4490c4aaedChange size of reference count field in __libcpp_nmstr from 32 bits to 64 bits for 64 bit targets. This is controls the data layout of all exceptions defined in <stdexcept>. This aligns the ABI with that of gcc-4.2.
Howard Hinnant
2012-08-08 16:17:31 +0000
584db4287bstd::equal operating on non-const __bit_iterators was not working. This fixes it.
Howard Hinnant
2012-08-05 21:43:11 +0000
e103a3d69bAndrew Morrow: The current CMake setup for libc++ incorrectly uses the variable LLVM_ENABLE_ASSERTIONS instead of LIBCXX_ENABLE_ASSERTIONS when figuring out what _DEBUG/NDEBUG defines to set. It also tries to test the non-existent variable 'uppercase_CMAKE_BUILD_TYPE', which the top level LLVM CMakeLists.txt sets up, but which the top level libc++ CMakeLists.txt currently does not. Changing the variable name tested and creating the uppercase release name variable allows libc++ to honor the LIBCXX_ENABLE_ASSERTIONS option correctly.
Howard Hinnant
2012-08-05 17:37:39 +0000
6886dd19c4Loosen up the timing requirements on 4 more tests.
Howard Hinnant
2012-08-04 00:47:42 +0000
4b2f4203a2Performance tweaking rotate.
Howard Hinnant
2012-08-03 18:01:20 +0000
cd99236231Andrew Morrow: The attached patch updates the initialization of the 'struct tm' in __time_get_storage<char> to match the initialization behavior in __time_get_storage<wchar>. Without the initialization, valgrind reports errors in the subsequent calls to strftime_l.
Howard Hinnant
2012-08-02 18:44:17 +0000
069bdd52c1Andrew Morrow: There are two tests under test/utilities/memory that heap allocate two integers which remain unused and are subsequently leaked, so the test fail when run under valgrind. Unless I'm overlooking a subtle reason why they are needed I think they can be removed, allowing these tests to pass under valgrind. The attached patch removes the variables. If there is a reason for them to exist, I can change this to just delete them at the end of the test.
Howard Hinnant
2012-08-02 18:39:48 +0000
63b2f4f2dcAndrew Morrow: The attached patch updates the lit.config for libc++ unit tests so that the valgrind configuration passed to lit.py is used to run .pass tests.
Howard Hinnant
2012-08-02 18:36:47 +0000
ef793f2513Andrew Morrow: Among the various libc++ tests that currently don't pass on Linux are localization/locale.categories/category.collate/category.ctype/locale.ctype.byname/is_1.pass.cpp and scan_is.pass.cpp. The tests fail when the character class being tested is compound, like ctype_base::alnum or ctype_base::graph, because the existing series of conditionals in do_is an do_scan_is will abort too early. For instance, if the character class being tested is alnum, and the character is numeric, do_is will return false because iswalpha_l will return false, 'result' becomes false, and the 'true' result from the later call to iswdigit_l ends up being ignored . A similar problem exists in do_scan_is.
Howard Hinnant
2012-08-02 18:35:07 +0000
403f91ad2cAndrew Morrow: The attached patch is an attempt to implement std:🧵:hardware_concurrency for platforms that don't offer sysctl, but do provide a POSIX sysconf and _SC_NPROCESSORS_ONLN.
Howard Hinnant
2012-08-02 18:17:49 +0000
6d39f9f356Andrew Morrow: This patch fixes test/input.output/iostream.format/output.streams/ostream.formatted/ostream.inserters.arithmetic/pointer.pass.cpp to accept '(nil)' as a valid representation for NULL so that the test passes on Linux. The same thing is already done in some other tests, like in /test/localization/locale.categories/category.numeric/locale.nm.put/facet.num.put.members/put_pointer.pass.cpp.
Howard Hinnant
2012-08-02 18:12:06 +0000
ee7a0bf265Andrew Morrow: Attached is a writeup of the current state of the libc++ test suite on Linux.
Howard Hinnant
2012-07-31 21:30:28 +0000
65f059b842Despite my pathological distrust of spin locks, the number just don't lie. I've put a small spin in __sp_mut::lock() on std::mutex::try_lock(), which is testing quite well. In my experience, putting in a yield for every failed iteration is also a major performance booster. This change makes one of the performance tests I was using (a highly contended one) run about 20 times faster.
