dis --- Python bytecode 的反組譯器

原始碼:Lib/dis.py


dis 模組支援反組譯分析 CPythonbytecode。CPython bytecode 作為輸入的模組被定義於Include/opcode.h 並且被編譯器和直譯器所使用。

CPython 實作細節: Bytecode is an implementation detail of the CPython interpreter. Noguarantees are made that bytecode will not be added, removed, or changedbetween versions of Python. Use of this module should not be considered towork across Python VMs or Python releases.

在 3.6 版的變更:Use 2 bytes for each instruction. Previously the number of bytes variedby instruction.

在 3.10 版的變更:The argument of jump, exception handling and loop instructions is nowthe instruction offset rather than the byte offset.

在 3.11 版的變更:Some instructions are accompanied by one or more inline cache entries,which take the form ofCACHE instructions. These instructionsare hidden by default, but can be shown by passingshow_caches=True toanydis utility. Furthermore, the interpreter now adapts thebytecode to specialize it for different runtime conditions. Theadaptive bytecode can be shown by passingadaptive=True.

在 3.12 版的變更:The argument of a jump is the offset of the target instruction relativeto the instruction that appears immediately after the jump instruction'sCACHE entries.

As a consequence, the presence of theCACHE instructions istransparent for forward jumps but needs to be taken into account whenreasoning about backward jumps.

在 3.13 版的變更:The output shows logical labels rather than instruction offsetsfor jump targets and exception handlers. The-O command lineoption and theshow_offsets argument were added.

Example: Given the functionmyfunc():

defmyfunc(alist):returnlen(alist)

the following command can be used to display the disassembly ofmyfunc():

>>>dis.dis(myfunc)  2           RESUME                   0  3           LOAD_GLOBAL              1 (len + NULL)              LOAD_FAST                0 (alist)              CALL                     1              RETURN_VALUE

(The "2" is a line number).

命令列介面

Thedis module can be invoked as a script from the command line:

python-mdis[-h][-C][-O][infile]

可接受以下選項:

-h,--help

Display usage and exit.

-C,--show-caches

Show inline caches.

在 3.13 版被加入.

-O,--show-offsets

Show offsets of instructions.

在 3.13 版被加入.

Ifinfile is specified, its disassembled code will be written to stdout.Otherwise, disassembly is performed on compiled source code received from stdin.

Bytecode analysis

在 3.4 版被加入.

The bytecode analysis API allows pieces of Python code to be wrapped in aBytecode object that provides easy access to details of the compiledcode.

classdis.Bytecode(x,*,first_line=None,current_offset=None,show_caches=False,adaptive=False,show_offsets=False)

Analyse the bytecode corresponding to a function, generator, asynchronousgenerator, coroutine, method, string of source code, or a code object (asreturned bycompile()).

This is a convenience wrapper around many of the functions listed below, mostnotablyget_instructions(), as iterating over aBytecodeinstance yields the bytecode operations asInstruction instances.

Iffirst_line is notNone, it indicates the line number that should bereported for the first source line in the disassembled code. Otherwise, thesource line information (if any) is taken directly from the disassembled codeobject.

Ifcurrent_offset is notNone, it refers to an instruction offset in thedisassembled code. Setting this meansdis() will display a "currentinstruction" marker against the specified opcode.

Ifshow_caches isTrue,dis() will display inline cacheentries used by the interpreter to specialize the bytecode.

Ifadaptive isTrue,dis() will display specialized bytecodethat may be different from the original bytecode.

Ifshow_offsets isTrue,dis() will include instructionoffsets in the output.

classmethodfrom_traceback(tb,*,show_caches=False)

Construct aBytecode instance from the given traceback, settingcurrent_offset to the instruction responsible for the exception.

codeobj

The compiled code object.

first_line

The first source line of the code object (if available)

dis()

Return a formatted view of the bytecode operations (the same as printed bydis.dis(), but returned as a multi-line string).

info()

Return a formatted multi-line string with detailed information about thecode object, likecode_info().

在 3.7 版的變更:This can now handle coroutine and asynchronous generator objects.

在 3.11 版的變更:新增show_cachesadaptive 參數。

範例:

>>>bytecode=dis.Bytecode(myfunc)>>>forinstrinbytecode:...print(instr.opname)...RESUMELOAD_GLOBALLOAD_FASTCALLRETURN_VALUE

Analysis functions

Thedis module also defines the following analysis functions that convertthe input directly to the desired output. They can be useful if only a singleoperation is being performed, so the intermediate analysis object isn't useful:

dis.code_info(x)

Return a formatted multi-line string with detailed code object informationfor the supplied function, generator, asynchronous generator, coroutine,method, source code string or code object.

Note that the exact contents of code info strings are highly implementationdependent and they may change arbitrarily across Python VMs or Pythonreleases.

在 3.2 版被加入.

在 3.7 版的變更:This can now handle coroutine and asynchronous generator objects.

dis.show_code(x,*,file=None)

Print detailed code object information for the supplied function, method,source code string or code object tofile (orsys.stdout iffileis not specified).

This is a convenient shorthand forprint(code_info(x),file=file),intended for interactive exploration at the interpreter prompt.

在 3.2 版被加入.

在 3.4 版的變更:新增file 參數。

dis.dis(x=None,*,file=None,depth=None,show_caches=False,adaptive=False)

Disassemble thex object.x can denote either a module, a class, amethod, a function, a generator, an asynchronous generator, a coroutine,a code object, a string of source code or a byte sequence of raw bytecode.For a module, it disassembles all functions. For a class, it disassemblesall methods (including class and static methods). For a code object orsequence of raw bytecode, it prints one line per bytecode instruction.It also recursively disassembles nested code objects. These can includegenerator expressions, nested functions, the bodies of nested classes,and the code objects used forannotation scopes.Strings are first compiled to code objects with thecompile()built-in function before being disassembled. If no object is provided, thisfunction disassembles the last traceback.

