pybind11/tests/test_builtin_casters.cpp

419 lines
17 KiB
C++
Raw Normal View History

/*
tests/test_builtin_casters.cpp -- Casters available without any additional headers
Copyright (c) 2017 Wenzel Jakob <wenzel.jakob@epfl.ch>
All rights reserved. Use of this source code is governed by a
BSD-style license that can be found in the LICENSE file.
*/
#include <pybind11/complex.h>
#include "pybind11_tests.h"
#include <utility>
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
struct ConstRefCasted {
int tag;
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
};
struct StringAttr {
explicit StringAttr(std::string v) : value(std::move(v)) {}
std::string value;
};
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
PYBIND11_NAMESPACE_BEGIN(pybind11)
PYBIND11_NAMESPACE_BEGIN(detail)
template <>
class type_caster<ConstRefCasted> {
public:
static constexpr auto name = const_name<ConstRefCasted>();
// Input is unimportant, a new value will always be constructed based on the
// cast operator.
bool load(handle, bool) { return true; }
explicit operator ConstRefCasted &&() {
value = {1};
// NOLINTNEXTLINE(performance-move-const-arg)
return std::move(value);
}
explicit operator ConstRefCasted &() {
value = {2};
return value;
}
explicit operator ConstRefCasted *() {
value = {3};
return &value;
}
explicit operator const ConstRefCasted &() {
value = {4};
return value;
}
explicit operator const ConstRefCasted *() {
value = {5};
return &value;
}
// custom cast_op to explicitly propagate types to the conversion operators.
template <typename T_>
using cast_op_type =
/// const
conditional_t<
std::is_same<remove_reference_t<T_>, const ConstRefCasted *>::value,
const ConstRefCasted *,
conditional_t<
std::is_same<T_, const ConstRefCasted &>::value,
const ConstRefCasted &,
/// non-const
conditional_t<std::is_same<remove_reference_t<T_>, ConstRefCasted *>::value,
ConstRefCasted *,
conditional_t<std::is_same<T_, ConstRefCasted &>::value,
ConstRefCasted &,
/* else */ ConstRefCasted &&>>>>;
private:
ConstRefCasted value = {0};
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
};
PYBIND11_NAMESPACE_END(detail)
PYBIND11_NAMESPACE_END(pybind11)
TEST_SUBMODULE(builtin_casters, m) {
// test_simple_string
m.def("string_roundtrip", [](const char *s) { return s; });
// test_unicode_conversion
// Some test characters in utf16 and utf32 encodings. The last one (the 𝐀) contains a null
// byte
char32_t a32 = 0x61 /*a*/, z32 = 0x7a /*z*/, ib32 = 0x203d /*‽*/, cake32 = 0x1f382 /*🎂*/,
mathbfA32 = 0x1d400 /*𝐀*/;
char16_t b16 = 0x62 /*b*/, z16 = 0x7a, ib16 = 0x203d, cake16_1 = 0xd83c, cake16_2 = 0xdf82,
mathbfA16_1 = 0xd835, mathbfA16_2 = 0xdc00;
std::wstring wstr;
wstr.push_back(0x61); // a
wstr.push_back(0x2e18); // ⸘
if (PYBIND11_SILENCE_MSVC_C4127(sizeof(wchar_t) == 2)) {
wstr.push_back(mathbfA16_1);
wstr.push_back(mathbfA16_2);
} // 𝐀, utf16
else {
wstr.push_back((wchar_t) mathbfA32);
} // 𝐀, utf32
wstr.push_back(0x7a); // z
m.def("good_utf8_string", []() {
return std::string((const char *) u8"Say utf8\u203d \U0001f382 \U0001d400");
}); // Say utf8‽ 🎂 𝐀
m.def("good_utf16_string", [=]() {
return std::u16string({b16, ib16, cake16_1, cake16_2, mathbfA16_1, mathbfA16_2, z16});
}); // b‽🎂𝐀z
m.def("good_utf32_string", [=]() {
return std::u32string({a32, mathbfA32, cake32, ib32, z32});
}); // a𝐀🎂‽z
m.def("good_wchar_string", [=]() { return wstr; }); // a‽𝐀z
m.def("bad_utf8_string", []() {
return std::string("abc\xd0"
"def");
});
m.def("bad_utf16_string", [=]() { return std::u16string({b16, char16_t(0xd800), z16}); });
// Under Python 2.7, invalid unicode UTF-32 characters didn't appear to trigger
// UnicodeDecodeError
m.def("bad_utf32_string", [=]() { return std::u32string({a32, char32_t(0xd800), z32}); });
if (PYBIND11_SILENCE_MSVC_C4127(sizeof(wchar_t) == 2)) {
m.def("bad_wchar_string", [=]() {
return std::wstring({wchar_t(0x61), wchar_t(0xd800)});
});
}
m.def("u8_Z", []() -> char { return 'Z'; });
m.def("u8_eacute", []() -> char { return '\xe9'; });
m.def("u16_ibang", [=]() -> char16_t { return ib16; });
m.def("u32_mathbfA", [=]() -> char32_t { return mathbfA32; });