Howard Hinnant
2012-07-30 17:13:21 +0000
7a7b6d8283Updated status
Howard Hinnant
2012-07-30 13:59:36 +0000
30055c68d4Updated the complete by-chapter graph
Howard Hinnant
2012-07-30 02:29:34 +0000
5fec82dc0dImplement [util.smartptr.shared.atomic]. This is the last unimplemented section in libc++. This requires a recompiled dylib. Failure to rebuild the dylib will result in a link-time error if and only if the functions from [util.smartptr.shared.atomic] are used.
Howard Hinnant
2012-07-30 01:40:57 +0000
116ce6a312Update CREDITS.TXT
Howard Hinnant
2012-07-26 20:22:37 +0000
afcac1ac46Patch by Andrew C. Morrow: shims to work around macroized getc and putc on linux. On my eglibc 2.13 based Debian system 'getc' is a macro defined in /usr/include/stdio.h. This decision to make it a macro doesn't seem to be guarded by any feature test macro as far as I can see.
Howard Hinnant
2012-07-26 20:01:13 +0000
8b5bb3c5afPatch by Andrew C. Morrow: Conditionally include cxxabi.h in new.cpp and typeinfo.cpp. Both new.cpp and typeinfo.cpp have code that is conditionally compiled based on the LIBCXXRT and _LIBCPPABI_VERSION defines, but those files do not currently include <cxxabi.h> in the non __APPLE__ case. The attached patch updates those files so that for non __APPLE__ builds <cxxabi.h> is included if available or if LIBCXXRT is set. I'm modeling this on the recent updates to exception.cpp.
Howard Hinnant
2012-07-26 17:42:39 +0000
ca8eb830dd<algorithm> no longer needs to include <cstdlib>, but can get away with just <cstddef>. This was brought to my attention by Salvatore Benedetto in his port to a bare-metal coretex-m3. This exposed two test bugs where an explicit #include <cstdlib> was needed.
Howard Hinnant
2012-07-26 17:09:09 +0000
f3d62ea57flocale::id really needs to be constructed at compile time.
Howard Hinnant
2012-07-26 16:14:37 +0000
0405cc4ae0libc++: switch from using _ATTRIBUTE(noreturn) (which conflicts with a platform-provided macro on some systems) to _LIBCPP_NORETURN.
Richard Smith
2012-07-26 02:04:22 +0000
8131a01a9cApple LWG 2067: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3318.html#2067 . This is the only actionable change that has been made to the C++ draft since C++11. In general it has not been decided exactly how libc++ will track changes made to C++11. New features and design changes will probably be #ifdef'd, especially if they are not backwards compatible. Defects and 'dumb mistakes' are more likely to just be put in. Decisions on telling one from the other will be made on a case by case basis.
Howard Hinnant
2012-07-21 19:34:12 +0000
8bf01ddd30noexcept applied to <future>.
Howard Hinnant
2012-07-21 17:46:55 +0000
6e1d851be8noexcept applied to <thread>.
Howard Hinnant
2012-07-21 16:50:47 +0000
c8f7413908noexcept applied to <condition_variable>.
Howard Hinnant
2012-07-21 16:32:53 +0000
499c61f999noexcept and constexpr applied to <mutex>.
Howard Hinnant
2012-07-21 16:13:09 +0000
46623a09eenoexcept and constexpr applied to <regex>.
Howard Hinnant
2012-07-21 01:31:58 +0000
f57bd564fdnoexcept and constexpr applied to <ios>.
Howard Hinnant
2012-07-21 01:03:40 +0000
bd143086acnoexcept applied to <valarray>.
Howard Hinnant
2012-07-21 00:51:28 +0000
410f2def47constexpr applied to <complex>.
Howard Hinnant
2012-07-20 22:18:27 +0000
c83960a9e4noexcept applied to <random>.
Howard Hinnant
2012-07-20 21:44:27 +0000
fe4c9dd747Relax the tolerances on some timing tests.
Howard Hinnant
2012-07-20 19:48:05 +0000
d06a640ba7noexcept applied to <iterator>.
