Commit Graph

778 Commits

Author SHA1 Message Date
Marshall Clow
028875aa7c Patch by Xing Xue to improve libc++ support for AIX
llvm-svn: 195144
2013-11-19 19:16:03 +00:00
Marshall Clow
7e1ea8d288 Patch by Bruce Mitchener. Change all references to EMSCRIPTEN to __EMSCRIPTEN__. If you're not using the PP symbol EMSCRIPTEN, then you should see no functionality change.
llvm-svn: 195136
2013-11-19 18:05:03 +00:00
Yaron Keren
fbeb63c0d1 This patch implements snprintf_l function in a way similar to the other
functions in src/support/win32/locale_win32.cpp and locale_win32.h, 
calling upon vsnprintf for which there is a MingW correct alternative.

Note! __USE_MINGW_ANSI_STDIO is not modified in this patch. In order to 
use the __mingw version it must be defined before including the MingW 
headers.

llvm-svn: 195044
2013-11-18 21:12:14 +00:00
Yaron Keren
c0299d7f53 Fix-it suggestion for fixing min or max defines on Windows.
llvm-svn: 194891
2013-11-15 23:41:01 +00:00
Yaron Keren
6e72b8ba8c Windows.h is not required.
llvm-svn: 194870
2013-11-15 22:54:15 +00:00
Marshall Clow
dfdac03c8f Move <optional> into include/experimental, and into the std::experimental namespace, since it's not part of C++14, but of an upcoming TS
llvm-svn: 194867
2013-11-15 22:42:10 +00:00
Justin Bogner
fd6cfe92d3 Add the CMakeLists.txt that was missed in r194825
llvm-svn: 194838
2013-11-15 18:34:43 +00:00
Howard Hinnant
f9fd0d6d11 This is a followup to r194536, which changed the pair copy constructor to be
trivial in C++03, thus making it trivial in both C++03 and C++11.

This patch allows one to opt-in/out of this decision with a macro.  You can
choose to have the pair copy constructor always be trivial, or always be
non-trivial.  The flag controlling this is now _LIBCPP_TRIVIAL_PAIR_COPY_CTOR.

The client can define this flag to 1, and the pair copy constructor will be
trivial (when possible of course), or to 0, and the pair copy constructor will
be nontrivial.

Default settings for this flag are set in <__config> (as usual).  With this
commit the default is _LIBCPP_TRIVIAL_PAIR_COPY_CTOR=1 for all platforms
except __APPLE__, which defaults to _LIBCPP_TRIVIAL_PAIR_COPY_CTOR=0.

llvm-svn: 194742
2013-11-14 22:52:25 +00:00
Marshall Clow
513ecaba91 Fixed bug in quoted strings implementation. Added test to be sure. Thanks to Peter Sommerlad for the report (and suggested fix)
llvm-svn: 194725
2013-11-14 20:01:38 +00:00
Marshall Clow
21fee96f69 Move <dynarray> into include/experimental, and into the std::experimental namespace, since it's not part of C++14, but of an upcoming TS
llvm-svn: 194614
2013-11-13 22:44:48 +00:00
Howard Hinnant
ccad8c32e0 This fixes a very subtle ABI problem concerning the copy constructor of
pair, and a couple of pair-like implementation detail types.  The
C++98/03 and 11 standards all specify that the copy constructor of
pair<int, int> is trivial. However as libc++ tracked the draft C++11
standard over the years, this copy constructor became non-trivial, and
then just recently was corrected back to trivial for C++11.

Unfortunately (for libc++1) the Itanium ABI specifies different calling
conventions for trivial and non-trivial copy constructors.  Therefore
currently the C++03 libc++ copy constructor for pair<int, int> is ABI
incompatible with the C++11 libc++ copy constructor for pair<int, int>.
This is Bad(tm).   This patch corrects the situation by making this copy
constructor trivial in C++03 mode as well.

Just in case it is needed for an incomplete C++11 compiler, libc++
retains the ability to support pair with rvalue references, but without
defaulted special members.  However the pair needs non-trivial special
members to implement this special case, (as it did when clang was in
this place a couple of years ago).

During this work a bug was also found and fixed in
is_trivially_constructible.

And there is a minor drive-by fix in <__config> regarding
__type_visibility__.