m.def("wchar_heart", []() -> wchar_t { return 0x2665; });
// test_single_char_arguments
m.attr("wchar_size") = py::cast(sizeof(wchar_t));
m.def("ord_char", [](char c) -> int { return static_cast<unsigned char>(c); });
m.def("ord_char_lv", [](char &c) -> int { return static_cast<unsigned char>(c); });
m.def("ord_char16", [](char16_t c) -> uint16_t { return c; });
m.def("ord_char16_lv", [](char16_t &c) -> uint16_t { return c; });
m.def("ord_char32", [](char32_t c) -> uint32_t { return c; });
m.def("ord_wchar", [](wchar_t c) -> int { return c; });
// test_bytes_to_string
m.def("strlen", [](char *s) { return strlen(s); });
m.def("string_length", [](const std::string &s) { return s.length(); });
#ifdef PYBIND11_HAS_U8STRING
m.attr("has_u8string") = true;
m.def("good_utf8_u8string", []() {
return std::u8string(u8"Say utf8\u203d \U0001f382 \U0001d400");
}); // Say utf8‽ 🎂 𝐀
m.def("bad_utf8_u8string", []() {
return std::u8string((const char8_t *) "abc\xd0"
"def");
});
m.def("u8_char8_Z", []() -> char8_t { return u8'Z'; });
// test_single_char_arguments
m.def("ord_char8", [](char8_t c) -> int { return static_cast<unsigned char>(c); });
m.def("ord_char8_lv", [](char8_t &c) -> int { return static_cast<unsigned char>(c); });
#endif
// test_string_view
#ifdef PYBIND11_HAS_STRING_VIEW
m.attr("has_string_view") = true;
m.def("string_view_print", [](std::string_view s) { py::print(s, s.size()); });
m.def("string_view16_print", [](std::u16string_view s) { py::print(s, s.size()); });
m.def("string_view32_print", [](std::u32string_view s) { py::print(s, s.size()); });
m.def("string_view_chars", [](std::string_view s) {
py::list l;
for (auto c : s) {
l.append((std::uint8_t) c);
}
return l;
});
m.def("string_view16_chars", [](std::u16string_view s) {
py::list l;
for (auto c : s) {
l.append((int) c);
}
return l;
});
m.def("string_view32_chars", [](std::u32string_view s) {
py::list l;
for (auto c : s) {
l.append((int) c);
}
return l;
});
m.def("string_view_return",
[]() { return std::string_view((const char *) u8"utf8 secret \U0001f382"); });
m.def("string_view16_return",
[]() { return std::u16string_view(u"utf16 secret \U0001f382"); });
m.def("string_view32_return",
[]() { return std::u32string_view(U"utf32 secret \U0001f382"); });
Expand std::string_view support to str, bytes, memoryview (#3521) * Expand string_view support to str, bytes, memoryview 1. Allows constructing a str or bytes implicitly from a string_view; this is essentially a small shortcut allowing a caller to write `py::bytes{sv}` rather than `py::bytes{sv.data(), sv.size()}`. 2. Allows implicit conversion *to* string_view from py::bytes -- this saves a fair bit more as currently there is no simple way to get such a view of the bytes without copying it (or resorting to Python API calls). (This is not done for `str` because when the str contains unicode we have to allocate to a temporary and so there might not be some string data we can properly view without owning.) 3. Allows `memoryview::from_memory` to accept a string_view. As with the other from_memory calls, it's entirely your responsibility to keep it alive. This also required moving the string_view availability detection into detail/common.h because this PR needs it in pytypes.h, which is higher up the include chain than cast.h where it was being detected currently. * Move string_view include to pytypes.h * CI-testing a fix for the "ambiguous conversion" issue. This change is known to fix the `tensorflow::tstring` issue reported under https://github.com/pybind/pybind11/pull/3521#issuecomment-985100965 TODO: Minimal reproducer for the `tensorflow::tstring` issue. * Make clang-tidy happy (hopefully). * Adding minimal reproducer for the `tensorflow::tstring` issue. Error without the enable_if trick: ``` /usr/local/google/home/rwgk/forked/pybind11/tests/test_builtin_casters.cpp:169:16: error: ambiguous conversion for functional-style cast from 'TypeWithBothOperatorStringAndStringView' to 'py::bytes' return py::bytes(TypeWithBothOperatorStringAndStringView()); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /usr/local/google/home/rwgk/forked/pybind11/include/pybind11/detail/../pytypes.h:1174:5: note: candidate constructor bytes(const std::string &s) : bytes(s.data(), s.size()) { } ^ /usr/local/google/home/rwgk/forked/pybind11/include/pybind11/detail/../pytypes.h:1191:5: note: candidate constructor bytes(std::string_view s) : bytes(s.data(), s.size()) { } ^ ``` * Adding missing NOLINTNEXTLINE * Also apply ambiguous conversion workaround to str() Co-authored-by: Ralf W. Grosse-Kunstleve <rwgk@google.com>
