1.. _type-conversions: 2 3Type conversions 4################ 5 6Apart from enabling cross-language function calls, a fundamental problem 7that a binding tool like pybind11 must address is to provide access to 8native Python types in C++ and vice versa. There are three fundamentally 9different ways to do this—which approach is preferable for a particular type 10depends on the situation at hand. 11 121. Use a native C++ type everywhere. In this case, the type must be wrapped 13 using pybind11-generated bindings so that Python can interact with it. 14 152. Use a native Python type everywhere. It will need to be wrapped so that 16 C++ functions can interact with it. 17 183. Use a native C++ type on the C++ side and a native Python type on the 19 Python side. pybind11 refers to this as a *type conversion*. 20 21 Type conversions are the most "natural" option in the sense that native 22 (non-wrapped) types are used everywhere. The main downside is that a copy 23 of the data must be made on every Python ↔ C++ transition: this is 24 needed since the C++ and Python versions of the same type generally won't 25 have the same memory layout. 26 27 pybind11 can perform many kinds of conversions automatically. An overview 28 is provided in the table ":ref:`conversion_table`". 29 30The following subsections discuss the differences between these options in more 31detail. The main focus in this section is on type conversions, which represent 32the last case of the above list. 33 34.. toctree:: 35 :maxdepth: 1 36 37 overview 38 strings 39 stl 40 functional 41 chrono 42 eigen 43 custom 44