feedc0de/boost_conversion
1
0
Fork 0
forked from boostorg/conversion
Code Pull Requests Activity

Do not force PGI compiler to use fallback version of lexical_cast (PGI fails tests even in fallback version)

Browse Source 
Trialing whitespaces removed

[SVN r78521]
This commit is contained in:
Antony Polukhin
2012-05-20 16:25:21 +00:00
parent e691b15b77
commit c6318e3819
11 changed files with 66 additions and 66 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
doc/Jamfile.v2
View File

@@ -6,9 +6,9 @@ using quickbook ;
import boostbook : boostbook ;
xml lexical_cast : lexical_cast.qbk ;
boostbook standalone
:
lexical_cast
boostbook standalone
:
lexical_cast
:
<xsl:param>boost.root=../../../..
<format>pdf:<xsl:param>boost.url.prefix=http://www.boost.org/doc/libs/release/doc/html

62
doc/lexical_cast.qbk
View File

@@ -34,7 +34,7 @@ The standard C++ library offers `stringstream` for the kind of in-core formattin
The `lexical_cast` function template offers a convenient and consistent form for supporting common conversions to and from arbitrary types when they are represented as text. The simplification it offers is in expression-level convenience for such conversions. For more involved conversions, such as where precision or formatting need tighter control than is offered by the default behavior of `lexical_cast`, the conventional `std::stringstream` approach is recommended. Where the conversions are numeric to numeric, __numericcast__ may offer more reasonable behavior than `lexical_cast`.
For a good discussion of the options and issues involved in string-based formatting, including comparison of `stringstream`, `lexical_cast`, and others, see Herb Sutter's article, [@http://www.gotw.ca/publications/mill19.htm The String Formatters of Manor Farm]. Also, take a look at the [link boost_lexical_cast.performance Performance] section.
For a good discussion of the options and issues involved in string-based formatting, including comparison of `stringstream`, `lexical_cast`, and others, see Herb Sutter's article, [@http://www.gotw.ca/publications/mill19.htm The String Formatters of Manor Farm]. Also, take a look at the [link boost_lexical_cast.performance Performance] section.
[endsect]
[section Examples]
@@ -105,8 +105,8 @@ The character type of the underlying stream is assumed to be `char` unless eithe
* `boost::iterator_range<WideCharPtr>`, where `WideCharPtr` is a pointer to wide-character or pointer to const wide-character
[important Many compilers and runtime libraries fail to make conversions using new Unicode characters. Make shure that the following code compiles and outputs nonzero values, before using new types:
``
std::cout
``
std::cout
<< booat::lexical_cast<std::u32string>(1.0).size()
<< " "
<< booat::lexical_cast<std::u16string>(1.0).size();
@@ -151,13 +151,13 @@ Consider the following example:
return data_length;
}
};
inline std::ostream& operator << (std::ostream& ostr, const example_class& rhs) {
return ostr << boost::make_iterator_range(rhs.data(), rhs.data() + rhs.size());
}
``
This is a good generic solution for most use cases.
This is a good generic solution for most use cases.
But we can make it even faster for some performance critical applications. During conversion, we loose speed at:
* `std::ostream` construction (it makes some heap allocations)
@@ -182,35 +182,35 @@ Now `boost::lexical_cast<some_type>(example_class_instance)` conversions won't c
[section Frequently Asked Questions]
* [*Question:] Why does `lexical_cast<int8_t>("127")` throw `bad_lexical_cast`?
* [*Answer:] The type `int8_t` is a `typedef` to `char` or `signed char`. Lexical conversion to these types is simply reading a byte from source but since the source has more than one byte, the exception is thrown.
Please use other integer types such as `int` or `short int`. If bounds checking is important, you can also
call __numericcast__:
