pybind11/tests/test_sequences_and_iterators.cpp

501 lines
19 KiB
C++
Raw Normal View History

2015-07-05 18:05:44 +00:00
/*
tests/test_sequences_and_iterators.cpp -- supporting Pythons' sequence protocol, iterators,
etc.
2015-07-05 18:05:44 +00:00
Copyright (c) 2016 Wenzel Jakob <wenzel.jakob@epfl.ch>
2015-07-05 18:05:44 +00:00
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_tests.h"
#include "constructor_stats.h"
#include <pybind11/operators.h>
#include <pybind11/stl.h>
2015-07-05 18:05:44 +00:00
#include <algorithm>
#include <utility>
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
#include <vector>
#ifdef PYBIND11_HAS_OPTIONAL
#include <optional>
#endif // PYBIND11_HAS_OPTIONAL
2016-08-24 22:30:00 +00:00
template<typename T>
class NonZeroIterator {
const T* ptr_;
public:
explicit NonZeroIterator(const T *ptr) : ptr_(ptr) {}
2016-08-24 22:30:00 +00:00
const T& operator*() const { return *ptr_; }
NonZeroIterator& operator++() { ++ptr_; return *this; }
};
class NonZeroSentinel {};
template<typename A, typename B>
bool operator==(const NonZeroIterator<std::pair<A, B>>& it, const NonZeroSentinel&) {
return !(*it).first || !(*it).second;
}
/* Iterator where dereferencing returns prvalues instead of references. */
template<typename T>
class NonRefIterator {
const T* ptr_;
public:
explicit NonRefIterator(const T *ptr) : ptr_(ptr) {}
T operator*() const { return T(*ptr_); }
NonRefIterator& operator++() { ++ptr_; return *this; }
bool operator==(const NonRefIterator &other) const { return ptr_ == other.ptr_; }
};
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
class NonCopyableInt {
public:
explicit NonCopyableInt(int value) : value_(value) {}
NonCopyableInt(const NonCopyableInt &) = delete;
NonCopyableInt(NonCopyableInt &&other) noexcept : value_(other.value_) {
other.value_ = -1; // detect when an unwanted move occurs
}
NonCopyableInt &operator=(const NonCopyableInt &) = delete;
NonCopyableInt &operator=(NonCopyableInt &&other) noexcept {
value_ = other.value_;
other.value_ = -1; // detect when an unwanted move occurs
return *this;
}
int get() const { return value_; }
void set(int value) { value_ = value; }
~NonCopyableInt() = default;
private:
int value_;
};
using NonCopyableIntPair = std::pair<NonCopyableInt, NonCopyableInt>;
PYBIND11_MAKE_OPAQUE(std::vector<NonCopyableInt>);
PYBIND11_MAKE_OPAQUE(std::vector<NonCopyableIntPair>);
template <typename PythonType>
py::list test_random_access_iterator(PythonType x) {
if (x.size() < 5)
throw py::value_error("Please provide at least 5 elements for testing.");
auto checks = py::list();
auto assert_equal = [&checks](py::handle a, py::handle b) {
auto result = PyObject_RichCompareBool(a.ptr(), b.ptr(), Py_EQ);
if (result == -1) { throw py::error_already_set(); }
checks.append(result != 0);
};
auto it = x.begin();
assert_equal(x[0], *it);
assert_equal(x[0], it[0]);
assert_equal(x[1], it[1]);
assert_equal(x[1], *(++it));
assert_equal(x[1], *(it++));
assert_equal(x[2], *it);
assert_equal(x[3], *(it += 1));
assert_equal(x[2], *(--it));
assert_equal(x[2], *(it--));
assert_equal(x[1], *it);
assert_equal(x[0], *(it -= 1));
assert_equal(it->attr("real"), x[0].attr("real"));
assert_equal((it + 1)->attr("real"), x[1].attr("real"));
assert_equal(x[1], *(it + 1));
assert_equal(x[1], *(1 + it));
it += 3;
assert_equal(x[1], *(it - 2));
checks.append(static_cast<std::size_t>(x.end() - x.begin()) == x.size());
checks.append((x.begin() + static_cast<std::ptrdiff_t>(x.size())) == x.end());
checks.append(x.begin() < x.end());
return checks;
}
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
TEST_SUBMODULE(sequences_and_iterators, m) {
// test_sliceable
class Sliceable{
public:
explicit Sliceable(int n) : size(n) {}
int start, stop, step;
int size;
};
py::class_<Sliceable>(m, "Sliceable")
.def(py::init<int>())
.def("__getitem__", [](const Sliceable &s, const py::slice &slice) {
py::ssize_t start = 0, stop = 0, step = 0, slicelength = 0;
if (!slice.compute(s.size, &start, &stop, &step, &slicelength))
throw py::error_already_set();