2021-12-03 17:20:32 +00:00
// The inner lambdas here are to also test implicit conversion
using namespace std::literals;
m.def("string_view_bytes",
[]() { return [](py::bytes b) { return b; }("abc \x80\x80 def"sv); });
m.def("string_view_str",
[]() { return [](py::str s) { return s; }("abc \342\200\275 def"sv); });
m.def("string_view_from_bytes",
[](const py::bytes &b) { return [](std::string_view s) { return s; }(b); });
Expand std::string_view support to str, bytes, memoryview (#3521) * Expand string_view support to str, bytes, memoryview 1. Allows constructing a str or bytes implicitly from a string_view; this is essentially a small shortcut allowing a caller to write `py::bytes{sv}` rather than `py::bytes{sv.data(), sv.size()}`. 2. Allows implicit conversion *to* string_view from py::bytes -- this saves a fair bit more as currently there is no simple way to get such a view of the bytes without copying it (or resorting to Python API calls). (This is not done for `str` because when the str contains unicode we have to allocate to a temporary and so there might not be some string data we can properly view without owning.) 3. Allows `memoryview::from_memory` to accept a string_view. As with the other from_memory calls, it's entirely your responsibility to keep it alive. This also required moving the string_view availability detection into detail/common.h because this PR needs it in pytypes.h, which is higher up the include chain than cast.h where it was being detected currently. * Move string_view include to pytypes.h * CI-testing a fix for the "ambiguous conversion" issue. This change is known to fix the `tensorflow::tstring` issue reported under https://github.com/pybind/pybind11/pull/3521#issuecomment-985100965 TODO: Minimal reproducer for the `tensorflow::tstring` issue. * Make clang-tidy happy (hopefully). * Adding minimal reproducer for the `tensorflow::tstring` issue. Error without the enable_if trick: ``` /usr/local/google/home/rwgk/forked/pybind11/tests/test_builtin_casters.cpp:169:16: error: ambiguous conversion for functional-style cast from 'TypeWithBothOperatorStringAndStringView' to 'py::bytes' return py::bytes(TypeWithBothOperatorStringAndStringView()); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /usr/local/google/home/rwgk/forked/pybind11/include/pybind11/detail/../pytypes.h:1174:5: note: candidate constructor bytes(const std::string &s) : bytes(s.data(), s.size()) { } ^ /usr/local/google/home/rwgk/forked/pybind11/include/pybind11/detail/../pytypes.h:1191:5: note: candidate constructor bytes(std::string_view s) : bytes(s.data(), s.size()) { } ^ ``` * Adding missing NOLINTNEXTLINE * Also apply ambiguous conversion workaround to str() Co-authored-by: Ralf W. Grosse-Kunstleve <rwgk@google.com>
2021-12-03 17:20:32 +00:00
m.def("string_view_memoryview", []() {
static constexpr auto val = "Have some \360\237\216\202"sv;
return py::memoryview::from_memory(val);
});
# ifdef PYBIND11_HAS_U8STRING
m.def("string_view8_print", [](std::u8string_view s) { py::print(s, s.size()); });
m.def("string_view8_chars", [](std::u8string_view s) {
py::list l;
for (auto c : s)
l.append((std::uint8_t) c);
return l;
});
m.def("string_view8_return", []() { return std::u8string_view(u8"utf8 secret \U0001f382"); });
m.def("string_view8_str", []() { return py::str{std::u8string_view{u8"abc ‽ def"}}; });
