr<wbr>ivial-struct-traits.m Tue Mar 27 17:12:08 2018<br>@@ -0,0 +1,25 @@<br>+// RUN: %clang_cc1 -fsyntax-only -fobjc-arc -verify %s<br>+<br>+// expected-no-diagnostics<br>+<br>+struct Trivial {<br>+ int x;<br>+};<br>+<br>+struct NonTrivial {<br>+ id x;<br>+};<br>+<br>+int trivial_assign[__has_trivial_a<wbr>ssign(struct Trivial) ? 1 : -1];<br>+int trivial_move_assign[__has_triv<wbr>ial_move_assign(struct Trivial) ? 1 : -1];<br>+int trivial_copy_constructor[__has<wbr>_trivial_copy(struct Trivial) ? 1 : -1];<br>+int trivial_move_constructor[__has<wbr>_trivial_move_constructor(stru<wbr>ct Trivial) ? 1 : -1];<br>+int trivial_constructor[__has_triv<wbr>ial_constructor(struct Trivial) ? 1 : -1];<br>+int trivial_destructor[__has_trivi<wbr>al_destructor(struct Trivial) ? 1 : -1];<br>+<br>+int non_trivial_assign[__has_trivi<wbr>al_assign(struct NonTrivial) ? -1 : 1];<br>+int non_trivial_move_assign[__has_<wbr>trivial_move_assign(struct NonTrivial) ? -1 : 1];<br>+int non_trivial_copy_constructor[_<wbr>_has_trivial_copy(struct NonTrivial) ? -1 : 1];<br>+int non_trivial_move_constructor[_<wbr>_has_trivial_move_constructor(<wbr>struct NonTrivial) ? -1 : 1];<br>+int non_trivial_constructor[__has_<wbr>trivial_constructor(struct NonTrivial) ? -1 : 1];<br>+int non_trivial_destructor[__has_t<wbr>rivial_destructor(struct NonTrivial) ? -1 : 1];<br><br><br>______________________________<wbr>_________________<br>cfe-commits mailing list<br><a href="mailto:cfe-commits@lists.llvm.org" target="_blank">cfe-commits@lists.llvm.org</a><br><a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/cfe-commits</a><br></blockquote></div><br></div></div></div></blockquote></div></div></div><br></div><br>______________________________<wbr>_________________<br>cfe-commits mailing list<br><a href="mailto:cfe-commits@lists.llvm.org" target="_blank">cfe-commits@lists.llvm.org</a><br><a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/cfe-commits</a></blockquote></div></div></div></div></blockquote></div></div></div><br></div><br>______________________________<wbr>_________________<br>
cfe-commits mailing list<br>
<a href="mailto:cfe-commits@lists.llvm.org">cfe-commits@lists.llvm.org</a><br>
<a href="http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits" rel="noreferrer" target="_blank">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/cfe-commits</a><br>
<br></blockquote></div><br></div></div>