int istart = static_cast<int>(start);
int istop = static_cast<int>(stop);
int istep = static_cast<int>(step);
return std::make_tuple(istart, istop, istep);
});
2016-08-24 22:30:00 +00:00
m.def("make_forward_slice_size_t", []() { return py::slice(0, -1, 1); });
m.def("make_reversed_slice_object", []() { return py::slice(py::none(), py::none(), py::int_(-1)); });
#ifdef PYBIND11_HAS_OPTIONAL
m.attr("has_optional") = true;
m.def("make_reversed_slice_size_t_optional_verbose", []() { return py::slice(std::nullopt, std::nullopt, -1); });
// Warning: The following spelling may still compile if optional<> is not present and give wrong answers.
// Please use with caution.
m.def("make_reversed_slice_size_t_optional", []() { return py::slice({}, {}, -1); });
#else
m.attr("has_optional") = false;
#endif
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
// test_sequence
class Sequence {
public:
explicit Sequence(size_t size) : m_size(size) {
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
print_created(this, "of size", m_size);
m_data = new float[size];
memset(m_data, 0, sizeof(float) * size);
}
explicit Sequence(const std::vector<float> &value) : m_size(value.size()) {
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
print_created(this, "of size", m_size, "from std::vector");
m_data = new float[m_size];
memcpy(m_data, &value[0], sizeof(float) * m_size);
}
Sequence(const Sequence &s) : m_size(s.m_size) {
print_copy_created(this);
m_data = new float[m_size];
memcpy(m_data, s.m_data, sizeof(float)*m_size);
}
Sequence(Sequence &&s) noexcept : m_size(s.m_size), m_data(s.m_data) {
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
print_move_created(this);
s.m_size = 0;
s.m_data = nullptr;
}
~Sequence() { print_destroyed(this); delete[] m_data; }
Sequence &operator=(const Sequence &s) {
if (&s != this) {
delete[] m_data;
m_size = s.m_size;
m_data = new float[m_size];
memcpy(m_data, s.m_data, sizeof(float)*m_size);
}
print_copy_assigned(this);
return *this;
}
Sequence &operator=(Sequence &&s) noexcept {
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
if (&s != this) {
delete[] m_data;
m_size = s.m_size;
m_data = s.m_data;
s.m_size = 0;
s.m_data = nullptr;
}
print_move_assigned(this);
return *this;
}
bool operator==(const Sequence &s) const {
if (m_size != s.size()) return false;
for (size_t i = 0; i < m_size; ++i)
if (m_data[i] != s[i])
return false;
return true;
}
bool operator!=(const Sequence &s) const { return !operator==(s); }
float operator[](size_t index) const { return m_data[index]; }
float &operator[](size_t index) { return m_data[index]; }
2015-07-05 18:05:44 +00:00
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
bool contains(float v) const {
for (size_t i = 0; i < m_size; ++i)
if (v == m_data[i])
return true;
return false;
}
Sequence reversed() const {
Sequence result(m_size);
for (size_t i = 0; i < m_size; ++i)
result[m_size - i - 1] = m_data[i];
return result;
}
size_t size() const { return m_size; }
const float *begin() const { return m_data; }
const float *end() const { return m_data+m_size; }
private:
size_t m_size;
float *m_data;
};
py::class_<Sequence>(m, "Sequence")
.def(py::init<size_t>())
.def(py::init<const std::vector<float> &>())
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
/// Bare bones interface
.def("__getitem__",
[](const Sequence &s, size_t i) {
if (i >= s.size())
throw py::index_error();
return s[i];
})
.def("__setitem__",
[](Sequence &s, size_t i, float v) {
if (i >= s.size())
throw py::index_error();
s[i] = v;
})
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
.def("__len__", &Sequence::size)
/// Optional sequence protocol operations
.def(
"__iter__",
[](const Sequence &s) { return py::make_iterator(s.begin(), s.end()); },
py::keep_alive<0, 1>() /* Essential: keep object alive while iterator exists */)
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
.def("__contains__", [](const Sequence &s, float v) { return s.contains(v); })
.def("__reversed__", [](const Sequence &s) -> Sequence { return s.reversed(); })