* [*Answer:] The type `int8_t` is a `typedef` to `char` or `signed char`. Lexical conversion to these types is simply reading a byte from source but since the source has more than one byte, the exception is thrown.
Please use other integer types such as `int` or `short int`. If bounds checking is important, you can also
call __numericcast__:
`numeric_cast<int8_t>(lexical_cast<int>("127"));`
[pre
]
* [*Question:] Why does `lexical_cast<unsigned char>("127")` throw `bad_lexical_cast`?
* [*Answer:] Lexical conversion to any char type is simply reading a byte from source. But since the source has more than one byte, the exception is thrown.
Please use other integer types such as `int` or `short int`. If bounds checking is important, you can also
call __numericcast__:
* [*Answer:] Lexical conversion to any char type is simply reading a byte from source. But since the source has more than one byte, the exception is thrown.
Please use other integer types such as `int` or `short int`. If bounds checking is important, you can also
call __numericcast__:
`numeric_cast<unsigned char>(lexical_cast<int>("127"));`
[pre
]
* [*Question:] What does `lexical_cast<std::string>` of an `int8_t` or `uint8_t` not do what I expect?
* [*Answer:] As above, note that int8_t and uint8_t are actually chars and are formatted as such. To avoid
* [*Answer:] As above, note that int8_t and uint8_t are actually chars and are formatted as such. To avoid
this, cast to an integer type first: `lexical_cast<std::string>(static_cast<int>(n));`
[pre
]
* [*Question:] The implementation always resets the `ios_base::skipws` flag of an underlying stream object.
* [*Question:] The implementation always resets the `ios_base::skipws` flag of an underlying stream object.
It breaks my `operator>>` that works only in presence of this flag. Can you remove code that resets the flag?
* [*Answer:] May be in a future version. There is no requirement in
__proposallong__ to reset the flag but
remember that __proposalshort__ is not yet accepted by the committee. By the way, it's a great opportunity to
* [*Answer:] May be in a future version. There is no requirement in
__proposallong__ to reset the flag but
remember that __proposalshort__ is not yet accepted by the committee. By the way, it's a great opportunity to
make your `operator>>` conform to the standard.
Read a good C++ book, study `std::sentry` and [@boost:libs/io/doc/ios_state.html `ios_state_saver`].
@@ -218,17 +218,17 @@ Read a good C++ book, study `std::sentry` and [@boost:libs/io/doc/ios_state.html
]
* [*Question:] Why `std::cout << boost::lexical_cast<unsigned int>("-1");` does not throw, but outputs 4294967295?
* [*Answer:] `boost::lexical_cast` has the behavior of `std::stringstream`, which uses `num_get` functions of
`std::locale` to convert numbers. If we look at the Programming languages — C++, we'll see, that `num_get` uses
the rules of `scanf` for conversions. And in the C99 standard for unsigned input value minus sign is optional, so
* [*Answer:] `boost::lexical_cast` has the behavior of `std::stringstream`, which uses `num_get` functions of
`std::locale` to convert numbers. If we look at the Programming languages — C++, we'll see, that `num_get` uses
the rules of `scanf` for conversions. And in the C99 standard for unsigned input value minus sign is optional, so
if a negative number is read, no errors will arise and the result will be the two's complement.
[pre
]
* [*Question:] Why `boost::lexical_cast<int>(L'A');` outputs 65 and `boost::lexical_cast<wchar_t>(L"65");` does not throw?
* [*Answer:] If you are using an old version of Visual Studio or compile code with /Zc:wchar_t- flag,
`boost::lexical_cast` sees single `wchar_t` character as `unsigned short`. It is not a `boost::lexical_cast` mistake, but a
* [*Answer:] If you are using an old version of Visual Studio or compile code with /Zc:wchar_t- flag,
