首页 > 代码库 > Python的运行机制--pyc文件格式浅析
Python的运行机制--pyc文件格式浅析
Python的原始代码在运行前都会被先编译成字节码,并把编译的结果保存到一个一个的PyCodeObject中,pyc 文件即是把PyCodeObject从内存中以marshal格式保存到文件后的结果。
下面我们来通过测试和工具来了解下pyc文件到底有些什么东西。
先写个简单的测试程序:
test.py
import dis myglobal = True def add(a): b = 1 a += b return a class world: def __init__(self): pass def sayHello(self): print 'hello,world' w = world() w.sayHello()
在这个例子里,全局变量,函数,类都有了,然后我们用下面的命令把它编译成pyc文件:
python -m compileall test.py
然后我们通过下面的代码分析一下test.pyc:
showfile.py
import dis, marshal, struct, sys, time, types def show_file(fname): f = open(fname, "rb") magic = f.read(4) moddate = f.read(4) modtime = time.asctime(time.localtime(struct.unpack('L', moddate)[0])) print "magic %s" % (magic.encode('hex')) print "moddate %s (%s)" % (moddate.encode('hex'), modtime) code = marshal.load(f) show_code(code) def show_code(code, indent=''): old_indent = indent print "%s<code>" % indent indent += ' ' print "%s<argcount> %d </argcount>" % (indent, code.co_argcount) print "%s<nlocals> %d</nlocals>" % (indent, code.co_nlocals) print "%s<stacksize> %d</stacksize>" % (indent, code.co_stacksize) print "%s<flags> %04x</flags>" % (indent, code.co_flags) show_hex("code", code.co_code, indent=indent) print "%s<dis>" % indent dis.disassemble(code) print "%s</dis>" % indent print "%s<names> %r</names>" % (indent, code.co_names) print "%s<varnames> %r</varnames>" % (indent, code.co_varnames) print "%s<freevars> %r</freevars>" % (indent, code.co_freevars) print "%s<cellvars> %r</cellvars>" % (indent, code.co_cellvars) print "%s<filename> %r</filename>" % (indent, code.co_filename) print "%s<name> %r</name>" % (indent, code.co_name) print "%s<firstlineno> %d</firstlineno>" % (indent, code.co_firstlineno) print "%s<consts>" % indent for const in code.co_consts: if type(const) == types.CodeType: show_code(const, indent+' ') else: print " %s%r" % (indent, const) print "%s</consts>" % indent show_hex("lnotab", code.co_lnotab, indent=indent) print "%s</code>" % old_indent def show_hex(label, h, indent): h = h.encode('hex') if len(h) < 60: print "%s<%s> %s</%s>" % (indent, label, h,label) else: print "%s<%s>" % (indent, label) for i in range(0, len(h), 60): print "%s %s" % (indent, h[i:i+60]) print "%s</%s>" % (indent, label) show_file(sys.argv[1])
运行下面的命令:
showfile.py test.pyc >test.xml
将会把输入保存到test.xml中,来看看test.xml中有什么内容:
magic 03f30d0a moddate a2567054 (Sat Nov 22 17:25:54 2014) <code> <argcount> 0 </argcount> <nlocals> 0</nlocals> <stacksize> 3</stacksize> <flags> 0040</flags> <code> 6400006401006c00005a00006501005a02006402008400005a0300640300 640500640400840000830000595a04006504008300005a05006505006a06 008300000164010053 </code> <dis> 1 0 LOAD_CONST 0 (-1) 3 LOAD_CONST 1 (None) 6 IMPORT_NAME 0 (dis) 9 STORE_NAME 0 (dis) 2 12 LOAD_NAME 1 (True) 15 STORE_NAME 2 (myglobal) 4 18 LOAD_CONST 2 (<code object add at 024E3B60, file "test.py", line 4>) 21 MAKE_FUNCTION 0 24 STORE_NAME 3 (add) 9 27 LOAD_CONST 3 ('world') 30 LOAD_CONST 5 (()) 33 LOAD_CONST 4 (<code object world at 024E3650, file "test.py", line 9>) 36 MAKE_FUNCTION 0 39 CALL_FUNCTION 0 42 BUILD_CLASS 43 STORE_NAME 4 (world) 15 46 LOAD_NAME 4 (world) 49 CALL_FUNCTION 0 52 STORE_NAME 5 (w) 16 55 LOAD_NAME 5 (w) 58 LOAD_ATTR 6 (sayHello) 61 CALL_FUNCTION 0 64 POP_TOP 65 LOAD_CONST 1 (None) 68 RETURN_VALUE </dis> <names> ('dis', 'True', 'myglobal', 'add', 'world', 'w', 'sayHello')</names> <varnames> ()</varnames> <freevars> ()</freevars> <cellvars> ()</cellvars> <filename> 'test.py'</filename> <name> '<module>'</name> <firstlineno> 1</firstlineno> <consts> -1 None <code> <argcount> 1 </argcount> <nlocals> 2</nlocals> <stacksize> 2</stacksize> <flags> 0043</flags> <code> 6401007d01007c00007c0100377d00007c000053</code> <dis> 5 0 LOAD_CONST 1 (1) 3 STORE_FAST 1 (b) 6 6 LOAD_FAST 0 (a) 9 LOAD_FAST 1 (b) 12 INPLACE_ADD 13 STORE_FAST 0 (a) 7 16 LOAD_FAST 0 (a) 19 RETURN_VALUE </dis> <names> ()</names> <varnames> ('a', 'b')</varnames> <freevars> ()</freevars> <cellvars> ()</cellvars> <filename> 'test.py'</filename> <name> 'add'</name> <firstlineno> 4</firstlineno> <consts> None 1 </consts> <lnotab> 000106010a01</lnotab> </code> 