/// Slicing protocol (optional)
.def("__getitem__",
[](const Sequence &s, const py::slice &slice) -> Sequence * {
size_t start = 0, stop = 0, step = 0, slicelength = 0;
if (!slice.compute(s.size(), &start, &stop, &step, &slicelength))
throw py::error_already_set();
auto *seq = new Sequence(slicelength);
for (size_t i = 0; i < slicelength; ++i) {
(*seq)[i] = s[start];
start += step;
}
return seq;
})
.def("__setitem__",
[](Sequence &s, const py::slice &slice, const Sequence &value) {
size_t start = 0, stop = 0, step = 0, slicelength = 0;
if (!slice.compute(s.size(), &start, &stop, &step, &slicelength))
throw py::error_already_set();
if (slicelength != value.size())
throw std::runtime_error(
"Left and right hand size of slice assignment have different sizes!");
for (size_t i = 0; i < slicelength; ++i) {
s[start] = value[i];
start += step;
}
})
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
/// Comparisons
.def(py::self == py::self)
.def(py::self != py::self)
// Could also define py::self + py::self for concatenation, etc.
;
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
// test_map_iterator
// Interface of a map-like object that isn't (directly) an unordered_map, but provides some basic
// map-like functionality.
class StringMap {
public:
StringMap() = default;
explicit StringMap(std::unordered_map<std::string, std::string> init)
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
: map(std::move(init)) {}
void set(const std::string &key, std::string val) { map[key] = std::move(val); }
std::string get(const std::string &key) const { return map.at(key); }
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
size_t size() const { return map.size(); }
private:
std::unordered_map<std::string, std::string> map;
public:
decltype(map.cbegin()) begin() const { return map.cbegin(); }
decltype(map.cend()) end() const { return map.cend(); }
};
py::class_<StringMap>(m, "StringMap")
.def(py::init<>())
.def(py::init<std::unordered_map<std::string, std::string>>())
.def("__getitem__",
[](const StringMap &map, const std::string &key) {
try {
return map.get(key);
} catch (const std::out_of_range &) {
throw py::key_error("key '" + key + "' does not exist");
}
})
.def("__setitem__", &StringMap::set)
.def("__len__", &StringMap::size)
.def(
"__iter__",
[](const StringMap &map) { return py::make_key_iterator(map.begin(), map.end()); },
py::keep_alive<0, 1>())
.def(
"items",
[](const StringMap &map) { return py::make_iterator(map.begin(), map.end()); },
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
py::keep_alive<0, 1>())
.def(
"values",
[](const StringMap &map) { return py::make_value_iterator(map.begin(), map.end()); },
py::keep_alive<0, 1>());
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
// test_generalized_iterators
class IntPairs {
public:
explicit IntPairs(std::vector<std::pair<int, int>> data) : data_(std::move(data)) {}
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
const std::pair<int, int>* begin() const { return data_.data(); }
// .end() only required for py::make_iterator(self) overload
const std::pair<int, int>* end() const { return data_.data() + data_.size(); }
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
private:
std::vector<std::pair<int, int>> data_;
};
2016-08-24 22:30:00 +00:00
py::class_<IntPairs>(m, "IntPairs")
.def(py::init<std::vector<std::pair<int, int>>>())
.def("nonzero", [](const IntPairs& s) {
return py::make_iterator(NonZeroIterator<std::pair<int, int>>(s.begin()), NonZeroSentinel());
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
}, py::keep_alive<0, 1>())
2016-08-24 22:30:00 +00:00
.def("nonzero_keys", [](const IntPairs& s) {
return py::make_key_iterator(NonZeroIterator<std::pair<int, int>>(s.begin()), NonZeroSentinel());
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
}, py::keep_alive<0, 1>())
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
.def("nonzero_values", [](const IntPairs& s) {
return py::make_value_iterator(NonZeroIterator<std::pair<int, int>>(s.begin()), NonZeroSentinel());
