This adds support for a `py::args_kw_only()` annotation that can be
specified between `py::arg` annotations to indicate that any following
arguments are keyword-only. This allows you to write:
m.def("f", [](int a, int b) { /* ... */ },
py::arg("a"), py::args_kw_only(), py::arg("b"));
and have it work like Python 3's:
def f(a, *, b):
# ...
with respect to how `a` and `b` arguments are accepted (that is, `a` can
be positional or by keyword; `b` can only be specified by keyword).
* test pair-copyability on C++17 upwards
The stdlib falsely detects containers like M=std::map<T, U>
as copyable, even when one of T and U is not copyable.
Therefore we cannot rely on the stdlib dismissing std::pair<T, M>
by itself, even on C++17.
* fix is_copy_assignable
bind_map used std::is_copy_assignable which suffers from the same problems
as std::is_copy_constructible, therefore the same fix has been applied.
* created tests for copyability
* fix: Avoid conversion to `int_` rhs argument of enum eq/ne
* test: compare unscoped enum with strings
* suppress comparison to None warning
* test unscoped enum arithmetic and comparision with unsupported type
* Make `overload_cast_impl` available in C++11 mode.
Narrow the scope of the `#if defined(PYBIND11_CPP14)` block around overload_cast to only
cover the parts where C++14 is stricly required. Thus, the implementation in
`pybind11::details::overload_cast_impl` is still available in C++11 mode.
* PR #1581: Modify test to use overload_cast_impl, update docs and change log
This is only necessary if `get_internals` is called for the first time in a given module when the running thread is in a GIL-released state.
Fixes#1364
* Test dtype field order in numpy dtype tests
When running tests with NumPy 1.14 or later this test exposes the
"invalid buffer descriptor" error reported in #1274.
* Create dtype_ptr with ordered fields
* Fix casting of time points with non-system-clock duration on Windows
Add explicit `time_point_cast` to time point with duration of system
clock. Fixes Visual Studio compile error.
* Add test case for custom time points casting
In def_readonly and def_readwrite, there is an assertion that the member comes
from the class or a base class:
static_assert(std::is_base_of<C, type>::value, "...");
However, if C and type are the same type, is_base_of will still only be true
if they are the same _non-union_ type. This means we can't define accessors
for the members of a union type because of this assertion.
Update the assertion to test
std::is_same<C, type>::value || std::is_base_of<C, type>::value
which will allow union types, or members of base classes.
Also add a basic unit test for accessing unions.
* Fix async Python functors invoking from multiple C++ threads (#1587)
Ensure GIL is held during functor destruction.
* Add async Python callbacks test that runs in separate Python thread
I found that the numpy array tests already contained an empty-shaped
array test, but none with data in it.
Following PEP 3118, scalars have an empty shape and ndim 0. This
works already and is now also documented/covered by a test.
* Fixing order of arguments in call to PyErr_GivenExceptionMatches in pybind11::error_already_set.matches
* Added tests on error_already_set::matches fix for exception base classes
* Fix warning that not including a cmake source or build dir will be a fatal error (it is now on newest CMakes)
* Fixes appveyor
* Travis uses CMake 3.9 for more than a year now
* Travis dropped sudo: false in December
* Dropping Sphinx 2
- clang7: Suppress self-assign warnings; fix missing virtual dtors
- pypy:
- Keep old version (newer stuff breaks)
- Pin packages to extra index for speed
- travis:
- Make docker explicit; remove docker if not needed
- Make commands more verbose (for debugging / repro)
- Make Ubuntu dist explicit per job
- Fix Windows
- Add names to travis
This avoids GIL deadlocking when pybind11 tries to acquire the GIL in a thread that already acquired it using standard Python API (e.g. when running from a Python thread).
* Adds std::deque to the types supported by list_caster in stl.h.
* Adds a new test_deque test in test_stl.{py,cpp}.
* Updates the documentation to include std::deque as a default
supported type.