# endif
Expand std::string_view support to str, bytes, memoryview (#3521) * Expand string_view support to str, bytes, memoryview 1. Allows constructing a str or bytes implicitly from a string_view; this is essentially a small shortcut allowing a caller to write `py::bytes{sv}` rather than `py::bytes{sv.data(), sv.size()}`. 2. Allows implicit conversion *to* string_view from py::bytes -- this saves a fair bit more as currently there is no simple way to get such a view of the bytes without copying it (or resorting to Python API calls). (This is not done for `str` because when the str contains unicode we have to allocate to a temporary and so there might not be some string data we can properly view without owning.) 3. Allows `memoryview::from_memory` to accept a string_view. As with the other from_memory calls, it's entirely your responsibility to keep it alive. This also required moving the string_view availability detection into detail/common.h because this PR needs it in pytypes.h, which is higher up the include chain than cast.h where it was being detected currently. * Move string_view include to pytypes.h * CI-testing a fix for the "ambiguous conversion" issue. This change is known to fix the `tensorflow::tstring` issue reported under https://github.com/pybind/pybind11/pull/3521#issuecomment-985100965 TODO: Minimal reproducer for the `tensorflow::tstring` issue. * Make clang-tidy happy (hopefully). * Adding minimal reproducer for the `tensorflow::tstring` issue. Error without the enable_if trick: ``` /usr/local/google/home/rwgk/forked/pybind11/tests/test_builtin_casters.cpp:169:16: error: ambiguous conversion for functional-style cast from 'TypeWithBothOperatorStringAndStringView' to 'py::bytes' return py::bytes(TypeWithBothOperatorStringAndStringView()); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /usr/local/google/home/rwgk/forked/pybind11/include/pybind11/detail/../pytypes.h:1174:5: note: candidate constructor bytes(const std::string &s) : bytes(s.data(), s.size()) { } ^ /usr/local/google/home/rwgk/forked/pybind11/include/pybind11/detail/../pytypes.h:1191:5: note: candidate constructor bytes(std::string_view s) : bytes(s.data(), s.size()) { } ^ ``` * Adding missing NOLINTNEXTLINE * Also apply ambiguous conversion workaround to str() Co-authored-by: Ralf W. Grosse-Kunstleve <rwgk@google.com>
2021-12-03 17:20:32 +00:00
struct TypeWithBothOperatorStringAndStringView {
// NOLINTNEXTLINE(google-explicit-constructor)
operator std::string() const { return "success"; }
// NOLINTNEXTLINE(google-explicit-constructor)
operator std::string_view() const { return "failure"; }
};
m.def("bytes_from_type_with_both_operator_string_and_string_view",
[]() { return py::bytes(TypeWithBothOperatorStringAndStringView()); });
m.def("str_from_type_with_both_operator_string_and_string_view",
[]() { return py::str(TypeWithBothOperatorStringAndStringView()); });
#endif
Fix unsigned error value casting When casting to an unsigned type from a python 2 `int`, we currently cast using `(unsigned long long) PyLong_AsUnsignedLong(src.ptr())`. If the Python cast fails, it returns (unsigned long) -1, but then we cast this to `unsigned long long`, which means we get 4294967295, but because that isn't equal to `(unsigned long long) -1`, we don't detect the failure. This commit moves the unsigned casting into a `detail::as_unsigned` function which, upon error, casts -1 to the final type, and otherwise casts the return value to the final type to avoid the problematic double cast when an error occurs. The error most commonly shows up wherever `long` is 32-bits (e.g. under both 32- and 64-bit Windows, and under 32-bit linux) when passing a negative value to a bound function taking an `unsigned long`. Fixes #929. The added tests also trigger a latent segfault under PyPy: when casting to an integer smaller than `long` (e.g. casting to a `uint32_t` on a 64-bit `long` architecture) we check both for a Python error and also that the resulting intermediate value will fit in the final type. If there is no conversion error, but we get a value that would overflow, we end up calling `PyErr_ExceptionMatches()` illegally: that call is only allowed when there is a current exception. Under PyPy, this segfaults the test suite. It doesn't appear to segfault under CPython, but the documentation suggests that it *could* do so. The fix is to only check for the exception match if we actually got an error.