Howard Hinnant
2012-07-20 19:36:34 +0000
08bce1754dconstexpr applied to <array>.
Howard Hinnant
2012-07-20 19:20:49 +0000
03d7181b0econstexpr applied to <string>.
Howard Hinnant
2012-07-20 19:09:12 +0000
e41f475a44Further tweaks on relaxing complete type checking for function.
Howard Hinnant
2012-07-20 18:56:07 +0000
7d87f6be1bJean-Daniel : clang now supports all required type_traits.
Howard Hinnant
2012-07-19 15:59:52 +0000
af3473404eJean-Daniel updates the libc++ index page to reflect not so recent changes in C++ standard status.
Howard Hinnant
2012-07-19 15:57:51 +0000
c425307238Relax the complete-type checks that are happening under __invokable<Fp, Args...> to only check Fp, and not Args... . This should be sufficient to give the desired high quality diagnostics under both bind and function. And this allows a test reported by Rich E on cfe-dev to pass. Tracked by <rdar://problem/11880602>.
Howard Hinnant
2012-07-16 16:17:34 +0000
473f838128Applied constexpr to <chrono>.
Howard Hinnant
2012-07-13 19:17:27 +0000
1ca23672a0Fixed a bug in wstring_convert concerning zero-length inputs. Thanks to Jonathan Coxhead for reporting this bug.
Howard Hinnant
2012-07-12 18:07:41 +0000
591e32d624Teach libc++ to check for libc++abi and use its features if they're available.
Richard Smith
2012-07-11 09:35:47 +0000
d586248597Add test for self-referencing emplace test.
Howard Hinnant
2012-07-09 02:47:43 +0000
a58402abb9Change emplace for vector and deque to create the temporary (when necessary) before any changes to the container are made. Nikolay Ivchenkov deserves the credit for pushing this problem and the solution for it.
Howard Hinnant
2012-07-08 23:23:04 +0000
46e9493c68Appy constexpr to <memory>. Picked up a few missing noexcepts as well.
Howard Hinnant
2012-07-07 20:56:04 +0000
384608e90dApply constexpr to the mutex constructor. As a conforming extension, apply constexpr to the condition_variable constructor. These are important because it enables the compiler to construct these types at compile time, even though the object will be non-const. Since they are constructed at compile time, there is no chance of a data race before they are constructed.
Howard Hinnant
2012-07-07 20:01:52 +0000
90d8723476Apply constexpr to <bitset>.
Howard Hinnant
2012-07-07 17:04:52 +0000
74f26f251bApply noexcept to tuple.
Howard Hinnant
2012-07-06 21:53:48 +0000
4eebfc3394As a conforming extension give tuple a noexcept default constructor conditionalized on its held types.
Howard Hinnant
2012-07-06 20:50:27 +0000
5394c1ed30Give tuple a constexpr default constructor.
Howard Hinnant
2012-07-06 20:39:45 +0000
a0b5befbd3New Windows libc++ test results provided by Ruben Van Boxem.
Howard Hinnant
2012-07-06 19:35:31 +0000
9b12f23e30Apply noexcept to those functions implemented in <cstdlib> as a conforming extension.
Howard Hinnant
2012-07-06 19:16:56 +0000
cac0c46abbApply noexcept to those functions implemented in <cmath> as a conforming extension.
Howard Hinnant
2012-07-06 19:13:50 +0000
71499ad176Add noexcept test for offsetof macro per [support.types]/p4.
Howard Hinnant
2012-07-06 18:39:01 +0000
7a44515588This commit establishes a new bucket_count policy in the unordered containers: The policy now allows a power-of-2 number of buckets to be requested (and that request honored) by the client. And if the number of buckets is set to a power of 2, then the constraint of the hash to the number of buckets uses & instead of %. If the client does not specify a number of buckets, then the policy remains unchanged: a prime number of buckets is selected. The growth policy is that the number of buckets is roughly doubled when needed. While growing, either the prime, or the power-of-2 strategy will be preserved. There is a small run time cost for putting in this switch. For very cheap hash functions, e.g. identity for int, the cost can be as high as 18%. However with more typical use cases, e.g. strings, the cost is in the noise level. I've measured cases with very cheap hash functions (int) that using a power-of-2 number of buckets can make look up about twice as fast. However I've also noted that a power-of-2 number of buckets is more susceptible to accidental catastrophic collisions. Though I've also noted that accidental catastrophic collisions are also possible when using a prime number of buckets (but seems far less likely). In short, this patch adds an extra tuning knob for those clients trying to get the last bit of performance squeezed out of their hash containers. Casual users of the hash containers will not notice the introduction of this tuning knob. Those clients who swear by power-of-2 hash containers can now opt-in to that strategy. Clients who prefer a prime number of buckets can continue as they have.
