python compare string_python源码学习（七）——String对象

系统管理员 2022-11-05 04:30 120阅读 0赞

有了前面对Python中整数对象的学习，再学习String对象就回容易理解些，首先我们来看一下Python中的PyStringObject和PySting\_Type，PyStringObject.h对象的定义如下：

typedef struct \{

PyObject\_VAR\_HEAD

long ob\_shash;

int ob\_sstate;

char ob\_sval\[1\];

/\* Invariants:

\* ob\_sval contains space for 'ob\_size+1' elements.

\* ob\_sval\[ob\_size\] == 0.

\* ob\_shash is the hash of the string or -1 if not computed yet.

\* ob\_sstate != 0 iff the string object is in stringobject.c's

\* 'interned' dictionary; in this case the two references

\* from 'interned' to this object are \*not counted\* in ob\_refcnt.

\*/

\} PyStringObject;

PyObject\_VAR\_HEAD中有个ob\_size变量，字符串的长度由ob\_size决定，ob\_size决定了这段内存的实际长度(字节)，这个机制是Python中所有变长对象的实现机制。Python中字符串都是以结尾。ob\_shash的初始值是-1，这个在Python的dict中有非常巨大的作用。hash值得算法如下：

static long

string\_hash(PyStringObject \*a)

\#ifdef Py\_DEBUG

assert(\_Py\_HashSecret\_Initialized);

\#endif

if (a->ob\_shash != -1)

return a->ob\_shash;

len = Py\_SIZE(a);

/\*

We make the hash of the empty string be 0, rather than using

(prefix ^ suffix), since this slightly obfuscates the hash secret

\*/

if (len == 0) \{

a->ob\_shash = 0;

return 0;

p = (unsigned char \*) a->ob\_sval;

x = \_Py\_HashSecret.prefix;

x ^= \*p << 7;

while (--len >= 0)

x = (1000003\*x) ^ \*p++;

x ^= Py\_SIZE(a);

x ^= \_Py\_HashSecret.suffix;

if (x == -1)

x = -2;

a->ob\_shash = x;

return x;

字符串的预存hash值和intern机制使得Python的执行效率提升了20%左右。再看一下PyString\_Type，代码如下：

PyTypeObject PyString\_Type = \{

PyVarObject\_HEAD\_INIT(&PyType\_Type, 0)

"str",

PyStringObject\_SIZE,

sizeof(char),

string\_dealloc, /\* tp\_dealloc \*/

(printfunc)string\_print, /\* tp\_print \*/

0, /\* tp\_getattr \*/

0, /\* tp\_setattr \*/

0, /\* tp\_compare \*/

string\_repr, /\* tp\_repr \*/

&string\_as\_number, /\* tp\_as\_number \*/

&string\_as\_sequence, /\* tp\_as\_sequence \*/

&string\_as\_mapping, /\* tp\_as\_mapping \*/

(hashfunc)string\_hash, /\* tp\_hash \*/

0, /\* tp\_call \*/

string\_str, /\* tp\_str \*/

PyObject\_GenericGetAttr, /\* tp\_getattro \*/

0, /\* tp\_setattro \*/

&string\_as\_buffer, /\* tp\_as\_buffer \*/

Py\_TPFLAGS\_DEFAULT | Py\_TPFLAGS\_CHECKTYPES |

Py\_TPFLAGS\_BASETYPE | Py\_TPFLAGS\_STRING\_SUBCLASS |

Py\_TPFLAGS\_HAVE\_NEWBUFFER, /\* tp\_flags \*/

string\_doc, /\* tp\_doc \*/

0, /\* tp\_traverse \*/

0, /\* tp\_clear \*/

(richcmpfunc)string\_richcompare, /\* tp\_richcompare \*/

0, /\* tp\_weaklistoffset \*/

0, /\* tp\_iter \*/

0, /\* tp\_iternext \*/

string\_methods, /\* tp\_methods \*/

0, /\* tp\_members \*/

0, /\* tp\_getset \*/

&PyBaseString\_Type, /\* tp\_base \*/

0, /\* tp\_dict \*/

0, /\* tp\_descr\_get \*/

0, /\* tp\_descr\_set \*/

0, /\* tp\_dictoffset \*/

0, /\* tp\_init \*/

0, /\* tp\_alloc \*/

string\_new, /\* tp\_new \*/

PyObject\_Del, /\* tp\_free \*/

\};

从上面的代码我们不难看出，String对象的tp\_as\_number、tp\_as\_sequence、tp\_as\_mapping三个域都被设置了，这说明了PyStringObject对数值操作

、序列操作和映射操作都支持。接下来我们看一下图片：

![p><p%20align=][p_p_20align]![b2ed57431548ac38b0e60e65305d2ae3.png][]

好了，通过以上两张图片我们就可以比较清楚的了解Python中string对象运行时的状态以及所占内存了，以后将会讲一个非常重要的机制，就是string的intern机制。

[p_p_20align]: http://www.androiddev.net/wp-content/uploads/2014/07/Screen-Shot-2014-07-05-at-%E4%B8%8B%E5%8D%883.55task%20management%20online.57.png%E2%80%9D%20alt=%E2%80%9DScreen%20Shot%202014-07-05%20at%20%E4%B8%8B%E5%8D%883.55.57%E2%80%B3%20width=%E2%80%9D656%E2%80%B3%20height=%E2%80%9D336%E2%80%B3%20class=%E2%80%9Dalignnone%20size-full%20wp-image-914%E2%80%B3%20srcset=%E2%80%9Dhttps://www.androiddev.net/wp-content/uploads/2014/07/Screen-Shot-2014-07-05-at-%E4%B8%8B%E5%8D%883.55.57.png%20656w,%20https://www.androiddev.net/wp-content/uploads/2014/07/Screen-Shot-2014-07-05-at-%E4%B8%8B%E5%8D%883.55.57-300%C3%97153.png%20300w%E2%80%9D%20sizes=%E2%80%9D%28max-width:%20656px%29%20100vw,%20656px%E2%80%9D%20/
[b2ed57431548ac38b0e60e65305d2ae3.png]: /images/20221023/5343423f663a476282b3b22f17baedc5.png