* Check default holder
-Recognize "std::unique_ptr<T, D>" as a default holder even if "D" doesn't match between base and derived holders
* Add test for unique_ptr<T, D> change
Pybind11 provides a cast operator between opaque void* pointers on the
C++ side and capsules on the Python side. The py::cast<void *>
expression was not aware of this possibility and incorrectly triggered a
compile-time assertion ("Unable to cast type to reference: value is
local to type caster") that is now fixed.
* Support C++17 aligned new statement
This patch makes pybind11 aware of nonstandard alignment requirements in
bound types and passes on this information to C++17 aligned 'new'
operator. Pre-C++17, the behavior is unchanged.
This PR brings the std::array<> caster in sync with the other STL type
casters: to accept an arbitrary sequence as input (rather than a list,
which is too restrictive).
* Fix for Issue #1258
list_caster::load method will now check for a Python string and prevent its automatic conversion to a list.
This should fix the issue "pybind11/stl.h converts string to vector<string> #1258" (https://github.com/pybind/pybind11/issues/1258)
* Added tests for fix of issue #1258
* Changelog: stl string auto-conversion
* Fix potential crash when calling an overloaded function
The crash would occur if:
- dispatcher() uses two-pass logic (because the target is overloaded and some arguments support conversions)
- the first pass (with conversions disabled) doesn't find any matching overload
- the second pass does find a matching overload, but its return value can't be converted to Python
The code for formatting the error message assumed `it` still pointed to the selected overload,
but during the second-pass loop `it` was nullptr. Fix by setting `it` correctly if a second-pass
call returns a nullptr `handle`. Add a new test that segfaults without this fix.
* Make overload iteration const-correct so we don't have to iterate again on second-pass error
* Change test_error_after_conversions dependencies to local classes/variables
This commit addresses an inefficiency in how enums are created in
pybind11. Most of the enum_<> implementation is completely generic --
however, being a template class, it ended up instantiating vast amounts
of essentially identical code in larger projects with many enums.
This commit introduces a generic non-templated helper class that is
compatible with any kind of enumeration. enum_ then becomes a thin
wrapper around this new class.
The new enum_<> API is designed to be 100% compatible with the old one.
object_api::operator[] has a powerful overload for py::handle that can
accept slices, tuples (for NumPy), etc.
Lists, sequences, and tuples provide their own specialized operator[],
which unfortunately disables this functionality. This is accidental, and
the purpose of this commit is to re-enable the more general behavior.
This commit is tangentially related to the previous one in that it makes
py::handle/py::object et al. behave more like their Python counterparts.
This commit revamps the object_api class so that it maps most C++
operators to their Python analogs. This makes it possible to, e.g.
perform arithmetic using a py::int_ or py::array.
* check for already existing enum value added; added test
* added enum value name to exception message
* test for defining enum with multiple identical names moved to test_enum.cpp/py
This PR adds a new py::ellipsis() method which can be used in
conjunction with NumPy's generalized slicing support. For instance,
the following is now valid (where "a" is a NumPy array):
py::array b = a[py::make_tuple(0, py::ellipsis(), 0)];
Catch v2 changed the `run(...)` signature to take a `char *argv[]`,
arguing partly that technically a `char *argv[]` type is the correct
`main()` signature rather than `const char *argv[]`.
Dropping the `const` here doesn't appear to cause any problems with
catch v1 (tested against both the cmake-downloaded 1.9.3 and Debian's
1.12.1 package) so we can follow suit.
* stl.h: propagate return value policies to type-specific casters
Return value policies for containers like those handled in in 'stl.h'
are currently broken.
The problem is that detail::return_value_policy_override<C>::policy()
always returns 'move' when given a non-pointer/reference type, e.g.
'std::vector<...>'.
This is sensible behavior for custom types that are exposed via
'py::class_<>', but it does not make sense for types that are handled by
other type casters (STL containers, Eigen matrices, etc.).
This commit changes the behavior so that
detail::return_value_policy_override only becomes active when the type
caster derives from type_caster_generic.
Furthermore, the override logic is called recursively in STL type
casters to enable key/value-specific behavior.