2017-07-01 20:31:49 +00:00
// test_integer_casting
m.def("i32_str", [](std::int32_t v) { return std::to_string(v); });
m.def("u32_str", [](std::uint32_t v) { return std::to_string(v); });
m.def("i64_str", [](std::int64_t v) { return std::to_string(v); });
m.def("u64_str", [](std::uint64_t v) { return std::to_string(v); });
// test_int_convert
m.def("int_passthrough", [](int arg) { return arg; });
m.def(
"int_passthrough_noconvert", [](int arg) { return arg; }, py::arg{}.noconvert());
// test_tuple
m.def(
"pair_passthrough",
[](const std::pair<bool, std::string> &input) {
return std::make_pair(input.second, input.first);
},
"Return a pair in reversed order");
m.def(
"tuple_passthrough",
[](std::tuple<bool, std::string, int> input) {
return std::make_tuple(std::get<2>(input), std::get<1>(input), std::get<0>(input));
},
"Return a triple in reversed order");
m.def("empty_tuple", []() { return std::tuple<>(); });
static std::pair<RValueCaster, RValueCaster> lvpair;
static std::tuple<RValueCaster, RValueCaster, RValueCaster> lvtuple;
static std::pair<RValueCaster, std::tuple<RValueCaster, std::pair<RValueCaster, RValueCaster>>>
lvnested;
m.def("rvalue_pair", []() { return std::make_pair(RValueCaster{}, RValueCaster{}); });
m.def("lvalue_pair", []() -> const decltype(lvpair) & { return lvpair; });
m.def("rvalue_tuple",
[]() { return std::make_tuple(RValueCaster{}, RValueCaster{}, RValueCaster{}); });
m.def("lvalue_tuple", []() -> const decltype(lvtuple) & { return lvtuple; });
m.def("rvalue_nested", []() {
return std::make_pair(
RValueCaster{},
std::make_tuple(RValueCaster{}, std::make_pair(RValueCaster{}, RValueCaster{})));
});
m.def("lvalue_nested", []() -> const decltype(lvnested) & { return lvnested; });
m.def(
"int_string_pair",
[]() {
// Using no-destructor idiom to side-step warnings from overzealous compilers.
static auto *int_string_pair = new std::pair<int, std::string>{2, "items"};
return int_string_pair;
},
py::return_value_policy::reference);
// test_builtins_cast_return_none
m.def("return_none_string", []() -> std::string * { return nullptr; });
m.def("return_none_char", []() -> const char * { return nullptr; });
m.def("return_none_bool", []() -> bool * { return nullptr; });
m.def("return_none_int", []() -> int * { return nullptr; });
m.def("return_none_float", []() -> float * { return nullptr; });
m.def("return_none_pair", []() -> std::pair<int, int> * { return nullptr; });
// test_none_deferred
m.def("defer_none_cstring", [](char *) { return false; });
m.def("defer_none_cstring", [](const py::none &) { return true; });
m.def("defer_none_custom", [](UserType *) { return false; });
m.def("defer_none_custom", [](const py::none &) { return true; });
m.def("nodefer_none_void", [](void *) { return true; });
m.def("nodefer_none_void", [](const py::none &) { return false; });
// test_void_caster
m.def("load_nullptr_t", [](std::nullptr_t) {}); // not useful, but it should still compile
m.def("cast_nullptr_t", []() { return std::nullptr_t{}; });
ci: Intel icc/icpc via oneAPI (#2573) * CI: Intel icc/icpc via oneAPI Add testing for Intel icc/icpc via the oneAPI images. Intel oneAPI is in a late beta stage, currently shipping oneAPI beta09 with ICC 20.2. * CI: Skip Interpreter Tests for Intel Cannot find how to add this, neiter the package `libc6-dev` nor `intel-oneapi-mkl-devel` help when installed to solve this: ``` -- Looking for C++ include pthread.h -- Looking for C++ include pthread.h - not found CMake Error at /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:165 (message): Could NOT find Threads (missing: Threads_FOUND) Call Stack (most recent call first): /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:458 (_FPHSA_FAILURE_MESSAGE) /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindThreads.cmake:234 (FIND_PACKAGE_HANDLE_STANDARD_ARGS) tests/test_embed/CMakeLists.txt:17 (find_package) ``` * CI: libc6-dev from GCC for ICC * CI: Run bare metal for oneAPI * CI: Ubuntu 18.04 for oneAPI * CI: Intel +Catch -Eigen * CI: CMake from Apt (ICC tests) * CI: Replace Intel Py with GCC Py * CI: Intel w/o GCC's Eigen * CI: ICC with verbose make * [Debug] Find core dump * tests: use arg{} instead of arg() for Intel * tests: adding a few more missing {} * fix: sync with @tobiasleibner's branch * fix: try ubuntu 20-04 * fix: drop exit 1 * style: clang tidy fix * style: fix missing NOLINT * ICC: Update Compiler Name Changed upstream with the last oneAPI release. * ICC CI: Downgrade pytest pytest 6 does not capture the `discard_as_unraisable` stderr and just writes a warning with its content instead. * Use new test pinning requirements.txt * tests: add notes about intel, cleanup Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>
2021-01-15 20:59:47 +00:00
// [workaround(intel)] ICC 20/21 breaks with py::arg().stuff, using py::arg{}.stuff works.