The disassembly is written as text to the suppliedfile argument ifprovided and tosys.stdout otherwise.

The maximal depth of recursion is limited bydepth unless it isNone.depth=0 means no recursion.

Ifshow_caches isTrue, this function will display inline cacheentries used by the interpreter to specialize the bytecode.

Ifadaptive isTrue, this function will display specialized bytecodethat may be different from the original bytecode.

在 3.4 版的變更:新增file 參數。

在 3.7 版的變更:Implemented recursive disassembling and addeddepth parameter.

在 3.7 版的變更:This can now handle coroutine and asynchronous generator objects.

在 3.11 版的變更:新增show_cachesadaptive 參數。

distb(tb=None,*,file=None,show_caches=False,adaptive=False,
show_offset=False)

Disassemble the top-of-stack function of a traceback, using the lasttraceback if none was passed. The instruction causing the exception isindicated.

The disassembly is written as text to the suppliedfile argument ifprovided and tosys.stdout otherwise.

在 3.4 版的變更:新增file 參數。

在 3.11 版的變更:新增show_cachesadaptive 參數。

在 3.13 版的變更:新增show_offsets 參數。

dis.disassemble(code,lasti=-1,*,file=None,show_caches=False,adaptive=False)
disco(code,lasti=-1,*,file=None,show_caches=False,adaptive=False,
show_offsets=False)

Disassemble a code object, indicating the last instruction iflasti wasprovided. The output is divided in the following columns:

  1. the line number, for the first instruction of each line

  2. the current instruction, indicated as-->,

  3. a labelled instruction, indicated with>>,

  4. the address of the instruction,

  5. the operation code name,

  6. operation parameters, and

  7. interpretation of the parameters in parentheses.

The parameter interpretation recognizes local and global variable names,constant values, branch targets, and compare operators.

The disassembly is written as text to the suppliedfile argument ifprovided and tosys.stdout otherwise.

在 3.4 版的變更:新增file 參數。

在 3.11 版的變更:新增show_cachesadaptive 參數。

在 3.13 版的變更:新增show_offsets 參數。

dis.get_instructions(x,*,first_line=None,show_caches=False,adaptive=False)

Return an iterator over the instructions in the supplied function, method,source code string or code object.

The iterator generates a series ofInstruction named tuples givingthe details of each operation in the supplied code.

Iffirst_line is notNone, it indicates the line number that should bereported for the first source line in the disassembled code. Otherwise, thesource line information (if any) is taken directly from the disassembled codeobject.

Theadaptive parameter works as it does indis().

在 3.4 版被加入.

在 3.11 版的變更:新增show_cachesadaptive 參數。

在 3.13 版的變更:Theshow_caches parameter is deprecated and has no effect. The iteratorgenerates theInstruction instances with thecache_infofield populated (regardless of the value ofshow_caches) and it no longergenerates separate items for the cache entries.

dis.findlinestarts(code)

This generator function uses theco_lines() methodof thecode objectcode to find the offsets whichare starts oflines in the source code. They are generated as(offset,lineno) pairs.

在 3.6 版的變更:Line numbers can be decreasing. Before, they were always increasing.

在 3.10 版的變更:ThePEP 626co_lines() method is used instead of theco_firstlineno andco_lnotabattributes of thecode object.

在 3.13 版的變更:Line numbers can beNone for bytecode that does not map to source lines.

dis.findlabels(code)

Detect all offsets in the raw compiled bytecode stringcode which are jump targets, andreturn a list of these offsets.

dis.stack_effect(opcode,oparg=None,*,jump=None)

Compute the stack effect ofopcode with argumentoparg.

If the code has a jump target andjump isTrue,stack_effect()will return the stack effect of jumping. Ifjump isFalse,it will return the stack effect of not jumping. And ifjump isNone (default), it will return the maximal stack effect of both cases.

在 3.4 版被加入.

在 3.8 版的變更:新增jump 參數。

在 3.13 版的變更:Ifoparg is omitted (orNone), the stack effect is now returnedforoparg=0. Previously this was an error for opcodes that use theirarg. It is also no longer an error to pass an integeroparg whentheopcode does not use it; theoparg in this case is ignored.

Python Bytecode Instructions

Theget_instructions() function andBytecode class providedetails of bytecode instructions asInstruction instances:

classdis.Instruction

Details for a bytecode operation

opcode

numeric code for operation, corresponding to the opcode values listedbelow and the bytecode values in theOpcode collections.

opname

human readable name for operation

baseopcode

numeric code for the base operation if operation is specialized;otherwise equal toopcode

baseopname

human readable name for the base operation if operation is specialized;otherwise equal toopname

arg

numeric argument to operation (if any), otherwiseNone

oparg

alias forarg

argval

resolved arg value (if any), otherwiseNone

argrepr

human readable description of operation argument (if any),otherwise an empty string.

offset

start index of operation within bytecode sequence

start_offset

start index of operation within bytecode sequence, including prefixedEXTENDED_ARG operations if present; otherwise equal tooffset

cache_offset

start index of the cache entries following the operation

end_offset

end index of the cache entries following the operation

starts_line

True if this opcode starts a source line, otherwiseFalse

line_number

source line number associated with this opcode (if any), otherwiseNone

is_jump_target

True if other code jumps to here, otherwiseFalse

jump_target

bytecode index of the jump target if this is a jump operation,otherwiseNone

positions

dis.Positions object holding thestart and end locations that are covered by this instruction.

cache_info

Information about the cache entries of this instruction, astriplets of the form(name,size,data), where thenameandsize describe the cache format and data is the contentsof the cache.cache_info isNone if the instruction does not havecaches.

在 3.4 版被加入.

在 3.11 版的變更:Fieldpositions is added.