* Add basic support for tag-based static polymorphism
Sometimes it is possible to look at a C++ object and know what its dynamic type is,
even if it doesn't use C++ polymorphism, because instances of the object and its
subclasses conform to some other mechanism for being self-describing; for example,
perhaps there's an enumerated "tag" or "kind" member in the base class that's always
set to an indication of the correct type. This might be done for performance reasons,
or to permit most-derived types to be trivially copyable. One of the most widely-known
examples is in LLVM: https://llvm.org/docs/HowToSetUpLLVMStyleRTTI.html
This PR permits pybind11 to be informed of such conventions via a new specializable
detail::polymorphic_type_hook<> template, which generalizes the previous logic for
determining the runtime type of an object based on C++ RTTI. Implementors provide
a way to map from a base class object to a const std::type_info* for the dynamic
type; pybind11 then uses this to ensure that casting a Base* to Python creates a
Python object that knows it's wrapping the appropriate sort of Derived.
There are a number of restrictions with this tag-based static polymorphism support
compared to pybind11's existing support for built-in C++ polymorphism:
- there is no support for this-pointer adjustment, so only single inheritance is permitted
- there is no way to make C++ code call new Python-provided subclasses
- when binding C++ classes that redefine a method in a subclass, the .def() must be
repeated in the binding for Python to know about the update
But these are not much of an issue in practice in many cases, the impact on the
complexity of pybind11's innards is minimal and localized, and the support for
automatic downcasting improves usability a great deal.
The property returns the enum_ value as a string.
For example:
>>> import module
>>> module.enum.VALUE
enum.VALUE
>>> str(module.enum.VALUE)
'enum.VALUE'
>>> module.enum.VALUE.name
'VALUE'
This is actually the equivalent of Boost.Python "name" property.
- PYBIND11_MAKE_OPAQUE now takes ... rather than a single argument and
expands it with __VA_ARGS__; this lets templated, comma-containing
types get through correctly.
- Adds a new macro PYBIND11_TYPE() that lets you pass the type into a
macro as a single argument, such as:
PYBIND11_OVERLOAD(PYBIND11_TYPE(R<1,2>), PYBIND11_TYPE(C<3,4>), func)
Unfortunately this only works for one macro call: to forward the
argument on to the next macro call (without the processor breaking it
up again) requires also adding the PYBIND11_TYPE(...) to type macro
arguments in the PYBIND11_OVERLOAD_... macro chain.
- updated the documentation with these two changes, and use them at a couple
places in the test suite to test that they work.
This fixes the test code on big-endian architectures: the array support
(PR #832) had hard-coded the little-endian '<' but we need to use '>' on
big-endian architectures.
This updates the `py::init` constructors to only use brace
initialization for aggregate initiailization if there is no constructor
with the given arguments.
This, in particular, fixes the regression in #1247 where the presence of
a `std::initializer_list<T>` constructor started being invoked for
constructor invocations in 2.2 even when there was a specific
constructor of the desired type.
The added test case demonstrates: without this change, it fails to
compile because the `.def(py::init<std::vector<int>>())` constructor
tries to invoke the `T(std::initializer_list<std::vector<int>>)`
constructor rather than the `T(std::vector<int>)` constructor.
By only using `new T{...}`-style construction when a `T(...)`
constructor doesn't exist, we should bypass this by while still allowing
`py::init<...>` to be used for aggregate type initialization (since such
types, by definition, don't have a user-declared constructor).
* Fix segfault when reloading interpreter with external modules
When embedding the interpreter and loading external modules in that
embedded interpreter, the external module correctly shares its
internals_ptr with the one in the embedded interpreter. When the
interpreter is shut down, however, only the `internals_ptr` local to
the embedded code is actually reset to nullptr: the external module
remains set.
The result is that loading an external pybind11 module, letting the
interpreter go through a finalize/initialize, then attempting to use
something in the external module fails because this external module is
still trying to use the old (destroyed) internals. This causes
undefined behaviour (typically a segfault).
This commit fixes it by adding a level of indirection in the internals
path, converting the local internals variable to `internals **` instead
of `internals *`. With this change, we can detect a stale internals
pointer and reload the internals pointer (either from a capsule or by
creating a new internals instance).
