简体繁体 English

Python 的 CFFI - 将浮点值 #define 指令传递给 cdef()

[英]Python's CFFI - Pass in floating point value #define directives to cdef()

原文 2021-12-17 05:36:09 2 1 python/ c/ python-cffi

CFFI's documentation says that it supports the use of constants and macros (like #define ) in the cdef "provided the macro is defined to be an integer value", but it doesn't explicitly say that floating point constants are not supported. CFFI 的文档说它支持在cdef中使用常量和宏（如#define ）“前提是宏被定义为 integer 值”，但它没有明确表示不支持浮点常量。 I've tried using the #define FOO... syntax with floats and #define FOO 0.5 but both fail, and have not found any more information on how I can possibly use floating point values in this manner.我尝试使用带有浮点数和#define FOO 0.5的#define FOO...语法，但都失败了，并且没有找到更多关于如何以这种方式使用浮点值的信息。 I'm trying to call functions from my C libraries from Python.我正在尝试从 Python 调用我的 C 库中的函数。

Is there a way to work around this limitation (if indeed it's a limitation) without touching the C code?有没有办法在不触及 C 代码的情况下解决这个限制（如果确实是一个限制）？

1 个解决方案

Only integers are supported.仅支持整数。 For things like floats, there are two solutions:对于浮点数之类的东西，有两种解决方案：

You can of course write the constant as pure Python code, if it has a known value, and use that from the rest of your Python code.您当然可以将常量写为纯 Python 代码（如果它具有已知值），并使用 Python 代码的 rest 中的值。 In other words, put it outside CFFI.换句话说，把它放在CFFI之外。
This works in API mode: static const float FOO;这适用于 API 模式： static const float FOO; declares FOO as something that can be read from in C, but not written to.将FOO声明为可以在 C 中读取但不能写入的内容。