在 3.13 版的變更:Changed fieldstarts_line.

Added fieldsstart_offset,cache_offset,end_offset,baseopname,baseopcode,jump_target,oparg,line_number andcache_info.

classdis.Positions

In case the information is not available, some fields might beNone.

lineno
end_lineno
col_offset
end_col_offset

在 3.11 版被加入.

The Python compiler currently generates the following bytecode instructions.

General instructions

In the following, We will refer to the interpreter stack asSTACK and describeoperations on it as if it was a Python list. The top of the stack corresponds toSTACK[-1] in this language.

NOP

Do nothing code. Used as a placeholder by the bytecode optimizer, and togenerate line tracing events.

POP_TOP

Removes the top-of-stack item:

STACK.pop()
END_FOR

Removes the top-of-stack item.Equivalent toPOP_TOP.Used to clean up at the end of loops, hence the name.

在 3.12 版被加入.

END_SEND

ImplementsdelSTACK[-2].Used to clean up when a generator exits.

在 3.12 版被加入.

COPY(i)

Push the i-th item to the top of the stack without removing it from its originallocation:

asserti>0STACK.append(STACK[-i])

在 3.11 版被加入.

SWAP(i)

Swap the top of the stack with the i-th element:

STACK[-i],STACK[-1]=STACK[-1],STACK[-i]

在 3.11 版被加入.

CACHE

Rather than being an actual instruction, this opcode is used to mark extraspace for the interpreter to cache useful data directly in the bytecodeitself. It is automatically hidden by alldis utilities, but can beviewed withshow_caches=True.

Logically, this space is part of the preceding instruction. Many opcodesexpect to be followed by an exact number of caches, and will instruct theinterpreter to skip over them at runtime.

Populated caches can look like arbitrary instructions, so great care shouldbe taken when reading or modifying raw, adaptive bytecode containingquickened data.

在 3.11 版被加入.

Unary operations

Unary operations take the top of the stack, apply the operation, and push theresult back on the stack.

UNARY_NEGATIVE

ImplementsSTACK[-1]=-STACK[-1].

UNARY_NOT

ImplementsSTACK[-1]=notSTACK[-1].

在 3.13 版的變更:This instruction now requires an exactbool operand.

UNARY_INVERT

ImplementsSTACK[-1]=~STACK[-1].

GET_ITER

ImplementsSTACK[-1]=iter(STACK[-1]).

GET_YIELD_FROM_ITER

IfSTACK[-1] is agenerator iterator orcoroutine objectit is left as is. Otherwise, implementsSTACK[-1]=iter(STACK[-1]).

在 3.5 版被加入.

TO_BOOL

ImplementsSTACK[-1]=bool(STACK[-1]).

在 3.13 版被加入.

Binary and in-place operations

Binary operations remove the top two items from the stack (STACK[-1] andSTACK[-2]). They perform the operation, then put the result back on the stack.

In-place operations are like binary operations, but the operation is done in-placewhenSTACK[-2] supports it, and the resultingSTACK[-1] may be (but doesnot have to be) the originalSTACK[-2].

BINARY_OP(op)

Implements the binary and in-place operators (depending on the value ofop):

rhs=STACK.pop()lhs=STACK.pop()STACK.append(lhsoprhs)

在 3.11 版被加入.

BINARY_SUBSCR

Implements:

key=STACK.pop()container=STACK.pop()STACK.append(container[key])
STORE_SUBSCR

Implements:

key=STACK.pop()container=STACK.pop()value=STACK.pop()container[key]=value
DELETE_SUBSCR

Implements:

key=STACK.pop()container=STACK.pop()delcontainer[key]
BINARY_SLICE

Implements:

end=STACK.pop()start=STACK.pop()container=STACK.pop()STACK.append(container[start:end])

在 3.12 版被加入.

STORE_SLICE

Implements:

end=STACK.pop()start=STACK.pop()container=STACK.pop()values=STACK.pop()container[start:end]=value

在 3.12 版被加入.

Coroutine opcodes

GET_AWAITABLE(where)

ImplementsSTACK[-1]=get_awaitable(STACK[-1]), whereget_awaitable(o)returnso ifo is a coroutine object or a generator object withtheCO_ITERABLE_COROUTINE flag, or resolveso.__await__.

If thewhere operand is nonzero, it indicates where the instructionoccurs:

  • 1: After a call to__aenter__

  • 2: After a call to__aexit__

在 3.5 版被加入.

在 3.11 版的變更:Previously, this instruction did not have an oparg.

GET_AITER

ImplementsSTACK[-1]=STACK[-1].__aiter__().

在 3.5 版被加入.

在 3.7 版的變更:Returning awaitable objects from__aiter__ is no longersupported.

GET_ANEXT

ImplementSTACK.append(get_awaitable(STACK[-1].__anext__())) to the stack.SeeGET_AWAITABLE for details aboutget_awaitable.

在 3.5 版被加入.

END_ASYNC_FOR

Terminates anasyncfor loop. Handles an exception raisedwhen awaiting a next item. The stack contains the async iterable inSTACK[-2] and the raised exception inSTACK[-1]. Both are popped.If the exception is notStopAsyncIteration, it is re-raised.

在 3.8 版被加入.

在 3.11 版的變更:Exception representation on the stack now consist of one, not three, items.

CLEANUP_THROW

Handles an exception raised during athrow() orclose() call through the current frame. IfSTACK[-1] is aninstance ofStopIteration, pop three values from the stack and pushitsvalue member. Otherwise, re-raiseSTACK[-1].

在 3.12 版被加入.

BEFORE_ASYNC_WITH

Resolves__aenter__ and__aexit__ fromSTACK[-1].Pushes__aexit__ and result of__aenter__() to the stack:

STACK.extend((__aexit__,__aenter__())

在 3.5 版被加入.