(No issue number: this was reported on gitter by @henryiii and @aoloe).
- UPDATEIFCOPY is deprecated, replaced with similar (but not identical)
WRITEBACKIFCOPY; trying to access the flag causes a deprecation
warning under numpy 1.14, so just check the new flag there.
- Numpy `repr` formatting of floats changed in 1.14.0 to `[1., 2., 3.]`
instead of the pre-1.14 `[ 1., 2., 3.]`. Updated the tests to
check for equality with the `repr(...)` value rather than the
hard-coded (and now version-dependent) string representation.
PEP8 indicates (correctly, IMO) that when an annotation is present, the
signature should include spaces around the equal sign, i.e.
def f(x: int = 1): ...
instead of
def f(x: int=1): ...
(in the latter case the equal appears to bind to the type, not to the
argument).
pybind11 signatures always includes a type annotation so we can always
add the spaces.
In the latest MSVC in C++17 mode including Eigen causes warnings:
warning C4996: 'std::unary_negate<_Fn>': warning STL4008: std::not1(),
std::not2(), std::unary_negate, and std::binary_negate are deprecated in
C++17. They are superseded by std::not_fn(). You can define
_SILENCE_CXX17_NEGATORS_DEPRECATION_WARNING or
_SILENCE_ALL_CXX17_DEPRECATION_WARNINGS to acknowledge that you have
received this warning.
This disables 4996 for the Eigen includes.
Catch generates a similar warning for std::uncaught_exception, so
disable the warning there, too.
In both cases this is temporary; we can (and should) remove the warnings
disabling once new upstream versions of Eigen and Catch are available
that address the warning. (The Catch one, in particular, looks to be
fixed in upstream master, so will probably be fixed in the next (2.0.2)
release).
Pybind11's default conversion to int always produces a long on Python 2 (`int`s and `long`s were unified in Python 3). This patch fixes `int` handling to match Python 2 on Python 2; for short types (`size_t` or smaller), the number will be returned as an `int` if possible, otherwise `long`. Requires Python 2.5+.
This is needed for things like `sys.exit`, which refuse to accept a `long`.
This commit turns on `-Wdeprecated` in the test suite and fixes several
associated deprecation warnings that show up as a result:
- in C++17 `static constexpr` members are implicitly inline; our
redeclaration (needed for C++11/14) is deprecated in C++17.
- various test suite classes have destructors and rely on implicit copy
constructors, but implicit copy constructor definitions when a
user-declared destructor is present was deprecated in C++11.
- Eigen also has various implicit copy constructors, so just disable
`-Wdeprecated` in `eigen.h`.
py::class_<T>'s `def_property` and `def_property_static` can now take a
`nullptr` as the getter to allow a write-only property to be established
(mirroring Python's `property()` built-in when `None` is given for the
getter).
This also updates properties to use the new nullptr constructor internally.
A few fixes related to how we set `__qualname__` and how we show the
type name in function signatures:
- `__qualname__` isn't supposed to have the module name at the
beginning, but we've been putting it there. This removes it, while
keeping the `Nested.Class` name chaining.
- print `__module__.__qualname__` rather than `type->tp_name`; the
latter doesn't work properly for nested classes, so we would get
`module.B` rather than `module.A.B` for a class `B` with parent `A`.
This also unifies the Python 3 and PyPy code. Fixes#1166.
- This now sets a `__qualname__` attribute on the type (as would happen
in Python 3.3+) for Python <3.3, including PyPy. While not particularly
important to have in earlier Python versions, it's useful for us to be
able to extracted the nested name, which is why `__qualname__` was
invented in the first place.
- Added tests for the above.
The just-updated flake8 package hits a bunch of:
E741 ambiguous variable name 'l'
warnings. This commit renames them all from `l` to `lst` (they are all
list values) to avoid the error.
- For the debian/buster docker build (GCC 7/C++17) install and use the
system `catch` package; this also renames "COMPILER_PACKAGES" to
"EXTRA_PACKAGES" since it now contains a non-compiler package.