`boost::lexical_cast` sees single `wchar_t` character as `unsigned short`. It is not a `boost::lexical_cast` mistake, but a
limitation of compiler options that you use.
[pre
@@ -236,7 +236,7 @@ limitation of compiler options that you use.
* [*Question:] Why `boost::lexical_cast<double>("-1.#IND");` throws `boost::bad_lexical_cast`?
* [*Answer:] `"-1.#IND"` is a compiler extension, that violates standard. You shall input `"-nan"`, `"nan"`, `"inf"`
, `"-inf"` (case insensitive) strings to get NaN and Inf values. `boost::lexical_cast<string>` outputs `"-nan"`, `"nan"`,
, `"-inf"` (case insensitive) strings to get NaN and Inf values. `boost::lexical_cast<string>` outputs `"-nan"`, `"nan"`,
`"inf"`, `"-inf"` strings, when has NaN or Inf input values.
* [*Question:] What is the fastest way to convert a non zero terminated string or a substring using `boost::lexical_cast`?
@@ -246,21 +246,21 @@ limitation of compiler options that you use.
[section Changes]
* [*boost 1.50.0 :]
* `boost::bad_lexical_cast` exception is now globaly visible and can be catched even if code is compiled with -fvisibility=hidden.
* Now it is possible to compile library with disabled exceptions.
* Better performance, less memory usage and bugfixes for `boost::iterator_range<character_type*>` conversions.
* [*boost 1.49.0 :]
* Restored work with typedefed wchar_t (compilation flag /Zc:wchar_t- for Visual Studio).
* Restored work with typedefed wchar_t (compilation flag /Zc:wchar_t- for Visual Studio).
* Better performance and less memory usage for `boost::container::basic_string` conversions.
* [*boost 1.48.0 :]
* Added code to work with Inf and NaN on any platform.
* Added code to work with Inf and NaN on any platform.
* Better performance and less memory usage for conversions to float type (and to double type, if `sizeof(double) < sizeof(long double)`).
* [*boost 1.47.0 :]
* Optimizations for "C" and other locales without number grouping.
@@ -283,7 +283,7 @@ limitation of compiler options that you use.
* The previous version of lexical_cast used the default stream precision for reading and writing floating-point numbers. For numerics that have a corresponding specialization of `std::numeric_limits`, the current version now chooses a precision to match.
* The previous version of lexical_cast did not support conversion to or from any wide-character-based types. For compilers with full language and library support for wide characters, `lexical_cast` now supports conversions from `wchar_t`, `wchar_t *`, and `std::wstring` and to `wchar_t` and `std::wstring`.
* The previous version of `lexical_cast` assumed that the conventional stream extractor operators were sufficient for reading values. However, string I/O is asymmetric, with the result that spaces play the role of I/O separators rather than string content. The current version fixes this error for `std::string` and, where supported, `std::wstring`: `lexical_cast<std::string>("Hello, World")` succeeds instead of failing with a `bad_lexical_cast` exception.
* The previous version of `lexical_cast` allowed unsafe and meaningless conversions to pointers. The current version now throws a `bad_lexical_cast` for conversions to pointers: `lexical_cast<char *>("Goodbye, World")` now throws an exception instead of causing undefined behavior.
* The previous version of `lexical_cast` allowed unsafe and meaningless conversions to pointers. The current version now throws a `bad_lexical_cast` for conversions to pointers: `lexical_cast<char *>("Goodbye, World")` now throws an exception instead of causing undefined behavior.
[endsect]
@@ -328,7 +328,7 @@ All the tests measure execution speed in milliseconds for 10000 iterations of th
``]
]
]
Fastest results are highlitened with "!!! *x* !!!".
Fastest results are highlitened with "!!! *x* !!!".
Do not use this results to compare compilers, because tests were taken on different hardware.
[endsect]