Howard Hinnant
2012-07-06 17:31:14 +0000
820e00755alink to Marshall's notes.
Howard Hinnant
2012-07-06 00:39:38 +0000
2d62229b96Fixed a bug regarding result_of reported by Sven Behne. The fix is C++11 only mainly because result_of is a variadic beast and working with variadics is just such a problem in C++03 mode. This should bring result_of up to full conformance with the C++11 spec.
Howard Hinnant
2012-06-26 17:37:15 +0000
67c8082a18fix help with bash
Nuno Lopes
2012-06-25 23:51:05 +0000
fe59276f04Revert pair constructors back to using is_convertible instead of is_constructible. This should pull things into alignment with the final draft. Fixes http://llvm.org/bugs/show_bug.cgi?id=13063#add_comment.
Howard Hinnant
2012-06-09 20:01:23 +0000
f9b6e7eeefFix warning flags for CMake builds, from Andrew C. Morrow!
Douglas Gregor
2012-06-07 22:26:00 +0000
caee2b093fFix a few testsuite bugs involving trailing null (or lack thereof) in strstream.
Howard Hinnant
2012-06-01 20:02:59 +0000
dbd9eacde0Fix dangling else clause. Bug found and fixed by Dimitry Andric.
Howard Hinnant
2012-05-31 23:12:03 +0000
6467aeb7c9Fix the new _ALIGNAS_TYPE per instructions supplied by Eli Friedman.
Howard Hinnant
2012-05-31 20:14:00 +0000
cbdd0896d3Protect use of alignas against older versions of clang
Howard Hinnant
2012-05-31 19:31:14 +0000
635ce1d127The rules for emplace in map, multimap, unordered_map and unordered_multimap changed a while back and I'm just now updating to these new rules. In a nutshell, you've got to know you're emplacing to a pair and use one of pair's constructors. I made one extension: If you want to emplace the key and default construct the mapped_type, you can just emplace(key), as opposed to emplace(piecewise_construct, forward_as_tuple(key), forward_as_tuple()).
Howard Hinnant
2012-05-25 22:04:21 +0000
3e3e5ebc72Fix memory corruption bug found and fixed by Andrew C. Morrow.
Howard Hinnant
2012-05-25 15:55:46 +0000
51065657aaAdd documentation regarding -fno-rtti.
Howard Hinnant
2012-05-20 13:03:53 +0000
0855ddeb24Revert my _LIBCPP_INLINE_VISIBILITY changes, r157097 and r157107
Douglas Gregor
2012-05-19 07:14:17 +0000
f20f0d3fc5valarray resize should not be _LIBCPP_INLINE_VISIBILITY
Douglas Gregor
2012-05-19 07:01:14 +0000
e9e4b855b8Move _LIBCPP_VISIBLE_INLINE from the out-of-line definitions of member functions to the original declarations, so that Clang will actually see them. Part of <rdar://problem/11489333>.
Douglas Gregor
2012-05-19 04:41:25 +0000
d2da6d2322Constrain __bind functor constructor such that it won't accidentally get used as a copy constructor from a non-const lvalue. Fixes <rdar://problem/11359080>.
Howard Hinnant
2012-05-04 17:21:02 +0000
f07a529b77Change std::abs from a template function to three overloads for float, double and long double.
Howard Hinnant
2012-05-03 14:58:34 +0000
87073e4bfbGreatly scale back ambitions of emulating move semantics in C++03 mode. It was causing more problems than it solved. This fixes http://llvm.org/bugs/show_bug.cgi?id=12704.