Miscellaneous opcodes

SET_ADD(i)

Implements:

item=STACK.pop()set.add(STACK[-i],item)

Used to implement set comprehensions.

LIST_APPEND(i)

Implements:

item=STACK.pop()list.append(STACK[-i],item)

Used to implement list comprehensions.

MAP_ADD(i)

Implements:

value=STACK.pop()key=STACK.pop()dict.__setitem__(STACK[-i],key,value)

Used to implement dict comprehensions.

在 3.1 版被加入.

在 3.8 版的變更:Map value isSTACK[-1] and map key isSTACK[-2]. Before, thosewere reversed.

For all of theSET_ADD,LIST_APPEND andMAP_ADDinstructions, while the added value or key/value pair is popped off, thecontainer object remains on the stack so that it is available for furtheriterations of the loop.

RETURN_VALUE

Returns withSTACK[-1] to the caller of the function.

RETURN_CONST(consti)

Returns withco_consts[consti] to the caller of the function.

在 3.12 版被加入.

YIELD_VALUE

YieldsSTACK.pop() from agenerator.

在 3.11 版的變更:oparg set to be the stack depth.

在 3.12 版的變更:oparg set to be the exception block depth, for efficient closing of generators.

在 3.13 版的變更:oparg is1 if this instruction is part of a yield-from or await, and0otherwise.

SETUP_ANNOTATIONS

Checks whether__annotations__ is defined inlocals(), if not it isset up to an emptydict. This opcode is only emitted if a classor module body containsvariable annotationsstatically.

在 3.6 版被加入.

POP_EXCEPT

Pops a value from the stack, which is used to restore the exception state.

在 3.11 版的變更:Exception representation on the stack now consist of one, not three, items.

RERAISE

Re-raises the exception currently on top of the stack. If oparg is non-zero,pops an additional value from the stack which is used to setf_lasti of the current frame.

在 3.9 版被加入.

在 3.11 版的變更:Exception representation on the stack now consist of one, not three, items.

PUSH_EXC_INFO

Pops a value from the stack. Pushes the current exception to the top of the stack.Pushes the value originally popped back to the stack.Used in exception handlers.

在 3.11 版被加入.

CHECK_EXC_MATCH

Performs exception matching forexcept. Tests whether theSTACK[-2]is an exception matchingSTACK[-1]. PopsSTACK[-1] and pushes the booleanresult of the test.

在 3.11 版被加入.

CHECK_EG_MATCH

Performs exception matching forexcept*. Appliessplit(STACK[-1]) onthe exception group representingSTACK[-2].

In case of a match, pops two items from the stack and pushes thenon-matching subgroup (None in case of full match) followed by thematching subgroup. When there is no match, pops one item (the matchtype) and pushesNone.

在 3.11 版被加入.

WITH_EXCEPT_START

Calls the function in position 4 on the stack with arguments (type, val, tb)representing the exception at the top of the stack.Used to implement the callcontext_manager.__exit__(*exc_info()) when an exceptionhas occurred in awith statement.

在 3.9 版被加入.

在 3.11 版的變更:The__exit__ function is in position 4 of the stack rather than 7.Exception representation on the stack now consist of one, not three, items.

LOAD_ASSERTION_ERROR

PushesAssertionError onto the stack. Used by theassertstatement.

在 3.9 版被加入.

LOAD_BUILD_CLASS

Pushesbuiltins.__build_class__() onto the stack. It is later calledto construct a class.

BEFORE_WITH

This opcode performs several operations before a with block starts. First,it loads__exit__() from the context manager and pushes it ontothe stack for later use byWITH_EXCEPT_START. Then,__enter__() is called. Finally, the result of calling the__enter__() method is pushed onto the stack.

在 3.11 版被加入.

GET_LEN

PerformSTACK.append(len(STACK[-1])). Used inmatch statements wherecomparison with structure of pattern is needed.

在 3.10 版被加入.

MATCH_MAPPING

IfSTACK[-1] is an instance ofcollections.abc.Mapping (or, moretechnically: if it has thePy_TPFLAGS_MAPPING flag set in itstp_flags), pushTrue onto the stack. Otherwise,pushFalse.

在 3.10 版被加入.

MATCH_SEQUENCE

IfSTACK[-1] is an instance ofcollections.abc.Sequence and isnot an instanceofstr/bytes/bytearray (or, more technically: if it hasthePy_TPFLAGS_SEQUENCE flag set in itstp_flags),pushTrue onto the stack. Otherwise, pushFalse.

在 3.10 版被加入.

MATCH_KEYS

STACK[-1] is a tuple of mapping keys, andSTACK[-2] is the match subject.IfSTACK[-2] contains all of the keys inSTACK[-1], push atuplecontaining the corresponding values. Otherwise, pushNone.

在 3.10 版被加入.

在 3.11 版的變更:Previously, this instruction also pushed a boolean value indicatingsuccess (True) or failure (False).

STORE_NAME(namei)

Implementsname=STACK.pop().namei is the index ofname in the attributeco_names of thecode object.The compiler tries to useSTORE_FAST orSTORE_GLOBAL if possible.

DELETE_NAME(namei)

Implementsdelname, wherenamei is the index intoco_namesattribute of thecode object.

UNPACK_SEQUENCE(count)

UnpacksSTACK[-1] intocount individual values, which are put onto the stackright-to-left. Require there to be exactlycount values.:

assert(len(STACK[-1])==count)STACK.extend(STACK.pop()[:-count-1:-1])
UNPACK_EX(counts)

Implements assignment with a starred target: Unpacks an iterable inSTACK[-1]into individual values, where the total number of values can be smaller than thenumber of items in the iterable: one of the new values will be a list of allleftover items.

The number of values before and after the list value is limited to 255.