'world' <code> <argcount> 0 </argcount> <nlocals> 0</nlocals> <stacksize> 1</stacksize> <flags> 0042</flags> <code> 6500005a01006400008400005a02006401008400005a03005253</code> <dis> 9 0 LOAD_NAME 0 (__name__) 3 STORE_NAME 1 (__module__) 10 6 LOAD_CONST 0 (<code object __init__ at 024E3B18, file "test.py", line 10>) 9 MAKE_FUNCTION 0 12 STORE_NAME 2 (__init__) 12 15 LOAD_CONST 1 (<code object sayHello at 024E3698, file "test.py", line 12>) 18 MAKE_FUNCTION 0 21 STORE_NAME 3 (sayHello) 24 LOAD_LOCALS 25 RETURN_VALUE </dis> <names> ('__name__', '__module__', '__init__', 'sayHello')</names> <varnames> ()</varnames> <freevars> ()</freevars> <cellvars> ()</cellvars> <filename> 'test.py'</filename> <name> 'world'</name> <firstlineno> 9</firstlineno> <consts> <code> <argcount> 1 </argcount> <nlocals> 1</nlocals> <stacksize> 1</stacksize> <flags> 0043</flags> <code> 64000053</code> <dis> 11 0 LOAD_CONST 0 (None) 3 RETURN_VALUE </dis> <names> ()</names> <varnames> ('self',)</varnames> <freevars> ()</freevars> <cellvars> ()</cellvars> <filename> 'test.py'</filename> <name> '__init__'</name> <firstlineno> 10</firstlineno> <consts> None </consts> <lnotab> 0001</lnotab> </code> <code> <argcount> 1 </argcount> <nlocals> 1</nlocals> <stacksize> 1</stacksize> <flags> 0043</flags> <code> 640100474864000053</code> <dis> 13 0 LOAD_CONST 1 ('hello,world') 3 PRINT_ITEM 4 PRINT_NEWLINE 5 LOAD_CONST 0 (None) 8 RETURN_VALUE </dis> <names> ()</names> <varnames> ('self',)</varnames> <freevars> ()</freevars> <cellvars> ()</cellvars> <filename> 'test.py'</filename> <name> 'sayHello'</name> <firstlineno> 12</firstlineno> <consts> None 'hello,world' </consts> <lnotab> 0001</lnotab> </code> </consts> <lnotab> 06010902</lnotab> </code> () </consts> <lnotab> 0c010602090513060901</lnotab> </code>
再对照Python代码中Code.h中定义的PyCodeObject结构:
/* Bytecode object */ typedef struct { PyObject_HEAD int co_argcount; /* #arguments, except *args */ int co_nlocals; /* #local variables */ int co_stacksize; /* #entries needed for evaluation stack */ int co_flags; /* CO_..., see below */ PyObject *co_code; /* instruction opcodes */ PyObject *co_consts; /* list (constants used) */ PyObject *co_names; /* list of strings (names used) */ PyObject *co_varnames; /* tuple of strings (local variable names) */ PyObject *co_freevars; /* tuple of strings (free variable names) */ PyObject *co_cellvars; /* tuple of strings (cell variable names) */ /* The rest doesn't count for hash/cmp */ PyObject *co_filename; /* string (where it was loaded from) */ PyObject *co_name; /* string (name, for reference) */ int co_firstlineno; /* first source line number */ PyObject *co_lnotab; /* string (encoding addr<->lineno mapping) See Objects/lnotab_notes.txt for details. */ void *co_zombieframe; /* for optimization only (see frameobject.c) */ PyObject *co_weakreflist; /* to support weakrefs to code objects */ } PyCodeObject;
我们可以看到,整个test.pyc就是一个嵌套的PyCodeObject结构的组合,对于每个函数,或者类的方法,都会生成一个对应的PyCodeObject结构,并且模块还会生成额外的一个PyCodeObject结构:
来简单解释下各字段的意义:
argcount:参数的个数
nlocals:局部变量的个数(包含参数在内)
stacksize:堆栈的大小
flags:用来表示参数中是否有*args或者 **kwargs
code:字节码
names:全局变量,函数,类,类的方法的名称
varnames:局部变量的名称(包含参数)
consts:一个常量表,在marshal.c中有定义所有的类型:
#define TYPE_NULL '0' #define TYPE_NONE 'N' #define TYPE_FALSE 'F' #define TYPE_TRUE 'T' #define TYPE_STOPITER 'S' #define TYPE_ELLIPSIS '.' #define TYPE_INT 'i' #define TYPE_INT64 'I' #define TYPE_FLOAT 'f' #define TYPE_BINARY_FLOAT 'g' #define TYPE_COMPLEX 'x' #define TYPE_BINARY_COMPLEX 'y' #define TYPE_LONG 'l' #define TYPE_STRING 's' #define TYPE_INTERNED 't' #define TYPE_STRINGREF 'R' #define TYPE_TUPLE '(' #define TYPE_LIST '[' #define TYPE_DICT '{' #define TYPE_CODE 'c' #define TYPE_UNICODE 'u' #define TYPE_UNKNOWN '?' #define TYPE_SET '<' #define TYPE_FROZENSET '>'
所有的PyCodeObject都是通过调用以下的函数得以运行的:
PyObject * PyEval_EvalFrameEx(PyFrameObject *f, int throwflag)
这个函数是Python的一个重量极的函数,它的作用即是执行中间码,Python的代码都是通过调用这个函数来运行的。
通过上面的分析,对于Python代码的运行应该有一个比较清醒的认识。
参考链接:
http://nedbatchelder.com/blog/200804/the_structure_of_pyc_files.html
http://daeken.com/2010-02-20_Python_Marshal_Format.html
Python的运行机制--pyc文件格式浅析