A test is updated to ensure that the copy constructor of pair<int, int>
is trivial in both C++03 and C++11.  This test will necessarily fail for
a compiler that implements rvalue references but not defaulted special
members.

llvm-svn: 194536
2013-11-13 00:39:22 +00:00
Howard Hinnant
88960d151e Tell libc++abi whether or not libc++ has declared bad_array_length.
llvm-svn: 194207
2013-11-07 17:15:51 +00:00
Marshall Clow
57b8f44c87 More duplicate code removal in <locale>. Hoist common parsing code into two templates: num_get::__do_get_signed and num_get::__do_get_unsigned, and make the do_get routines call them. No functionality change.
llvm-svn: 194185
2013-11-07 01:00:50 +00:00
Howard Hinnant
4478b25ade Fix several tuple bugs that were exposed by clang's implementation of CWG 1402. This fixes http://llvm.org/bugs/show_bug.cgi?id=17798.
llvm-svn: 194154
2013-11-06 17:45:43 +00:00
Marshall Clow
e427322327 Fix an off-by-one error in basic_string::__grow_by, where it would incorrectly throw length_error (instead of bad_alloc) when attempting to resize the string to 'max_size()'. Add tests for resizing to max_size +/-1
llvm-svn: 194151
2013-11-06 14:24:38 +00:00
Marshall Clow
9d66b72aa9 Refactor floating point code for num_get::do_get into a template. No functionality change
llvm-svn: 194080
2013-11-05 14:28:52 +00:00
Marshall Clow
99c096472d Better inline marking for __does_policy_contain. Thanks to Chongyu Zhu for the catch
llvm-svn: 193963
2013-11-03 22:06:53 +00:00
Marshall Clow
222d1c7f13 Mark __does_policy_contain as 'inline'. Thanks to Chongyu Zhu for the catch
llvm-svn: 193962
2013-11-03 20:07:47 +00:00
Marshall Clow
b1915875d0 Fix LWG Issue 2078. Make std::async(policy,...) try multiple policies until one succeeds.
llvm-svn: 193960
2013-11-03 15:43:35 +00:00
Marshall Clow
e1bedf4e93 LWG issue 2341; Make the two variants of basic_ostream::seekp and basic_istream::seekg behave consistently; update tests to make sure
llvm-svn: 193814
2013-10-31 22:20:45 +00:00
Marshall Clow
f5fa53882f Fixes PR17148
llvm-svn: 193772
2013-10-31 17:23:08 +00:00
Marshall Clow
07c28fe026 Mark seed_seq default constructor and size() as noexcept. This is implied, but not required by LWG issue 2180
llvm-svn: 193227
2013-10-23 05:56:47 +00:00
Marshall Clow
e604469e5c Patch by GM: apparently '__value' (two underscores) is a special name in Visual Studio, so rename the private method in <regex> with that name. GM's patch used '___value' (three underscores), but I changed that to '__regex_traits_value' because I've been burned in the past by identifiers that appear identical but are not.
llvm-svn: 193087
2013-10-21 15:43:25 +00:00
Marshall Clow
5b40666c6c Patch by GM: Making implicit conversion to bool explicit in <ios> and <__locale>
llvm-svn: 193085
2013-10-21 14:41:05 +00:00
Marshall Clow
d8cfc7dce9 Patch by GM: Adding MSVC support to __bit_reference
llvm-svn: 193084
2013-10-21 14:29:37 +00:00
Marshall Clow
9f21325ac7 Patch from GM to make more implicit bools explicit since we can't stop MSVC warning about this in headers and to warn is the MSVC default. No functionality change.
llvm-svn: 192548
2013-10-13 01:02:45 +00:00
Marshall Clow
c3776b1ce0 LWG Issue 2087: iostream_category() and noexcept
llvm-svn: 192545
2013-10-12 22:49:56 +00:00
Marshall Clow
d58daf9433 LWG Issue 2097: packaged_task constructors should be constrained
llvm-svn: 192544
2013-10-12 22:49:17 +00:00
Marshall Clow
f28fd284f8 LWG issue 2143: ios_base::xalloc should be thread-safe
llvm-svn: 192539
2013-10-12 19:13:52 +00:00
Marshall Clow
8de32cb3dc Implement national body comment GB9: remove std::gets
llvm-svn: 192538
2013-10-12 19:09:47 +00:00
Marshall Clow
0354b92992 patch by Yaron: Uses rand_s() from stdlib.h (when building for Windows)
llvm-svn: 192325
2013-10-09 21:49:03 +00:00
Marshall Clow
db78c7049e Fix LWG Issue 2141: common_type trait produces reference types
llvm-svn: 192142
2013-10-07 23:43:33 +00:00
Howard Hinnant
04c63bdb36 G M: Rename local variable from __except to avoid MSVC keyword clash.
llvm-svn: 192072
2013-10-06 21:00:29 +00:00
Howard Hinnant
32f9a776fb Use _LIBCPP_NEW_DELETE_VIS instead of LIBCPP_FUNC_VIS in src/new.cpp.
llvm-svn: 192071
2013-10-06 20:53:24 +00:00
Howard Hinnant
f66a81ad56 Yaron Keren: Add missing comment.
llvm-svn: 192068
2013-10-06 19:48:40 +00:00
Marshall Clow
dbd2ebb07e LWG Issue 2247: Implement type trait 'is_null_pointer'
llvm-svn: 192049
2013-10-05 21:21:17 +00:00
Marshall Clow
ea7c7cc521 Implement literal suffixes for compled
llvm-svn: 192048
2013-10-05 21:19:49 +00:00
Marshall Clow
3ceafc7f01 Mark namespaces for user defined literals as 'inline'
llvm-svn: 192047
2013-10-05 21:18:32 +00:00
Marshall Clow
de9320aa2b Implement LWG issue 2275 'forward_as_tuple should be constexpr'
llvm-svn: 192038
2013-10-05 18:46:37 +00:00
Howard Hinnant
a942f2ffd7 G M: The attached patch is for libcxx's new.cpp and __config files. The patch's intent is to make new.cpp compile using MS's cl.exe compiler without changing the meaning of anything for any other compiler.
The issue this patch seeks to address is that MS's compiler (cl.exe) doesn't support the __attribute__((__weak__)) or __atribute__((__visibility__("default")) syntax; so a solution must be found where cl.exe doesn't see this syntax.