- Add a status message indicating the catch version being used for
compiling the embedded tests
- Simplify some bash code by using VAR+=" foo" to append (rather than
VAR="${VAR} foo"
- Fix CMAKE_INCLUDE_PATH appending: it was prepending the ':' but not
the existing $CMAKE_INCLUDE_PATH value and so would end up with
":/eigen-path" if CMAKE_INCLUDE_PATH was already set. (This wasn't
bug that was actually noticed since currently nothing else sets it).
This fixes a bug introduced in b68959e822
when passing in a two-dimensional, but conformable, array as the value
for a compile-time Eigen vector (such as VectorXd or RowVectorXd). The
commit switched to using numpy to copy into the eigen data, but this
broke the described case because numpy refuses to broadcast a (N,1)
into a (N).
This commit fixes it by squeezing the input array whenever the output
array is 1-dimensional, which will let the problematic case through.
(This shouldn't squeeze inappropriately as dimension compatibility is
already checked for conformability before getting to the copy code).
This changes the caster to return a reference to a (new) local `CharT`
type caster member so that binding lvalue-reference char arguments
works (currently it results in a compilation failure).
Fixes#1116
This also matches the Eigen example for the row-major case.
This also enhances one of the tests to trigger a failure (and fixes it in the PR). (This isn't really a flaw in pybind itself, but rather fixes wrong code in the test code and docs).
The current C++14 constexpr signatures don't require relaxed constexpr,
but only `auto` return type deduction. To get around this in C++11,
the type caster's `name()` static member functions are turned into
`static constexpr auto` variables.
E.g. trying to convert a `list` to a `std::vector<int>` without
including <pybind11/stl.h> will now raise an error with a note that
suggests checking the headers.
The note is only appended if `std::` is found in the function
signature. This should only be the case when a header is missing.
E.g. when stl.h is included, the signature would contain `List[int]`
instead of `std::vector<int>` while using stl_bind.h would produce
something like `MyVector`. Similarly for `std::map`/`Dict`, `complex`,
`std::function`/`Callable`, etc.
There's a possibility for false positives, but it's pretty low.
To avoid an ODR violation in the test suite while testing
both `stl.h` and `std_bind.h` with `std::vector<bool>`,
the `py::bind_vector<std::vector<bool>>` test is moved to
the secondary module (which does not include `stl.h`).
There are two separate additions:
1. `py::hash(obj)` is equivalent to the Python `hash(obj)`.
2. `.def(hash(py::self))` registers the hash function defined by
`std::hash<T>` as the Python hash function.
The lookup of the `self` type and value pointer are moved out of
template code and into `dispatcher`. This brings down the binary
size of constructors back to the level of the old placement-new
approach. (It also avoids a second lookup for `init_instance`.)
With this implementation, mixing old- and new-style constructors
in the same overload set may result in some runtime overhead for
temporary allocations/deallocations, but this should be fine as
old style constructors are phased out.
Creating an instance of of a pybind11-bound type caused a reference leak in the
associated Python type object, which could prevent these from being collected
upon interpreter shutdown. This commit fixes that issue for all types that are
defined in a scope (e.g. a module). Unscoped anonymous types (e.g. custom
iterator types) always retain a positive reference count to prevent their
collection.
The `latest` build remains as is, but all others are modified to:
* Use regular Python instead of conda. `pip install` is much faster
than conda, but scipy isn't available. Numpy is still tested.
* Compile in debug mode instead of release.
* Skip CMake build tests. For some reason, CMake configuration is very
slow on AppVeyor and these tests are almost entirely CMake.
The changes reduce build time to about 1/3 of the original. The `latest`
config still covers scipy, release mode and the CMake build tests, so
the others don't need to.
The main point of `py::module_local` is to make the C++ -> Python cast
unique so that returning/casting a C++ instance is well-defined.
Unfortunately it also makes loading unique, but this isn't particularly
desirable: when an instance contains `Type` instance there's no reason
it shouldn't be possible to pass that instance to a bound function
taking a `Type` parameter, even if that function is in another module.