10
perf/Jamfile.v2
View File

@@ -13,17 +13,17 @@ path-constant TEST_DIR : . ;
project performance/test
: source-location ./
: requirements
# <library>/boost/chrono//boost_chrono
: requirements
# <library>/boost/chrono//boost_chrono
# <library>/boost/system//boost_system
<link>static
<target-os>freebsd:<linkflags>"-lrt"
<target-os>freebsd:<linkflags>"-lrt"
<target-os>linux:<linkflags>"-lrt"
<toolset>gcc:<cxxflags>-fvisibility=hidden
<toolset>intel-linux:<cxxflags>-fvisibility=hidden
<toolset>sun:<cxxflags>-xldscope=hidden
: default-build release
;
run performance_test.cpp : $(TEST_DIR) ;

10
test/Jamfile.v2
View File

@@ -11,20 +11,20 @@ import feature ;
project
: requirements
<library>/boost/test//boost_unit_test_framework
<link>static
<link>static
<toolset>gcc-4.8:<define>BOOST_STL_SUPPORTS_NEW_UNICODE_LOCALES
;
# Thanks to Steven Watanabe for helping with <nowchar> feature
# Thanks to Steven Watanabe for helping with <nowchar> feature
feature.feature nowchar : on :
composite optional propagated link-incompatible ;
feature.compose <nowchar>on : <cxxflags>/Zc:wchar_t- ;
test-suite conversion
: [ run implicit_cast.cpp ]
[ compile-fail implicit_cast_fail.cpp ]
[ run ../cast_test.cpp ]
[ run ../numeric_cast_test.cpp ]
[ run ../numeric_cast_test.cpp ]
[ run ../lexical_cast_test.cpp ]
[ run lexical_cast_loopback_test.cpp ]
[ run lexical_cast_abstract_test.cpp ]
@@ -39,7 +39,7 @@ test-suite conversion
[ compile lexical_cast_typedefed_wchar_test.cpp : <toolset>msvc:<nowchar>on ]
[ run lexical_cast_typedefed_wchar_test_runtime.cpp : : : <toolset>msvc:<nowchar>on <toolset>msvc,<stdlib>stlport:<build>no ]
[ run lexical_cast_no_locale_test.cpp : : : <define>BOOST_NO_STD_LOCALE <define>BOOST_LEXICAL_CAST_ASSUME_C_LOCALE ]
[ run lexical_cast_no_exceptions_test.cpp : : : <define>BOOST_NO_EXCEPTIONS
[ run lexical_cast_no_exceptions_test.cpp : : : <define>BOOST_NO_EXCEPTIONS
<toolset>gcc-4.3:<cflags>-fno-exceptions
<toolset>gcc-4.4:<cflags>-fno-exceptions
<toolset>gcc-4.5:<cflags>-fno-exceptions

2
test/implicit_cast.cpp
View File

@@ -28,7 +28,7 @@ int main()
type<foo> f = check_return(boost::implicit_cast<foo>("hello"));
type<long> z = check_return(boost::implicit_cast<long>(foo("hello")));
// warning supression:
(void)x;
(void)f;

2
test/lexical_cast_containers_test.cpp
View File

@@ -31,7 +31,7 @@ boost::unit_test::test_suite *init_unit_test_suite(int, char *[])
}
void testing_boost_containers_basic_string()
{
{
BOOST_CHECK("100" == lexical_cast<boost::container::string>("100"));
BOOST_CHECK(L"100" == lexical_cast<boost::container::wstring>(L"100"));

2
test/lexical_cast_float_types_test.cpp
View File

@@ -237,7 +237,7 @@ void test_converion_to_float_types()
CHECK_CLOSE_ABS_DIFF(-10101.0E-011, test_t);
CHECK_CLOSE_ABS_DIFF(-10101093, test_t);
CHECK_CLOSE_ABS_DIFF(10101093, test_t);
CHECK_CLOSE_ABS_DIFF(-.34, test_t);
CHECK_CLOSE_ABS_DIFF(.34, test_t);
CHECK_CLOSE_ABS_DIFF(.34e10, test_t);

12
test/lexical_cast_iterator_range_test.cpp
View File

@@ -138,7 +138,7 @@ void test_it_range_using_char(CharT* one, CharT* eleven)
#endif
}
void test_char_iterator_ranges()
void test_char_iterator_ranges()
{
typedef char test_char_type;
test_char_type data1[] = "1";
@@ -149,7 +149,7 @@ void test_char_iterator_ranges()
void test_unsigned_char_iterator_ranges()
void test_unsigned_char_iterator_ranges()
{
typedef unsigned char test_char_type;
test_char_type data1[] = "1";
@@ -158,7 +158,7 @@ void test_unsigned_char_iterator_ranges()
test_it_range_using_char(data1, data2);
}
void test_signed_char_iterator_ranges()
void test_signed_char_iterator_ranges()
{
typedef signed char test_char_type;
test_char_type data1[] = "1";
@@ -167,7 +167,7 @@ void test_signed_char_iterator_ranges()
test_it_range_using_char(data1, data2);
}
void test_wchar_iterator_ranges()
void test_wchar_iterator_ranges()
{
#ifndef BOOST_LCAST_NO_WCHAR_T
typedef wchar_t test_char_type;
@@ -179,7 +179,7 @@ void test_wchar_iterator_ranges()
BOOST_CHECK(true);
}
void test_char16_iterator_ranges()
void test_char16_iterator_ranges()
{
#if !defined(BOOST_NO_CHAR16_T) && !defined(BOOST_NO_UNICODE_LITERALS)
typedef char16_t test_char_type;
@@ -191,7 +191,7 @@ void test_char16_iterator_ranges()
BOOST_CHECK(true);
}
void test_char32_iterator_ranges()
void test_char32_iterator_ranges()
{
#if !defined(BOOST_NO_CHAR32_T) && !defined(BOOST_NO_UNICODE_LITERALS)
typedef char32_t test_char_type;