This patch seeks to solve this problem by changing code patterned like this:
__attribute__((__weak__, __visibility__("default")))
void* operator new(size_t size, const std::nothrow_t&) _NOEXCEPT { /*snip*/; return p; }

to code like this:
_LIBCPP_WEAK
void* operator new(size_t size, const std::nothrow_t&) _NOEXCEPT { return p; }

Howard:  Thanks for all the comments regarding the default visibility
tag on the definition.  I agree it isn't needed, and that there are lots
of other places where it is missing.  That being said, I'm not wanting
to rock the boat on that issue right now.  So I've added it back to the
definition via _LIBCPP_FUNC_VIS.  A later pass dedicated just to this
issue can bring things in to a consistent state one way or the other. 
Note that we do not want to have the exact same attributes on the
declaration and defintion in this case.  The declaration should not be
marked weak, whereas the definition should (which is what G M's patch
did). I've fully tested on OS X to ensure that the resultant attribute
syntax actually works.

llvm-svn: 192007
2013-10-04 23:56:37 +00:00
Howard Hinnant
3af48ef76e G M: Changes all references to "x inline" to "inline x" where x = _libcpp_always_inline or _libcpp_inline_visibility macros.
The patch touches these files:

locale
array
deque
new
string
utility
vector
__bit_reference
__split_buffer
locale_win32.h
 
There is no intended functionality change and it is expected that reversing the position of the inline keyword with regard to the other keywords does not change the meaning of anything, least not for apple/Linux etc.
 
It is intended to make libcxx more consistent with itself and to prevent the 1000 or so
"inline.cpp(3) : warning C4141: 'inline' : used more than once" warnings that MS's cl.exe compiler emits without this patch, i.e. if inline is not the first keyword before a function name etc.
 
Prefer "inline [other inline related keyword]" over "[other related keyword] inline".
After this patch, libcxx should be consistent to this pattern.

llvm-svn: 191987
2013-10-04 22:09:00 +00:00
Howard Hinnant
f7a8c4f347 G M: Fix libcxx's detection of rtti disablement for g++.exe and cl.exe. When RTTI is NOT enabled, _LIBCPP_NO_RTTI is defined.
llvm-svn: 191981
2013-10-04 21:24:21 +00:00
Howard Hinnant
80b84d4c26 G M: Provides the _LIBCPP_WARNING macro, to be used for MSVC only, since that compiler doesn't support #warning.
llvm-svn: 191980
2013-10-04 21:14:44 +00:00
Marshall Clow
ad864049d6 Remove non-printable chars that snuck in back in July; thanks to Yaron Keren for the catch
llvm-svn: 191756
2013-10-01 13:28:20 +00:00
Marshall Clow
45b983c4d0 Part 8 of LWG Issue 2210' unordered_set and unordered multiset; this got missed when I went on vacation
llvm-svn: 191705
2013-09-30 21:33:51 +00:00
Howard Hinnant
9fd9f84f48 SCARY/N2913 iterator support between the multi and non-multi versions of the associative and unordered containers. I beleive lack of support for this was accidentally recently introduced (by me) and this is fixing a regression. This time tests are put in to prevent such a regression in the future.
llvm-svn: 191692
2013-09-30 19:08:22 +00:00
Marshall Clow
3d5134dd52 Implement n3789; constexpr support in named function objects
llvm-svn: 191626
2013-09-28 19:06:12 +00:00
Marshall Clow
8392ab209e Add a set of macros for using the new '[[[deprecated]]' attribute. No code changes (yet)
llvm-svn: 191625
2013-09-28 18:35:31 +00:00
Howard Hinnant
7a828034c6 Peter Collingbourne: If a pointer is passed as the third argument of the (iterator,
iterator, allocator) constructor with the intention of it being
implicitly converted to the allocator type, it is possible for overload
resolution to favour the (iterator, iterator, enable_if) constructor.
Eliminate this possibility by moving the enable_if to one of the
existing arguments and removing the third argument.

llvm-svn: 191145
2013-09-21 21:13:54 +00:00
Howard Hinnant
ff29d538ea Apply LWG 2048. It is amazing to me that this actually works, but the existing tests confirm that it does. c++1y status page now showing libc++ is complete for c++1y modulo dynarray issues.
llvm-svn: 191142
2013-09-21 19:25:37 +00:00