This commit solves the issue by allowing foreign module (and global)
type loaders have a chance to load the value if the local module loader
fails. The implementation here does this by storing a module-local
loading function in a capsule in the python type, which we can then call
if the local (and possibly global, if the local type is masking a global
type) version doesn't work.
This reimplements the py::init<...> implementations using the various
functions added to support `py::init(...)`, and moves the implementing
structs into `detail/init.h` from `pybind11.h`. It doesn't simply use a
factory directly, as this is a very common case and implementation
without an extra lambda call is a small but useful optimization.
This, combined with the previous lazy initialization, also avoids
needing placement new for `py::init<...>()` construction: such
construction now occurs via an ordinary `new Type(...)`.
A consequence of this is that it also fixes a potential bug when using
multiple inheritance from Python: it was very easy to write classes
that double-initialize an existing instance which had the potential to
leak for non-pod classes. With the new implementation, an attempt to
call `__init__` on an already-initialized object is now ignored. (This
was already done in the previous commit for factory constructors).
This change exposed a few warnings (fixed here) from deleting a pointer
to a base class with virtual functions but without a virtual destructor.
These look like legitimate warnings that we shouldn't suppress; this
adds virtual destructors to the appropriate classes.
This allows you to use:
cls.def(py::init(&factory_function));
where `factory_function` returns a pointer, holder, or value of the
class type (or a derived type). Various compile-time checks
(static_asserts) are performed to ensure the function is valid, and
various run-time type checks where necessary.
Some other details of this feature:
- The `py::init` name doesn't conflict with the templated no-argument
`py::init<...>()`, but keeps the naming consistent: the existing
templated, no-argument one wraps constructors, the no-template,
function-argument one wraps factory functions.
- If returning a CppClass (whether by value or pointer) when an CppAlias
is required (i.e. python-side inheritance and a declared alias), a
dynamic_cast to the alias is attempted (for the pointer version); if
it fails, or if returned by value, an Alias(Class &&) constructor
is invoked. If this constructor doesn't exist, a runtime error occurs.
- for holder returns when an alias is required, we try a dynamic_cast of
the wrapped pointer to the alias to see if it is already an alias
instance; if it isn't, we raise an error.
- `py::init(class_factory, alias_factory)` is also available that takes
two factories: the first is called when an alias is not needed, the
second when it is.
- Reimplement factory instance clearing. The previous implementation
failed under python-side multiple inheritance: *each* inherited
type's factory init would clear the instance instead of only setting
its own type value. The new implementation here clears just the
relevant value pointer.
- dealloc is updated to explicitly set the leftover value pointer to
nullptr and the `holder_constructed` flag to false so that it can be
used to clear preallocated value without needing to rebuild the
instance internals data.
- Added various tests to test out new allocation/deallocation code.
- With preallocation now done lazily, init factory holders can
completely avoid the extra overhead of needing an extra
allocation/deallocation.
- Updated documentation to make factory constructors the default
advanced constructor style.
- If an `__init__` is called a second time, we have two choices: we can
throw away the first instance, replacing it with the second; or we can
ignore the second call. The latter is slightly easier, so do that.
An alias can be used for two main purposes: to override virtual methods,
and to add some extra data to a class needed for the pybind-wrapper.
Both of these absolutely require that the wrapped class be polymorphic
so that virtual dispatch and destruction, respectively, works.
In C++11 mode, `boost::apply_visitor` requires an explicit `result_type`.
This also adds optional tests for `boost::variant` in C++11/14, if boost
is available. In C++17 mode, `std::variant` is tested instead.
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.
Attempting to mix py::module_local and non-module_local classes results
in some unexpected/undesirable behaviour:
- if a class is registered non-local by some other module, a later
attempt to register it locally fails. It doesn't need to: it is
perfectly acceptable for the local registration to simply override
the external global registration.
- going the other way (i.e. module `A` registers a type `T` locally,
then `B` registers the same type `T` globally) causes a more serious
issue: `A.T`'s constructors no longer work because the `self` argument
gets converted to a `B.T`, which then fails to resolve.
Changing the cast precedence to prefer local over global fixes this and
makes it work more consistently, regardless of module load order.