The number of values before the list value is encoded in the argument of theopcode. The number of values after the list if any is encoded using anEXTENDED_ARG. As a consequence, the argument can be seen as a two bytes valueswhere the low byte ofcounts is the number of values before the list value, thehigh byte ofcounts the number of values after it.

The extracted values are put onto the stack right-to-left, i.e.a,*b,c=dwill be stored after execution asSTACK.extend((a,b,c)).

STORE_ATTR(namei)

Implements:

obj=STACK.pop()value=STACK.pop()obj.name=value

wherenamei is the index of name inco_names of thecode object.

DELETE_ATTR(namei)

Implements:

obj=STACK.pop()delobj.name

wherenamei is the index of name intoco_names of thecode object.

STORE_GLOBAL(namei)

Works asSTORE_NAME, but stores the name as a global.

DELETE_GLOBAL(namei)

Works asDELETE_NAME, but deletes a global name.

LOAD_CONST(consti)

Pushesco_consts[consti] onto the stack.

LOAD_NAME(namei)

Pushes the value associated withco_names[namei] onto the stack.The name is looked up within the locals, then the globals, then the builtins.

LOAD_LOCALS

Pushes a reference to the locals dictionary onto the stack. This is usedto prepare namespace dictionaries forLOAD_FROM_DICT_OR_DEREFandLOAD_FROM_DICT_OR_GLOBALS.

在 3.12 版被加入.

LOAD_FROM_DICT_OR_GLOBALS(i)

Pops a mapping off the stack and looks up the value forco_names[namei].If the name is not found there, looks it up in the globals and then the builtins,similar toLOAD_GLOBAL.This is used for loading global variables inannotation scopes within class bodies.

在 3.12 版被加入.

BUILD_TUPLE(count)

Creates a tuple consumingcount items from the stack, and pushes theresulting tuple onto the stack:

ifcount==0:value=()else:value=tuple(STACK[-count:])STACK=STACK[:-count]STACK.append(value)
BUILD_LIST(count)

Works asBUILD_TUPLE, but creates a list.

BUILD_SET(count)

Works asBUILD_TUPLE, but creates a set.

BUILD_MAP(count)

Pushes a new dictionary object onto the stack. Pops2*count itemsso that the dictionary holdscount entries:{...,STACK[-4]:STACK[-3],STACK[-2]:STACK[-1]}.

在 3.5 版的變更:The dictionary is created from stack items instead of creating anempty dictionary pre-sized to holdcount items.

BUILD_CONST_KEY_MAP(count)

The version ofBUILD_MAP specialized for constant keys. Pops thetop element on the stack which contains a tuple of keys, then starting fromSTACK[-2], popscount values to form values in the built dictionary.

在 3.6 版被加入.

BUILD_STRING(count)

Concatenatescount strings from the stack and pushes the resulting stringonto the stack.

在 3.6 版被加入.

LIST_EXTEND(i)

Implements:

seq=STACK.pop()list.extend(STACK[-i],seq)

Used to build lists.

在 3.9 版被加入.

SET_UPDATE(i)

Implements:

seq=STACK.pop()set.update(STACK[-i],seq)

Used to build sets.

在 3.9 版被加入.

DICT_UPDATE(i)

Implements:

map=STACK.pop()dict.update(STACK[-i],map)

Used to build dicts.

在 3.9 版被加入.

DICT_MERGE(i)

LikeDICT_UPDATE but raises an exception for duplicate keys.

在 3.9 版被加入.

LOAD_ATTR(namei)

If the low bit ofnamei is not set, this replacesSTACK[-1] withgetattr(STACK[-1],co_names[namei>>1]).

If the low bit ofnamei is set, this will attempt to load a method namedco_names[namei>>1] from theSTACK[-1] object.STACK[-1] is popped.This bytecode distinguishes two cases: ifSTACK[-1] has a method with thecorrect name, the bytecode pushes the unbound method andSTACK[-1].STACK[-1] will be used as the first argument (self) byCALLorCALL_KW when calling the unbound method.Otherwise,NULL and the object returned bythe attribute lookup are pushed.

在 3.12 版的變更:If the low bit ofnamei is set, then aNULL orself ispushed to the stack before the attribute or unbound method respectively.

LOAD_SUPER_ATTR(namei)

This opcode implementssuper(), both in its zero-argument andtwo-argument forms (e.g.super().method(),super().attr andsuper(cls,self).method(),super(cls,self).attr).

It pops three values from the stack (from top of stack down):

  • self: the first argument to the current method

  • cls: the class within which the current method was defined

  • the globalsuper

With respect to its argument, it works similarly toLOAD_ATTR,except thatnamei is shifted left by 2 bits instead of 1.

The low bit ofnamei signals to attempt a method load, as withLOAD_ATTR, which results in pushingNULL and the loaded method.When it is unset a single value is pushed to the stack.

The second-low bit ofnamei, if set, means that this was a two-argumentcall tosuper() (unset means zero-argument).

在 3.12 版被加入.

COMPARE_OP(opname)

Performs a Boolean operation. The operation name can be found incmp_op[opname>>5]. If the fifth-lowest bit ofopname is set(opname&16), the result should be coerced tobool.

在 3.13 版的變更:The fifth-lowest bit of the oparg now indicates a forced conversion tobool.

IS_OP(invert)

Performsis comparison, orisnot ifinvert is 1.

在 3.9 版被加入.

CONTAINS_OP(invert)

Performsin comparison, ornotin ifinvert is 1.

在 3.9 版被加入.

IMPORT_NAME(namei)

Imports the moduleco_names[namei].STACK[-1] andSTACK[-2] arepopped and provide thefromlist andlevel arguments of__import__().The module object is pushed onto the stack. The current namespace is not affected: for a proper import statement, a subsequentSTORE_FAST instructionmodifies the namespace.

IMPORT_FROM(namei)

Loads the attributeco_names[namei] from the module found inSTACK[-1].The resulting object is pushed onto the stack, to be subsequently stored by aSTORE_FAST instruction.