// test_bool_caster
m.def("bool_passthrough", [](bool arg) { return arg; });
m.def(
"bool_passthrough_noconvert", [](bool arg) { return arg; }, py::arg{}.noconvert());
fix: Intel ICC C++17 compatibility (#2729) * CI: Intel icc/icpc via oneAPI Add testing for Intel icc/icpc via the oneAPI images. Intel oneAPI is in a late beta stage, currently shipping oneAPI beta09 with ICC 20.2. CI: Skip Interpreter Tests for Intel Cannot find how to add this, neiter the package `libc6-dev` nor `intel-oneapi-mkl-devel` help when installed to solve this: ``` -- Looking for C++ include pthread.h -- Looking for C++ include pthread.h - not found CMake Error at /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:165 (message): Could NOT find Threads (missing: Threads_FOUND) Call Stack (most recent call first): /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:458 (_FPHSA_FAILURE_MESSAGE) /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindThreads.cmake:234 (FIND_PACKAGE_HANDLE_STANDARD_ARGS) tests/test_embed/CMakeLists.txt:17 (find_package) ``` CI: libc6-dev from GCC for ICC CI: Run bare metal for oneAPI CI: Ubuntu 18.04 for oneAPI CI: Intel +Catch -Eigen CI: CMake from Apt (ICC tests) CI: Replace Intel Py with GCC Py CI: Intel w/o GCC's Eigen CI: ICC with verbose make [Debug] Find core dump tests: use arg{} instead of arg() for Intel tests: adding a few more missing {} fix: sync with @tobiasleibner's branch fix: try ubuntu 20-04 fix: drop exit 1 docs: Apply suggestions from code review Co-authored-by: Tobias Leibner <tobias.leibner@googlemail.com> Workaround for ICC enable_if issues Another workaround for ICC's enable_if issues fix error in previous commit Disable one test for the Intel compiler in C++17 mode Add back one instance of py::arg().noconvert() Add NOLINT to fix clang-tidy check Work around for ICC internal error in PYBIND11_EXPAND_SIDE_EFFECTS in C++17 mode CI: Intel ICC with C++17 docs: pybind11/numpy.h does not require numpy at build time. (#2720) This is nice enough to be mentioned explicitly in the docs. docs: Update warning about Python 3.9.0 UB, now that 3.9.1 has been released (#2719) Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python. ci: drop pypy2 linux, PGI 20.7, add Python 10 dev (#2724) * ci: drop pypy2 linux, add Python 10 dev * ci: fix mistake * ci: commented-out PGI 20.11, drop 20.7 fix: regression with installed pybind11 overriding local one (#2716) * fix: regression with installed pybind11 overriding discovered one Closes #2709 * docs: wording incorrect style: remove redundant instance->owned = true (#2723) which was just before set to True in instance->allocate_layout() fix: also throw in the move-constructor added by the PYBIND11_OBJECT macro, after the argument has been moved-out (if necessary) (#2701) Make args_are_all_* ICC workarounds unconditional Disable test_aligned on Intel ICC Fix test_aligned on Intel ICC Skip test_python_alreadyset_in_destructor on Intel ICC Fix test_aligned again ICC CI: Downgrade pytest pytest 6 does not capture the `discard_as_unraisable` stderr and just writes a warning with its content instead. * refactor: simpler Intel workaround, suggested by @laramiel * fix: try version with impl to see if it is easier to compile * docs: update README for ICC Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja> Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>
2021-01-18 00:53:07 +00:00
// TODO: This should be disabled and fixed in future Intel compilers
#if !defined(__INTEL_COMPILER)
// Test "bool_passthrough_noconvert" again, but using () instead of {} to construct py::arg
// When compiled with the Intel compiler, this results in segmentation faults when importing
// the module. Tested with icc (ICC) 2021.1 Beta 20200827, this should be tested again when
// a newer version of icc is available.