14
test/lexical_cast_no_exceptions_test.cpp
View File

@@ -22,7 +22,7 @@
#ifndef BOOST_NO_EXCEPTIONS
#error "This test must be compiled with -DBOOST_NO_EXCEPTIONS"
#endif
#endif
bool g_was_exception = false;
@@ -60,25 +60,25 @@ inline std::istream& operator>> (std::istream& i, Escape& rhs)
void test_exceptions_off()
{
Escape v("");
g_was_exception = false;
g_was_exception = false;
lexical_cast<char>(v);
BOOST_CHECK(g_was_exception);
g_was_exception = false;
lexical_cast<unsigned char>(v);
BOOST_CHECK(g_was_exception);
v = lexical_cast<Escape>(100);
BOOST_CHECK_EQUAL(lexical_cast<int>(v), 100);
BOOST_CHECK_EQUAL(lexical_cast<unsigned int>(v), 100u);
v = lexical_cast<Escape>(0.0);
BOOST_CHECK_EQUAL(lexical_cast<double>(v), 0.0);
BOOST_CHECK_EQUAL(lexical_cast<short>(100), 100);
BOOST_CHECK_EQUAL(lexical_cast<float>(0.0), 0.0);
g_was_exception = false;
lexical_cast<short>(700000);
BOOST_CHECK(g_was_exception);

10
test/lexical_cast_no_locale_test.cpp
View File

@@ -23,12 +23,12 @@
using namespace boost;
// Testing compilation and some basic usage with BOOST_NO_STD_LOCALE
// Tests are mainly copyied from lexical_cast_empty_input_test.cpp (something
// Tests are mainly copyied from lexical_cast_empty_input_test.cpp (something
// new added to test_empty_3)
#ifndef BOOST_NO_STD_LOCALE
#error "This test must be compiled with -DBOOST_NO_STD_LOCALE"
#endif
#endif
template <class T>
@@ -106,15 +106,15 @@ void test_empty_3()
{
Escape v("");
do_test_on_empty_input(v);
BOOST_CHECK_THROW(lexical_cast<char>(v), bad_lexical_cast);
BOOST_CHECK_THROW(lexical_cast<unsigned char>(v), bad_lexical_cast);
BOOST_CHECK_THROW(lexical_cast<signed char>(v), bad_lexical_cast);
v = lexical_cast<Escape>(100);
BOOST_CHECK_EQUAL(lexical_cast<int>(v), 100);
BOOST_CHECK_EQUAL(lexical_cast<unsigned int>(v), 100u);
v = lexical_cast<Escape>(0.0);
BOOST_CHECK_EQUAL(lexical_cast<double>(v), 0.0);
}

2
test/lexical_cast_typedefed_wchar_test_runtime.cpp
View File

@@ -30,7 +30,7 @@ void test_typedefed_wchar_t_runtime()
BOOST_CHECK_EQUAL(boost::lexical_cast<int>(L'A'), 65);
BOOST_CHECK_EQUAL(boost::lexical_cast<int>(L'B'), 66);
BOOST_CHECK_EQUAL(boost::lexical_cast<wchar_t>(L"65"), 65);
BOOST_CHECK_EQUAL(boost::lexical_cast<wchar_t>(L"66"), 66);
#endif