JUMP_FORWARD(delta)

Increments bytecode counter bydelta.

JUMP_BACKWARD(delta)

Decrements bytecode counter bydelta. Checks for interrupts.

在 3.11 版被加入.

JUMP_BACKWARD_NO_INTERRUPT(delta)

Decrements bytecode counter bydelta. Does not check for interrupts.

在 3.11 版被加入.

POP_JUMP_IF_TRUE(delta)

IfSTACK[-1] is true, increments the bytecode counter bydelta.STACK[-1] is popped.

在 3.11 版的變更:The oparg is now a relative delta rather than an absolute target.This opcode is a pseudo-instruction, replaced in final bytecode bythe directed versions (forward/backward).

在 3.12 版的變更:This is no longer a pseudo-instruction.

在 3.13 版的變更:This instruction now requires an exactbool operand.

POP_JUMP_IF_FALSE(delta)

IfSTACK[-1] is false, increments the bytecode counter bydelta.STACK[-1] is popped.

在 3.11 版的變更:The oparg is now a relative delta rather than an absolute target.This opcode is a pseudo-instruction, replaced in final bytecode bythe directed versions (forward/backward).

在 3.12 版的變更:This is no longer a pseudo-instruction.

在 3.13 版的變更:This instruction now requires an exactbool operand.

POP_JUMP_IF_NOT_NONE(delta)

IfSTACK[-1] is notNone, increments the bytecode counter bydelta.STACK[-1] is popped.

在 3.11 版被加入.

在 3.12 版的變更:This is no longer a pseudo-instruction.

POP_JUMP_IF_NONE(delta)

IfSTACK[-1] isNone, increments the bytecode counter bydelta.STACK[-1] is popped.

在 3.11 版被加入.

在 3.12 版的變更:This is no longer a pseudo-instruction.

FOR_ITER(delta)

STACK[-1] is aniterator. Call its__next__() method.If this yields a new value, push it on the stack (leaving the iterator belowit). If the iterator indicates it is exhausted then the byte code counter isincremented bydelta.

在 3.12 版的變更:Up until 3.11 the iterator was popped when it was exhausted.

LOAD_GLOBAL(namei)

Loads the global namedco_names[namei>>1] onto the stack.

在 3.11 版的變更:If the low bit ofnamei is set, then aNULL is pushed to thestack before the global variable.

LOAD_FAST(var_num)

Pushes a reference to the localco_varnames[var_num] onto the stack.

在 3.12 版的變更:This opcode is now only used in situations where the local variable isguaranteed to be initialized. It cannot raiseUnboundLocalError.

LOAD_FAST_LOAD_FAST(var_nums)

Pushes references toco_varnames[var_nums>>4] andco_varnames[var_nums&15] onto the stack.

在 3.13 版被加入.

LOAD_FAST_CHECK(var_num)

Pushes a reference to the localco_varnames[var_num] onto the stack,raising anUnboundLocalError if the local variable has not beeninitialized.

在 3.12 版被加入.

LOAD_FAST_AND_CLEAR(var_num)

Pushes a reference to the localco_varnames[var_num] onto the stack (orpushesNULL onto the stack if the local variable has not beeninitialized) and setsco_varnames[var_num] toNULL.

在 3.12 版被加入.

STORE_FAST(var_num)

StoresSTACK.pop() into the localco_varnames[var_num].

STORE_FAST_STORE_FAST(var_nums)

StoresSTACK[-1] intoco_varnames[var_nums>>4]andSTACK[-2] intoco_varnames[var_nums&15].

在 3.13 版被加入.

STORE_FAST_LOAD_FAST(var_nums)

StoresSTACK.pop() into the localco_varnames[var_nums>>4]and pushes a reference to the localco_varnames[var_nums&15]onto the stack.

在 3.13 版被加入.

DELETE_FAST(var_num)

Deletes localco_varnames[var_num].

MAKE_CELL(i)

Creates a new cell in sloti. If that slot is nonempty thenthat value is stored into the new cell.

在 3.11 版被加入.

LOAD_DEREF(i)

Loads the cell contained in sloti of the "fast locals" storage.Pushes a reference to the object the cell contains on the stack.

在 3.11 版的變更:i is no longer offset by the length ofco_varnames.

LOAD_FROM_DICT_OR_DEREF(i)

Pops a mapping off the stack and looks up the name associated withsloti of the "fast locals" storage in this mapping.If the name is not found there, loads it from the cell contained insloti, similar toLOAD_DEREF. This is used for loadingclosure variables in class bodies (which previously usedLOAD_CLASSDEREF) and inannotation scopes within class bodies.

在 3.12 版被加入.

STORE_DEREF(i)

StoresSTACK.pop() into the cell contained in sloti of the "fast locals"storage.

在 3.11 版的變更:i is no longer offset by the length ofco_varnames.

DELETE_DEREF(i)

Empties the cell contained in sloti of the "fast locals" storage.Used by thedel statement.

在 3.2 版被加入.

在 3.11 版的變更:i is no longer offset by the length ofco_varnames.

COPY_FREE_VARS(n)

Copies thenfree (closure) variables from the closureinto the frame. Removes the need for special code on the caller's side when callingclosures.

在 3.11 版被加入.

RAISE_VARARGS(argc)

Raises an exception using one of the 3 forms of theraise statement,depending on the value ofargc:

  • 0:raise (re-raise previous exception)

  • 1:raiseSTACK[-1] (raise exception instance or type atSTACK[-1])

  • 2:raiseSTACK[-2]fromSTACK[-1] (raise exception instance or type atSTACK[-2] with__cause__ set toSTACK[-1])

CALL(argc)

Calls a callable object with the number of arguments specified byargc.On the stack are (in ascending order):

  • The callable

  • self orNULL

  • The remaining positional arguments

argc is the total of the positional arguments, excludingself.