}, py::keep_alive<0, 1>())
// test iterator that returns values instead of references
.def("nonref", [](const IntPairs& s) {
return py::make_iterator(NonRefIterator<std::pair<int, int>>(s.begin()),
NonRefIterator<std::pair<int, int>>(s.end()));
}, py::keep_alive<0, 1>())
.def("nonref_keys", [](const IntPairs& s) {
return py::make_key_iterator(NonRefIterator<std::pair<int, int>>(s.begin()),
NonRefIterator<std::pair<int, int>>(s.end()));
}, py::keep_alive<0, 1>())
.def("nonref_values", [](const IntPairs& s) {
return py::make_value_iterator(NonRefIterator<std::pair<int, int>>(s.begin()),
NonRefIterator<std::pair<int, int>>(s.end()));
}, py::keep_alive<0, 1>())
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
// test single-argument make_iterator
.def("simple_iterator", [](IntPairs& self) {
return py::make_iterator(self);
}, py::keep_alive<0, 1>())
.def("simple_keys", [](IntPairs& self) {
return py::make_key_iterator(self);
}, py::keep_alive<0, 1>())
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
.def("simple_values", [](IntPairs& self) {
return py::make_value_iterator(self);
}, py::keep_alive<0, 1>())
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
// Test iterator with an Extra (doesn't do anything useful, so not used
// at runtime, but tests need to be able to compile with the correct
// overload. See PR #3293.
.def("_make_iterator_extras", [](IntPairs& self) {
return py::make_iterator(self, py::call_guard<int>());
}, py::keep_alive<0, 1>())
.def("_make_key_extras", [](IntPairs& self) {
return py::make_key_iterator(self, py::call_guard<int>());
}, py::keep_alive<0, 1>())
.def("_make_value_extras", [](IntPairs& self) {
return py::make_value_iterator(self, py::call_guard<int>());
}, py::keep_alive<0, 1>())
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
;
2016-08-24 22:30:00 +00:00
feat: reapply fixed version of #3271 (#3293) * Add make_value_iterator (#3271) * Add make_value_iterator This is the counterpart to make_key_iterator, and will allow implementing a `value` method in `bind_map` (although doing so is left for a subsequent PR). I made a few design changes to reduce copy-and-paste boilerplate. Previously detail::iterator_state had a boolean template parameter to indicate whether it was being used for make_iterator or make_key_iterator. I replaced the boolean with a class that determines how to dereference the iterator. This allows for a generic implementation of `__next__`. I also added the ValueType and Extra... parameters to the iterator_state template args, because I think it was a bug that they were missing: if make_iterator is called twice with different values of these, only the first set has effect (because the state class is only registered once). There is still a potential issue in that the *values* of the extra arguments are latched on the first call, but since most policies are empty classes this should be even less common. * Add some remove_cv_t to appease clang-tidy * Make iterator_access and friends take reference For some reason I'd accidentally made it take a const value, which caused some issues with third-party packages. * Another attempt to remove remove_cv_t from iterators Some of the return types were const (non-reference) types because of the pecularities of decltype: `decltype((*it).first)` is the *declared* type of the member of the pair, rather than the type of the expression. So if the reference type of the iterator is `pair<const int, int> &`, then the decltype is `const int`. Wrapping an extra set of parentheses to form `decltype(((*it).first))` would instead give `const int &`. This means that the existing make_key_iterator actually returns by value from `__next__`, rather than by reference. Since for mapping types, keys are always const, this probably hasn't been noticed, but it will affect make_value_iterator if the Python code tries to mutate the returned objects. I've changed things to use double parentheses so that make_iterator, make_key_iterator and make_value_iterator