m.def(
"bool_passthrough_noconvert2", [](bool arg) { return arg; }, py::arg().noconvert());
fix: Intel ICC C++17 compatibility (#2729) * CI: Intel icc/icpc via oneAPI Add testing for Intel icc/icpc via the oneAPI images. Intel oneAPI is in a late beta stage, currently shipping oneAPI beta09 with ICC 20.2. CI: Skip Interpreter Tests for Intel Cannot find how to add this, neiter the package `libc6-dev` nor `intel-oneapi-mkl-devel` help when installed to solve this: ``` -- Looking for C++ include pthread.h -- Looking for C++ include pthread.h - not found CMake Error at /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:165 (message): Could NOT find Threads (missing: Threads_FOUND) Call Stack (most recent call first): /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindPackageHandleStandardArgs.cmake:458 (_FPHSA_FAILURE_MESSAGE) /__t/cmake/3.18.4/x64/cmake-3.18.4-Linux-x86_64/share/cmake-3.18/Modules/FindThreads.cmake:234 (FIND_PACKAGE_HANDLE_STANDARD_ARGS) tests/test_embed/CMakeLists.txt:17 (find_package) ``` CI: libc6-dev from GCC for ICC CI: Run bare metal for oneAPI CI: Ubuntu 18.04 for oneAPI CI: Intel +Catch -Eigen CI: CMake from Apt (ICC tests) CI: Replace Intel Py with GCC Py CI: Intel w/o GCC's Eigen CI: ICC with verbose make [Debug] Find core dump tests: use arg{} instead of arg() for Intel tests: adding a few more missing {} fix: sync with @tobiasleibner's branch fix: try ubuntu 20-04 fix: drop exit 1 docs: Apply suggestions from code review Co-authored-by: Tobias Leibner <tobias.leibner@googlemail.com> Workaround for ICC enable_if issues Another workaround for ICC's enable_if issues fix error in previous commit Disable one test for the Intel compiler in C++17 mode Add back one instance of py::arg().noconvert() Add NOLINT to fix clang-tidy check Work around for ICC internal error in PYBIND11_EXPAND_SIDE_EFFECTS in C++17 mode CI: Intel ICC with C++17 docs: pybind11/numpy.h does not require numpy at build time. (#2720) This is nice enough to be mentioned explicitly in the docs. docs: Update warning about Python 3.9.0 UB, now that 3.9.1 has been released (#2719) Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python. ci: drop pypy2 linux, PGI 20.7, add Python 10 dev (#2724) * ci: drop pypy2 linux, add Python 10 dev * ci: fix mistake * ci: commented-out PGI 20.11, drop 20.7 fix: regression with installed pybind11 overriding local one (#2716) * fix: regression with installed pybind11 overriding discovered one Closes #2709 * docs: wording incorrect style: remove redundant instance->owned = true (#2723) which was just before set to True in instance->allocate_layout() fix: also throw in the move-constructor added by the PYBIND11_OBJECT macro, after the argument has been moved-out (if necessary) (#2701) Make args_are_all_* ICC workarounds unconditional Disable test_aligned on Intel ICC Fix test_aligned on Intel ICC Skip test_python_alreadyset_in_destructor on Intel ICC Fix test_aligned again ICC CI: Downgrade pytest pytest 6 does not capture the `discard_as_unraisable` stderr and just writes a warning with its content instead. * refactor: simpler Intel workaround, suggested by @laramiel * fix: try version with impl to see if it is easier to compile * docs: update README for ICC Co-authored-by: Axel Huebl <axel.huebl@plasma.ninja> Co-authored-by: Henry Schreiner <henryschreineriii@gmail.com>
2021-01-18 00:53:07 +00:00
#endif
// test_reference_wrapper
m.def("refwrap_builtin", [](std::reference_wrapper<int> p) { return 10 * p.get(); });
m.def("refwrap_usertype", [](std::reference_wrapper<UserType> p) { return p.get().value(); });
m.def("refwrap_usertype_const",
[](std::reference_wrapper<const UserType> p) { return p.get().value(); });
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
m.def("refwrap_lvalue", []() -> std::reference_wrapper<UserType> {
static UserType x(1);
return std::ref(x);
});
m.def("refwrap_lvalue_const", []() -> std::reference_wrapper<const UserType> {
static UserType x(1);
return std::cref(x);
});
// Not currently supported (std::pair caster has return-by-value cast operator);
// triggers static_assert failure.
// m.def("refwrap_pair", [](std::reference_wrapper<std::pair<int, int>>) { });
m.def(
"refwrap_list",
[](bool copy) {
static IncType x1(1), x2(2);
py::list l;
for (const auto &f : {std::ref(x1), std::ref(x2)}) {
l.append(py::cast(
f, copy ? py::return_value_policy::copy : py::return_value_policy::reference));
}
return l;
},
"copy"_a);
m.def("refwrap_iiw", [](const IncType &w) { return w.value(); });
m.def("refwrap_call_iiw", [](IncType &w, const py::function &f) {
py::list l;
l.append(f(std::ref(w)));
l.append(f(std::cref(w)));
IncType x(w.value());
l.append(f(std::ref(x)));
IncType y(w.value());
auto r3 = std::ref(y);
l.append(f(r3));
return l;
});
// test_complex
m.def("complex_cast", [](float x) { return "{}"_s.format(x); });
m.def("complex_cast",
[](std::complex<float> x) { return "({}, {})"_s.format(x.real(), x.imag()); });
// test int vs. long (Python 2)
m.def("int_cast", []() { return (int) 42; });
m.def("long_cast", []() { return (long) 42; });
m.def("longlong_cast", []() { return ULLONG_MAX; });
/// test void* cast operator
m.def("test_void_caster", []() -> bool {
void *v = (void *) 0xabcd;
py::object o = py::cast(v);
return py::cast<void *>(o) == v;
});
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
// Tests const/non-const propagation in cast_op.