Howard Hinnant
2012-05-01 15:37:54 +0000
c756f5b4e8libc++: only #include <cxxabi.h> if it exists. This allows libc++ to build out of the box on Linux systems. If you're building against libc++abi, you still need to make sure it can find <cxxabi.h> so it knows not to export symbols which libc++abi provides.
Richard Smith
2012-04-19 01:36:12 +0000
9efdc0bd5flibc++: Add some missing #includes to atomics tests. libc++ doesn't need these at the moment, but they allow these tests to be used to test clang against libstdc++. Add myself to the credits file, as suggested by Howard.
Richard Smith
2012-04-19 00:50:47 +0000
300c67ab92Apply noexcept and constexpr to <atomic>.
Howard Hinnant
2012-04-11 20:14:21 +0000
6186c7fe6aSwitch libc++ from __atomic_* builtins to __c11_atomic_* builtins. Per discussion with Howard, we are not interested in maintaining compatibility with older versions of clang.
Richard Smith
2012-04-11 18:55:46 +0000
b2292091cbNow that clang supports doing the right thing with regard to atomic initialisation, do the right thing with regard to atomic initialisation.
David Chisnall
2012-04-11 17:26:23 +0000
d3eca759a2Fix the remaining atomic tests, all of which were wrong for the case where a compare-and-exchange failed (it should update the expected value to the current value, and the tests were checking that it didn't...).
David Chisnall
2012-04-05 13:48:16 +0000
1ee87fa41aFix test cases that were trying to make atomic things that are not trivially copyable.
David Chisnall
2012-04-05 13:23:08 +0000
0341c820a0Fix use of __atomic_is_lock_free() intrinsic.
David Chisnall
2012-04-05 13:13:24 +0000
616e92d748Put std::piecewise_construct_t back into the dylib for ABI stability. When clients are in C++11/constexpr mode this will be safely ignored because piecewise_construct is then declared with internal linkage.
Howard Hinnant
2012-04-03 23:45:46 +0000
2a5349ba66constexpr support for <utility>. Patch contributed by Jonathan Sauer.
Howard Hinnant
2012-04-03 21:09:48 +0000
271a1c03c3Updated documentation contributed by Christopher Jefferson.
Howard Hinnant
2012-04-03 15:08:42 +0000
8efd3dac5dUpdate <random> with constexpr support. Patch contributed by Jonathan Sauer.
Howard Hinnant
2012-04-02 21:00:45 +0000
60cb7d267fUpdate <limits> with constexpr support. Patch contributed by Jonathan Sauer.
Howard Hinnant
2012-04-02 19:23:15 +0000
27b4fd30efThis is an initial commit of constexpr support as proposed by Richard Smith. This by no means completes constexpr support. Indeed, it hardly scratches the surface. All it does is lay the foundation in <__config> and changes those few places in the library that are already using that foundation.
Howard Hinnant
2012-04-02 00:40:41 +0000
dc1345fd44I believe tuple is still under development in the standard. Daniel Krugler is/will be making convincing arguments that a modified form of LWG 2051 (currently NAD Future) is easily acheivable and desirable. He has demonstrated that a tuple<T...> where all of the T are implicitly convertible from U... should have a tuple constructor that is also implicit, instead of explicit. This would support the use cases in LWG 2051 while not undermining T... with explicit conversions from U.... This check-in is an experimental implementation of Daniel's work. I believe this work to be mature enough to warrant inclusion into libc++. If anyone sees real-world problems that this check in causes, please let me know and I will revert it, and provide the feedback to the LWG.
Howard Hinnant
2012-04-01 23:10:42 +0000
9aa4e11451It appears that the standard accidentally removed the default constructor for error_category. I'm putting it back in. This fixes http://llvm.org/bugs/show_bug.cgi?id=12321.
Howard Hinnant
2012-03-21 16:18:57 +0000
87c61a61feAllow libc++ to be built with CMake from within the LLVM tree. The libc++ part is just some renaming as the variable was already in use, conflicting with something else in the LLVM tree. Contributed by Ruben Van Boxem.
Howard Hinnant
2012-03-19 15:40:23 +0000
1378397721Alter the terminal streams such that they do not get added to the atexit chain, and thus never get destructed.
Howard Hinnant
2012-03-16 15:13:51 +0000
Howard Hinnant
2012-08-13 12:29:17 +0000