should now all return the reference type of the iterator. I'll still need to add a test for that; for now I'm just checking whether I can keep Clang-Tidy happy. * Add back some NOLINTNEXTLINE to appease Clang-Tidy This is favoured over using remove_cv_t because in some cases a const value return type is deliberate (particularly for Eigen). * Add a unit test for iterator referencing Ensure that make_iterator, make_key_iterator and make_value_iterator return references to the container elements, rather than copies. The test for make_key_iterator fails to compile on master, which gives me confidence that this branch has fixed it. * Make the iterator_access etc operator() const I'm actually a little surprised it compiled at all given that the operator() is called on a temporary, but I don't claim to fully understand all the different value types in C++11. * Attempt to work around compiler bugs https://godbolt.org/ shows an example where ICC gets the wrong result for a decltype used as the default for a template argument, and CI also showed problems with PGI. This is a shot in the dark to see if it fixes things. * Make a test constructor explicit (Clang-Tidy) * Fix unit test on GCC 4.8.5 It seems to require the arguments to the std::pair constructor to be implicitly convertible to the types in the pair, rather than just requiring is_constructible. * Remove DOXYGEN_SHOULD_SKIP_THIS guards Now that a complex decltype expression has been replaced by a simpler nested type, I'm hoping Doxygen will be able to build it without issues. * Add comment to explain iterator_state template params * fix: regression in #3271 Co-authored-by: Bruce Merry <1963944+bmerry@users.noreply.github.com>
2021-09-23 19:06:07 +00:00
// test_iterater_referencing
py::class_<NonCopyableInt>(m, "NonCopyableInt")
.def(py::init<int>())
.def("set", &NonCopyableInt::set)
.def("__int__", &NonCopyableInt::get)
;
py::class_<std::vector<NonCopyableInt>>(m, "VectorNonCopyableInt")
.def(py::init<>())
.def("append", [](std::vector<NonCopyableInt> &vec, int value) {
vec.emplace_back(value);
})
.def("__iter__", [](std::vector<NonCopyableInt> &vec) {
return py::make_iterator(vec.begin(), vec.end());
})
;
py::class_<std::vector<NonCopyableIntPair>>(m, "VectorNonCopyableIntPair")
.def(py::init<>())
.def("append", [](std::vector<NonCopyableIntPair> &vec, const std::pair<int, int> &value) {
vec.emplace_back(NonCopyableInt(value.first), NonCopyableInt(value.second));
})
.def("keys", [](std::vector<NonCopyableIntPair> &vec) {
return py::make_key_iterator(vec.begin(), vec.end());
})
.def("values", [](std::vector<NonCopyableIntPair> &vec) {
return py::make_value_iterator(vec.begin(), vec.end());
})
;
#if 0
// Obsolete: special data structure for exposing custom iterator types to python
// kept here for illustrative purposes because there might be some use cases which
// are not covered by the much simpler py::make_iterator
struct PySequenceIterator {
PySequenceIterator(const Sequence &seq, py::object ref) : seq(seq), ref(ref) { }
float next() {
if (index == seq.size())
throw py::stop_iteration();
return seq[index++];
}
const Sequence &seq;
py::object ref; // keep a reference
size_t index = 0;
};
2015-07-05 18:05:44 +00:00
py::class_<PySequenceIterator>(seq, "Iterator")
.def("__iter__", [](PySequenceIterator &it) -> PySequenceIterator& { return it; })
.def("__next__", &PySequenceIterator::next);
On the actual Sequence object, the iterator would be constructed as follows:
.def("__iter__", [](py::object s) { return PySequenceIterator(s.cast<const Sequence &>(), s); })
#endif
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
// test_python_iterator_in_cpp
m.def("object_to_list", [](const py::object &o) {
auto l = py::list();
for (auto item : o) {
l.append(item);
}
return l;
});
m.def("iterator_to_list", [](py::iterator it) {
auto l = py::list();
while (it != py::iterator::sentinel()) {
l.append(*it);
++it;
}
return l;
});
// test_sequence_length: check that Python sequences can be converted to py::sequence.