CALL pops all arguments and the callable object off the stack,calls the callable object with those arguments, and pushes the return valuereturned by the callable object.

在 3.11 版被加入.

在 3.13 版的變更:The callable now always appears at the same position on the stack.

在 3.13 版的變更:Calls with keyword arguments are now handled byCALL_KW.

CALL_KW(argc)

Calls a callable object with the number of arguments specified byargc,including one or more named arguments. On the stack are (in ascending order):

  • The callable

  • self orNULL

  • The remaining positional arguments

  • The named arguments

  • Atuple of keyword argument names

argc is the total of the positional and named arguments, excludingself.The length of the tuple of keyword argument names is the number of named arguments.

CALL_KW pops all arguments, the keyword names, and the callable objectoff the stack, calls the callable object with those arguments, and pushes thereturn value returned by the callable object.

在 3.13 版被加入.

CALL_FUNCTION_EX(flags)

Calls a callable object with variable set of positional and keywordarguments. If the lowest bit offlags is set, the top of the stackcontains a mapping object containing additional keyword arguments.Before the callable is called, the mapping object and iterable objectare each "unpacked" and their contents passed in as keyword andpositional arguments respectively.CALL_FUNCTION_EX pops all arguments and the callable object off the stack,calls the callable object with those arguments, and pushes the return valuereturned by the callable object.

在 3.6 版被加入.

PUSH_NULL

Pushes aNULL to the stack.Used in the call sequence to match theNULL pushed byLOAD_METHOD for non-method calls.

在 3.11 版被加入.

MAKE_FUNCTION

Pushes a new function object on the stack built from the code object atSTACK[-1].

在 3.10 版的變更:Flag value0x04 is a tuple of strings instead of dictionary

在 3.11 版的變更:Qualified name atSTACK[-1] was removed.

在 3.13 版的變更:Extra function attributes on the stack, signaled by oparg flags, wereremoved. They now useSET_FUNCTION_ATTRIBUTE.

SET_FUNCTION_ATTRIBUTE(flag)

Sets an attribute on a function object. Expects the function atSTACK[-1]and the attribute value to set atSTACK[-2]; consumes both and leaves thefunction atSTACK[-1]. The flag determines which attribute to set:

  • 0x01 a tuple of default values for positional-only andpositional-or-keyword parameters in positional order

  • 0x02 a dictionary of keyword-only parameters' default values

  • 0x04 a tuple of strings containing parameters' annotations

  • 0x08 a tuple containing cells for free variables, making a closure

在 3.13 版被加入.

BUILD_SLICE(argc)

Pushes a slice object on the stack.argc must be 2 or 3. If it is 2, implements:

end=STACK.pop()start=STACK.pop()STACK.append(slice(start,end))

if it is 3, implements:

step=STACK.pop()end=STACK.pop()start=STACK.pop()STACK.append(slice(start,end,step))

See theslice() built-in function for more information.

EXTENDED_ARG(ext)

Prefixes any opcode which has an argument too big to fit into the default onebyte.ext holds an additional byte which act as higher bits in the argument.For each opcode, at most three prefixalEXTENDED_ARG are allowed, formingan argument from two-byte to four-byte.

CONVERT_VALUE(oparg)

Convert value to a string, depending onoparg:

value=STACK.pop()result=func(value)STACK.append(result)

Used for implementing formatted string literals (f-strings).

在 3.13 版被加入.

FORMAT_SIMPLE

Formats the value on top of stack:

value=STACK.pop()result=value.__format__("")STACK.append(result)

Used for implementing formatted string literals (f-strings).

在 3.13 版被加入.

FORMAT_WITH_SPEC

Formats the given value with the given format spec:

spec=STACK.pop()value=STACK.pop()result=value.__format__(spec)STACK.append(result)

Used for implementing formatted string literals (f-strings).

在 3.13 版被加入.

MATCH_CLASS(count)

STACK[-1] is a tuple of keyword attribute names,STACK[-2] is the classbeing matched against, andSTACK[-3] is the match subject.count is thenumber of positional sub-patterns.

PopSTACK[-1],STACK[-2], andSTACK[-3]. IfSTACK[-3] is aninstance ofSTACK[-2] and has the positional and keyword attributesrequired bycount andSTACK[-1], push a tuple of extracted attributes.Otherwise, pushNone.

在 3.10 版被加入.

在 3.11 版的變更:Previously, this instruction also pushed a boolean value indicatingsuccess (True) or failure (False).

RESUME(context)

A no-op. Performs internal tracing, debugging and optimization checks.

Thecontext oparand consists of two parts. The lowest two bitsindicate where theRESUME occurs:

  • 0 The start of a function, which is neither a generator, coroutinenor an async generator

  • 1 After ayield expression

  • 2 After ayieldfrom expression

  • 3 After anawait expression

The next bit is1 if the RESUME is at except-depth1, and0otherwise.

在 3.11 版被加入.

在 3.13 版的變更:The oparg value changed to include information about except-depth

RETURN_GENERATOR

Create a generator, coroutine, or async generator from the current frame.Used as first opcode of in code object for the above mentioned callables.Clear the current frame and return the newly created generator.

在 3.11 版被加入.

SEND(delta)

Equivalent toSTACK[-1]=STACK[-2].send(STACK[-1]). Used inyieldfromandawait statements.

If the call raisesStopIteration, pop the top value from the stack,push the exception'svalue attribute, and increment the bytecode counterbydelta.

在 3.11 版被加入.

HAVE_ARGUMENT

This is not really an opcode. It identifies the dividing line betweenopcodes in the range [0,255] which don't use their argument and thosethat do (<HAVE_ARGUMENT and>=HAVE_ARGUMENT, respectively).

