Python Reference ยท Standard Library: Concurrency, Network and Tools
The dis module
Reference for Python's dis module: dis.dis(), get_instructions, Bytecode, opnames, code objects, and why bytecode differs between Python versions.
When Python runs your code, it first compiles it to bytecode, a compact list of instructions for the interpreter's virtual machine. The dis module turns that bytecode back into readable form, so you can see what a function really does.
It is a learning and debugging tool. Bytecode is a CPython implementation detail that changes between releases, so never write programs that depend on specific instructions.
import dis def add(a, b): return a + b dis.dis(add)
Try it
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.
Syntax
dis.dis(x=None, *, file=None, depth=None, show_caches=False, adaptive=False) dis.get_instructions(x, *, first_line=None, show_caches=False, adaptive=False) dis.Bytecode(x, *, first_line=None, current_offset=None, show_caches=False, adaptive=False) dis.code_info(x) dis.show_code(x, *, file=None) dis.opname
| Name | Description |
|---|---|
x |
A function, method, class, generator, code object, source string or bytes of raw code. With no argument, dis.dis() shows the last traceback. |
file |
Where dis.dis() writes its output. Defaults to sys.stdout. |
depth |
How many levels of nested code objects (inner functions, lambdas) to disassemble. None means all. |
adaptive |
Show the specialized instructions that CPython 3.11+ may have swapped in after warm-up. |
Return value: dis.dis() prints and returns None. get_instructions() returns an iterator of Instruction named tuples with fields such as opname, opcode, arg, argval, argrepr, offset and starts_line. Bytecode(x) is an iterable object with .dis() that returns the listing as a string. code_info() returns a string.
Raises: TypeError if x is an object that has no code.
Usage notes
- Output is version-specific. Python 3.11 added specialized and inline-cache instructions, 3.12 and later renamed or merged several opcodes, and 3.13 and 3.14 changed more. Compare listings only from the same version.
- Each function has a code object in
fn.__code__. Useful attributes includeco_varnames(local names),co_names(global and attribute names),co_consts(constants) andco_argcount. - Constant folding. The compiler evaluates simple constant expressions such as
60 * 60 * 24ahead of time, so you will find86400among the constants. - Nested code is included. A function containing a lambda or inner
defshows those as separate code objects;depth=1limits the listing to the outer function. dis.Bytecode(fn).dis()returns a string, handy when you want to compare or store the listing instead of printing it.- Reading a listing: each line has the source line number (on the first instruction of a line), the byte offset, the opcode name and, for many opcodes, the argument followed by its meaning in parentheses. Jump targets are marked with
>>. - Do not use bytecode length as a speed measure. Fewer instructions can still be slower: a call into C such as
sum()is one instruction that does a lot of work. - Not a security boundary. Reading bytecode is fine, but loading untrusted bytecode with
marshalcan execute arbitrary code.