m.def("sequence_length", [](const py::sequence &seq) { return seq.size(); });
// Make sure that py::iterator works with std algorithms
m.def("count_none", [](const py::object &o) {
return std::count_if(o.begin(), o.end(), [](py::handle h) { return h.is_none(); });
});
m.def("find_none", [](const py::object &o) {
auto it = std::find_if(o.begin(), o.end(), [](py::handle h) { return h.is_none(); });
return it->is_none();
});
m.def("count_nonzeros", [](const py::dict &d) {
return std::count_if(d.begin(), d.end(), [](std::pair<py::handle, py::handle> p) {
return p.second.cast<int>() != 0;
});
});
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
m.def("tuple_iterator", &test_random_access_iterator<py::tuple>);
m.def("list_iterator", &test_random_access_iterator<py::list>);
m.def("sequence_iterator", &test_random_access_iterator<py::sequence>);
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
// test_iterator_passthrough
// #181: iterator passthrough did not compile
m.def("iterator_passthrough", [](py::iterator s) -> py::iterator {
return py::make_iterator(std::begin(s), std::end(s));
});
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
// test_iterator_rvp
// #388: Can't make iterators via make_iterator() with different r/v policies
static std::vector<int> list = { 1, 2, 3 };
m.def("make_iterator_1", []() { return py::make_iterator<py::return_value_policy::copy>(list); });
m.def("make_iterator_2", []() { return py::make_iterator<py::return_value_policy::automatic>(list); });
Update all remaining tests to new test styles This udpates all the remaining tests to the new test suite code and comment styles started in #898. For the most part, the test coverage here is unchanged, with a few minor exceptions as noted below. - test_constants_and_functions: this adds more overload tests with overloads with different number of arguments for more comprehensive overload_cast testing. The test style conversion broke the overload tests under MSVC 2015, prompting the additional tests while looking for a workaround. - test_eigen: this dropped the unused functions `get_cm_corners` and `get_cm_corners_const`--these same tests were duplicates of the same things provided (and used) via ReturnTester methods. - test_opaque_types: this test had a hidden dependence on ExampleMandA which is now fixed by using the global UserType which suffices for the relevant test. - test_methods_and_attributes: this required some additions to UserType to make it usable as a replacement for the test's previous SimpleType: UserType gained a value mutator, and the `value` property is not mutable (it was previously readonly). Some overload tests were also added to better test overload_cast (as described above). - test_numpy_array: removed the untemplated mutate_data/mutate_data_t: the templated versions with an empty parameter pack expand to the same thing. - test_stl: this was already mostly in the new style; this just tweaks things a bit, localizing a class, and adding some missing `// test_whatever` comments. - test_virtual_functions: like `test_stl`, this was mostly in the new test style already, but needed some `// test_whatever` comments. This commit also moves the inherited virtual example code to the end of the file, after the main set of tests (since it is less important than the other tests, and rather length); it also got renamed to `test_inherited_virtuals` (from `test_inheriting_repeat`) because it tests both inherited virtual approaches, not just the repeat approach.
2017-07-25 20:47:36 +00:00
}