If your application uses pseudo instructions or specialized instructions,use thehasarg collection instead.

在 3.6 版的變更:Now every instruction has an argument, but opcodes<HAVE_ARGUMENTignore it. Before, only opcodes>=HAVE_ARGUMENT had an argument.

在 3.12 版的變更:Pseudo instructions were added to thedis module, and for themit is not true that comparison withHAVE_ARGUMENT indicates whetherthey use their arg.

在 3.13 版之後被棄用:Usehasarg instead.

CALL_INTRINSIC_1

Calls an intrinsic function with one argument. PassesSTACK[-1] as theargument and setsSTACK[-1] to the result. Used to implementfunctionality that is not performance critical.

The operand determines which intrinsic function is called:

運算元

描述

INTRINSIC_1_INVALID

Not valid

INTRINSIC_PRINT

Prints the argument to standardout. Used in the REPL.

INTRINSIC_IMPORT_STAR

Performsimport* for thenamed module.

INTRINSIC_STOPITERATION_ERROR

Extracts the return value from aStopIteration exception.

INTRINSIC_ASYNC_GEN_WRAP

Wraps an async generator value

INTRINSIC_UNARY_POSITIVE

Performs the unary+operation

INTRINSIC_LIST_TO_TUPLE

Converts a list to a tuple

INTRINSIC_TYPEVAR

Creates atyping.TypeVar

INTRINSIC_PARAMSPEC

Creates atyping.ParamSpec

INTRINSIC_TYPEVARTUPLE

Creates atyping.TypeVarTuple

INTRINSIC_SUBSCRIPT_GENERIC

Returnstyping.Genericsubscripted with the argument

INTRINSIC_TYPEALIAS

Creates atyping.TypeAliasType;used in thetypestatement. The argument is a tupleof the type alias's name,type parameters, and value.

在 3.12 版被加入.

CALL_INTRINSIC_2

Calls an intrinsic function with two arguments. Used to implement functionalitythat is not performance critical:

arg2=STACK.pop()arg1=STACK.pop()result=intrinsic2(arg1,arg2)STACK.append(result)

The operand determines which intrinsic function is called:

運算元

描述

INTRINSIC_2_INVALID

Not valid

INTRINSIC_PREP_RERAISE_STAR

Calculates theExceptionGroup to raisefrom atry-except*.

INTRINSIC_TYPEVAR_WITH_BOUND

Creates atyping.TypeVarwith a bound.

INTRINSIC_TYPEVAR_WITH_CONSTRAINTS

Creates atyping.TypeVar withconstraints.

INTRINSIC_SET_FUNCTION_TYPE_PARAMS

Sets the__type_params__attribute of a function.

在 3.12 版被加入.

Pseudo-instructions

These opcodes do not appear in Python bytecode. They are used by the compilerbut are replaced by real opcodes or removed before bytecode is generated.

SETUP_FINALLY(target)

Set up an exception handler for the following code block. If an exceptionoccurs, the value stack level is restored to its current state and controlis transferred to the exception handler attarget.

SETUP_CLEANUP(target)

LikeSETUP_FINALLY, but in case of an exception also pushes the lastinstruction (lasti) to the stack so thatRERAISE can restore it.If an exception occurs, the value stack level and the last instruction onthe frame are restored to their current state, and control is transferredto the exception handler attarget.

SETUP_WITH(target)

LikeSETUP_CLEANUP, but in case of an exception one more item is poppedfrom the stack before control is transferred to the exception handler attarget.

This variant is used inwith andasyncwithconstructs, which push the return value of the context manager's__enter__() or__aenter__() to the stack.

POP_BLOCK

Marks the end of the code block associated with the lastSETUP_FINALLY,SETUP_CLEANUP orSETUP_WITH.

JUMP
JUMP_NO_INTERRUPT

Undirected relative jump instructions which are replaced by theirdirected (forward/backward) counterparts by the assembler.

LOAD_CLOSURE(i)

Pushes a reference to the cell contained in sloti of the "fast locals"storage.

Note thatLOAD_CLOSURE is replaced withLOAD_FAST in the assembler.

在 3.13 版的變更:This opcode is now a pseudo-instruction.

LOAD_METHOD

Optimized unbound method lookup. Emitted as aLOAD_ATTR opcodewith a flag set in the arg.

Opcode collections

These collections are provided for automatic introspection of bytecodeinstructions:

在 3.12 版的變更:The collections now contain pseudo instructions and instrumentedinstructions as well. These are opcodes with values>=MIN_PSEUDO_OPCODEand>=MIN_INSTRUMENTED_OPCODE.

dis.opname

Sequence of operation names, indexable using the bytecode.

dis.opmap

Dictionary mapping operation names to bytecodes.

dis.cmp_op

Sequence of all compare operation names.

dis.hasarg

Sequence of bytecodes that use their argument.

在 3.12 版被加入.

dis.hasconst

Sequence of bytecodes that access a constant.

dis.hasfree

Sequence of bytecodes that access afree (closure) variable.'free' in this context refers to names in the current scope that arereferenced by inner scopes or names in outer scopes that are referencedfrom this scope. It doesnot include references to global or builtin scopes.

dis.hasname

Sequence of bytecodes that access an attribute by name.

dis.hasjump

Sequence of bytecodes that have a jump target. All jumpsare relative.

在 3.13 版被加入.

dis.haslocal

Sequence of bytecodes that access a local variable.

dis.hascompare

Sequence of bytecodes of Boolean operations.

dis.hasexc

Sequence of bytecodes that set an exception handler.

在 3.12 版被加入.

dis.hasjrel

Sequence of bytecodes that have a relative jump target.

在 3.13 版之後被棄用:All jumps are now relative. Usehasjump.

dis.hasjabs

Sequence of bytecodes that have an absolute jump target.

在 3.13 版之後被棄用:All jumps are now relative. This list is empty.