This commit adds a `py::module_local` attribute that lets you confine a
registered type to the module (more technically, the shared object) in
which it is defined, by registering it with:
py::class_<C>(m, "C", py::module_local())
This will allow the same C++ class `C` to be registered in different
modules with independent sets of class definitions. On the Python side,
two such types will be completely distinct; on the C++ side, the C++
type resolves to a different Python type in each module.
This applies `py::module_local` automatically to `stl_bind.h` bindings
when the container value type looks like something global: i.e. when it
is a converting type (for example, when binding a `std::vector<int>`),
or when it is a registered type itself bound with `py::module_local`.
This should help resolve potential future conflicts (e.g. if two
completely unrelated modules both try to bind a `std::vector<int>`.
Users can override the automatic selection by adding a
`py::module_local()` or `py::module_local(false)`.
Note that this does mildly break backwards compatibility: bound stl
containers of basic types like `std::vector<int>` cannot be bound in one
module and returned in a different module. (This can be re-enabled with
`py::module_local(false)` as described above, but with the potential for
eventual load conflicts).
The builtin exception handler currently doesn't work across modules
under clang/libc++ for builtin pybind exceptions like
`pybind11::error_already_set` or `pybind11::stop_iteration`: under
RTLD_LOCAL module loading clang considers each module's exception
classes distinct types. This then means that the base exception
translator fails to catch the exceptions and the fall through to the
generic `std::exception` handler, which completely breaks things like
`stop_iteration`: only the `stop_iteration` of the first module loaded
actually works properly; later modules raise a RuntimeError with no
message when trying to invoke their iterators.
For example, two modules defined like this exhibit the behaviour under
clang++/libc++:
z1.cpp:
#include <pybind11/pybind11.h>
#include <pybind11/stl_bind.h>
namespace py = pybind11;
PYBIND11_MODULE(z1, m) {
py::bind_vector<std::vector<long>>(m, "IntVector");
}
z2.cpp:
#include <pybind11/pybind11.h>
#include <pybind11/stl_bind.h>
namespace py = pybind11;
PYBIND11_MODULE(z2, m) {
py::bind_vector<std::vector<double>>(m, "FloatVector");
}
Python:
import z1, z2
for i in z2.FloatVector():
pass
results in:
Traceback (most recent call last):
File "zs.py", line 2, in <module>
for i in z2.FloatVector():
RuntimeError
This commit fixes the issue by adding a new exception translator each
time the internals pointer is initialized from python builtins: this
generally means the internals data was initialized by some other
module. (The extra translator(s) are skipped under libstdc++).
This adds the infrastructure for a separate test plugin for cross-module
tests. (This commit contains no tests that actually use it, but the
following commits do; this is separated simply to provide a cleaner
commit history).
Currently types that are capable of conversion always call their convert
function when invoked with a `py::object` which is actually the correct
type. This means that code such as `py::cast<py::list>(obj)` and
`py::list l(obj.attr("list"))` make copies, which was an oversight
rather than an intentional feature.
While at first glance there might be something behind having
`py::list(obj)` make a copy (as it would in Python), this would be
inconsistent when you dig a little deeper because `py::list(l)`
*doesn't* make a copy for an existing `py::list l`, and having an
inconsistency within C++ would be worse than a C++ <-> Python
inconsistency.
It is possible to get around the copying using a
`reinterpret_borrow<list>(o)` (and this commit fixes one place, in
`embed.h`, that does so), but that seems a misuse of
`reinterpret_borrow`, which is really supposed to be just for dealing
with raw python-returned values, not `py::object`-derived wrappers which
are supposed to be higher level.
This changes the constructor of such converting types (i.e. anything
using PYBIND11_OBJECT_CVT -- `str`, `bool_`, `int_`, `float_`, `tuple`,
`dict`, `list`, `set`, `memoryview`) to reference rather than copy when
the check function passes.
It also adds an `object &&` constructor that is slightly more efficient
by avoiding an inc_ref when the check function passes.
`error_already_set` is more complicated than it needs to be, partly
because it manages reference counts itself rather than using
`py::object`, and partly because it tries to do more exception clearing
than is needed. This commit greatly simplifies it, and fixes#927.