m.def("takes", [](ConstRefCasted x) { return x.tag; });
m.def("takes_move", [](ConstRefCasted &&x) { return x.tag; });
m.def("takes_ptr", [](ConstRefCasted *x) { return x->tag; });
m.def("takes_ref", [](ConstRefCasted &x) { return x.tag; });
Adjusting `type_caster<std::reference_wrapper<T>>` to support const/non-const propagation in `cast_op`. (#2705) * Allow type_caster of std::reference_wrapper<T> to be the same as a native reference. Before, both std::reference_wrapper<T> and std::reference_wrapper<const T> would invoke cast_op<type>. This doesn't allow the type_caster<> specialization for T to distinguish reference_wrapper types from value types. After, the type_caster<> specialization invokes cast_op<type&>, which allows reference_wrapper to behave in the same way as a native reference type. * Add tests/examples for std::reference_wrapper<const T> * Add tests which use mutable/immutable variants This test is a chimera; it blends the pybind11 casters with a custom pytype implementation that supports immutable and mutable calls. In order to detect the immutable/mutable state, the cast_op needs to propagate it, even through e.g. std::reference<const T> Note: This is still a work in progress; some things are crashing, which likely means that I have a refcounting bug or something else missing. * Add/finish tests that distinguish const& from & Fixes the bugs in my custom python type implementation, demonstrate test that requires const& and reference_wrapper<const T> being treated differently from Non-const. * Add passing a const to non-const method. * Demonstrate non-const conversion of reference_wrapper in tests. Apply formatting presubmit check. * Fix build errors from presubmit checks. * Try and fix a few more CI errors * More CI fixes. * More CI fixups. * Try and get PyPy to work. * Additional minor fixups. Getting close to CI green. * More ci fixes? * fix clang-tidy warnings from presubmit * fix more clang-tidy warnings * minor comment and consistency cleanups * PyDECREF -> Py_DECREF * copy/move constructors * Resolve codereview comments * more review comment fixes * review comments: remove spurious & * Make the test fail even when the static_assert is commented out. This expands the test_freezable_type_caster a bit by: 1/ adding accessors .is_immutable and .addr to compare identity from python. 2/ Changing the default cast_op of the type_caster<> specialization to return a non-const value. In normal codepaths this is a reasonable default. 3/ adding roundtrip variants to exercise the by reference, by pointer and by reference_wrapper in all call paths. In conjunction with 2/, this demonstrates the failure case of the existing std::reference_wrpper conversion, which now loses const in a similar way that happens when using the default cast_op_type<>. * apply presubmit formatting * Revert inclusion of test_freezable_type_caster There's some concern that this test is a bit unwieldly because of the use of the raw <Python.h> functions. Removing for now. * Add a test that validates const references propagation. This test verifies that cast_op may be used to correctly detect const reference types when used with std::reference_wrapper. * mend * Review comments based changes. 1. std::add_lvalue_reference<type> -> type& 2. Simplify the test a little more; we're never returning the ConstRefCaster type so the class_ definition can be removed. * formatted files again. * Move const_ref_caster test to builtin_casters * Review comments: use cast_op and adjust some comments. * Simplify ConstRefCasted test I like this version better as it moves the assertion that matters back into python.
2020-12-16 00:53:55 +00:00
m.def("takes_ref_wrap", [](std::reference_wrapper<ConstRefCasted> x) { return x.get().tag; });
m.def("takes_const_ptr", [](const ConstRefCasted *x) { return x->tag; });
m.def("takes_const_ref", [](const ConstRefCasted &x) { return x.tag; });
m.def("takes_const_ref_wrap",
[](std::reference_wrapper<const ConstRefCasted> x) { return x.get().tag; });
// test return_value_policy::_return_as_bytes
m.def(
"invalid_utf8_string_as_bytes",
[]() { return std::string("\xba\xd0\xba\xd0"); },
py::return_value_policy::_return_as_bytes);
m.def("invalid_utf8_string_as_str", []() { return std::string("\xba\xd0\xba\xd0"); });
m.def(
"invalid_utf8_char_array_as_bytes",
[]() { return "\xba\xd0\xba\xd0"; },
py::return_value_policy::_return_as_bytes);
py::class_<StringAttr>(m, "StringAttr")
.def(py::init<std::string>())
.def_property(
"value",
py::cpp_function([](StringAttr &self) { return self.value; },
py::return_value_policy::_return_as_bytes),
py::cpp_function([](StringAttr &self, std::string v) { self.value = std::move(v); }));
#ifdef PYBIND11_HAS_STRING_VIEW
m.def(
"invalid_utf8_string_view_as_bytes",
[]() { return std::string_view("\xba\xd0\xba\xd0"); },
py::return_value_policy::_return_as_bytes);
#endif
}