Using `py::object` instead of `PyObject *` means we can rely on
implicit copy/move constructors.
The current logic did both a `PyErr_Clear` on deletion *and* a
`PyErr_Fetch` on creation. I can't see how the `PyErr_Clear` on
deletion is ever useful: the `Fetch` on creation itself clears the
error, so the only way doing a `PyErr_Clear` on deletion could do
anything if is some *other* exception was raised while the
`error_already_set` object was alive--but in that case, clearing some
other exception seems wrong. (Code that is worried about an exception
handler raising another exception would already catch a second
`error_already_set` from exception code).
The destructor itself called `clear()`, but `clear()` was a little bit
more paranoid that needed: it called `restore()` to restore the
currently captured error, but then immediately cleared it, using the
`PyErr_Restore` to release the references. That's unnecessary: it's
valid for us to release the references manually. This updates the code
to simply release the references on the three objects (preserving the
gil acquire).
`clear()`, however, also had the side effect of clearing the current
error, even if the current `error_already_set` didn't have a current
error (e.g. because of a previous `restore()` or `clear()` call). I
don't really see how clearing the error here can ever actually be
useful: the only way the current error could be set is if you called
`restore()` (in which case the current stored error-related members have
already been released), or if some *other* code raised the error, in
which case `clear()` on *this* object is clearing an error for which it
shouldn't be responsible.
Neither of those seem like intentional or desirable features, and
manually requesting deletion of the stored references similarly seems
pointless, so I've just made `clear()` an empty method and marked it
deprecated.
This also fixes a minor potential issue with the destruction: it is
technically possible for `value` to be null (though this seems likely to
be rare in practice); this updates the check to look at `type` which
will always be non-null for a `Fetch`ed exception.
This also adds error_already_set round-trip throw tests to the test
suite.
The instance registration for offset base types fails (under macOS, with
a segfault) in the presense of virtual base types. The issue occurs
when trying to `static_cast<Base *>(derived_ptr)` when `derived_ptr` has
been allocated (via `operator new`) but not initialized.
This commit fixes the issue by moving the addition to
`registered_instances` into `init_holder` rather than immediately after
value pointer allocation.
This also renames it to `init_instance` since it does more than holder
initialization now. (I also further renamed `init_holder_helper` to
`init_holder` since `init_holder` isn't used anymore).
Fixes#959.
This adds support for implicit conversions to bool from Python types
with `__bool__` (Python 3) or `__nonzero__` (Python 2) attributes, and
adds direct (i.e. non-converting) support for numpy bools.
If a class doesn't provide a `T::operator delete(void *)` but does have
a `T::operator delete(void *, size_t)` the latter is invoked by a
`delete someT`. Pybind currently only look for and call the former;
this commit adds detection and calling of the latter when the former
doesn't exist.
This changes the pointer `cast()` in `PYBIND11_TYPE_CASTER` to recognize
the `take_ownership` policy: if casting a pointer with take-ownership,
the `cast()` now recalls `cast()` with a dereferenced rvalue (rather
than the previous code, which was always calling it with a const lvalue
reference), and deletes the pointer after the chained `cast()` is
complete.
This makes code like:
m.def("f", []() { return new std::vector<int>(100, 1); },
py::return_value_policy::take_ownership);
do the expected thing by taking over ownership of the returned pointer
(which is deleted once the chained cast completes).
PR #936 broke the ability to return a pointer to a stl container (and,
likewise, to a tuple) because the added deduced type matched a
non-const pointer argument: the pointer-accepting `cast` in
PYBIND11_TYPE_CASTER had a `const type *`, which is a worse match for a
non-const pointer than the universal reference template #936 added.
This changes the provided TYPE_CASTER cast(ptr) to take the pointer by
template arg (so that it will accept either const or non-const pointer).
It has two other effects: it slightly reduces .so size (because many
type casters never actually need the pointer cast at all), and it allows
type casters to provide their untemplated pointer `cast()` that will
take precedence over the templated version provided in the macro.
