File: //proc/self/root/usr/lib/python2.4/pickletools.pyo
m
=5Dc C @ s d Z d d g Z d Z d Z d Z d e f d YZ d k l Z d Z
e d
d d d
d e
d d Z d Z e d
d d d d e d d Z
d Z e d
d d d d e d d Z e e d Z e d
d d e d e d d Z d Z e d
d d e d e d d Z d Z e d
d d e d e d d! Z d" Z e d
d# d e d e d d$ Z d% Z e d
d& d e d e d d' Z d( Z e d
d) d e d e d d* Z d+ Z e d
d, d e d e d d- Z d. Z d/ Z e d
d0 d e d e d d1 Z! e d
d2 d e d e d d3 Z" d4 Z# e d
d5 d e d e# d d6 Z$ d7 Z% e d
d8 d d9 d e% d d: Z& d; k' l( Z( d< Z) e d
d= d e d e) d d> Z* d? Z+ e d
d@ d e d e+ d dA Z, dB e f dC YZ- e- d
dD dE e. d dF Z/ e- d
dG dE e0 d dH Z1 e- d
dI dE e. e0 e2 f d dJ Z3 e- d
dK dE e2 f d dL Z4 e- d
dM dE e5 d dN Z6 e- d
dO dE e7 d dP Z8 e- d
dQ dE e9 d dR Z: e- d
dS dE e; d d dT Z= e- d
dU dE e> d dV Z? e- d
dW dE e@ d dX ZA e- d
dY dE eB d dZ ZC e- d
d[ dE e d d\ ZD e- d
d] dE e- d d^ ZE e- d
d_ dE e- d d` ZF da e f db YZG eG ZH eH d
dc dd de df e! dg g dh e3 g di dj d dk eH d
dl dd dm df e dg g dh e/ g di d
d dn eH d
do dd dp df e dg g dh e/ g di d
d dq eH d
dr dd ds df e
dg g dh e/ g di d
d dt eH d
du dd dv df e" dg g dh e1 g di dj d dw eH d
dx dd dy df e* dg g dh e1 g di d d dz eH d
d{ dd d| df e, dg g dh e1 g di d d d} eH d
d~ dd d df e dg g dh e8 g di dj d d eH d
d dd d df e dg g dh e8 g di d
d d eH d
d dd d df e dg g dh e8 g di d
d d eH d
d dd d df ddg g dh e= g di dj d d eH d
d dd d df ddg g dh e4 g di d d d eH d
d dd d df ddg g dh e4 g di d d d eH d
d dd d df e dg g dh e: g di dj d d eH d
d dd d df e dg g dh e: g di d
d d eH d
d dd d df e$ dg g dh e6 g di dj d d eH d
d dd d df e& dg g dh e6 g di d
d d eH d
d dd d df ddg g dh eA g di d
d d eH d
d dd d df ddg eA eD g dh eA g di dj d d eH d
d dd d df ddg eA eE eF g dh eA g di d
d d eH d
d dd d df ddg eE eF g dh eA g di dj d d eH d
d dd d df ddg g dh e? g di d
d d eH d
d dd d df ddg eE eF g dh e? g di dj d d eH d
d dd d df ddg eD g dh e? g di d d d eH d
d dd d df ddg eD eD g dh e? g di d d d eH d
d dd d df ddg eD eD eD g dh e? g di d d d eH d
d dd d df ddg g dh eC g di d
d d eH d
d dd d df ddg eE eF g dh eC g di dj d d eH d
d dd d df ddg eC eD eD g dh eC g di dj d d eH d
d dd d df ddg eC eE eF g dh eC g di d
d d eH d
d dd d df ddg eD g dh g di dj d d eH d
d dd d df ddg eD g dh eD eD g di dj d d eH d
d dd d df ddg g dh eE g di dj d d eH d
d dd d df ddg eE eF g dh g di dj d d eH d
d dd d df e! dg g dh eD g di dj d d eH d
d dd d df e dg g dh eD g di d
d d eH d
d dd d df e dg g dh eD g di d
d d eH d
d dd d df e! dg g dh g di dj d d eH d
d dd d df e dg g dh g di d
d d eH d
d dd d df e dg g dh g di d
d d eH d
d dd d df e dg g dh eD g di d d d eH d
d dd d df e
dg g dh eD g di d d d eH d
d dd d df e dg g dh eD g di d d d eH d
d dd d df e dg g dh eD g di dj d d eH d
d dd d df ddg eD eD g dh eD g di dj d d eH d
d dd d df ddg eD eD g dh eD g di dj d d eH d
d dd d df e dg eE eF g dh eD g di dj d d eH d
d dd d df ddg eE eD eF g dh eD g di d
d d eH d
d dd d df ddg eD eD g dh eD g di d d d eH d
d dd d df e dg g dh g di d d d eH d
d dd d df ddg eD g dh g di dj d d eH d
ddd ddf e dg g dh eD g di dj d d eH d
ddd ddf ddg eD g dh eD g di d
d d g5 ZI [H h ZJ h ZK x eL eI D] \ ZM ZN eN iO eJ j o' eP deN iO eJ eN iO eM f n eN iQ eK j o' eP d eN iQ eK eN iQ eM f n eM eJ eN iO <eM eK eN iQ <qW[J [K [M [N h ZR x eI D] ZN eN eR eN iQ <qW[N eS d
ZT eT [T d ZU ddd d ZV d
ZW dZX h deW <deX <ZY d ZZ e[ dj o eZ n dS( sr "Executable documentation" for the pickle module.
Extensive comments about the pickle protocols and pickle-machine opcodes
can be found here. Some functions meant for external use:
genops(pickle)
Generate all the opcodes in a pickle, as (opcode, arg, position) triples.
dis(pickle, out=None, memo=None, indentlevel=4)
Print a symbolic disassembly of a pickle.
t dist genopsiiit ArgumentDescriptorc B s t Z d Z d Z RS( Nt namet nt readert docc C s( | | _ | | _ | | _ | | _ d S( N( R t selfR R R ( R R R R R ( ( t! /usr/lib/python2.4/pickletools.pyt __init__ s ( s nameR s readers doc( t __name__t
__module__t __slots__R ( ( ( R R s ( s unpackc C s4 | i d } | o t | Sn t d d S( sS
>>> import StringIO
>>> read_uint1(StringIO.StringIO('\xff'))
255
i s' not enough data in stream to read uint1N( t ft readt datat ordt
ValueError( R
R ( ( R t
read_uint1 s
R t uint1R i R R s One-byte unsigned integer.c C sG | i d } t | d j o t d | d Sn t d d S( s
>>> import StringIO
>>> read_uint2(StringIO.StringIO('\xff\x00'))
255
>>> read_uint2(StringIO.StringIO('\xff\xff'))
65535
i s <Hi s' not enough data in stream to read uint2N( R
R R t lent _unpackR ( R
R ( ( R t
read_uint2 s
t uint2i s) Two-byte unsigned integer, little-endian.c C sG | i d } t | d j o t d | d Sn t d d S( s
>>> import StringIO
>>> read_int4(StringIO.StringIO('\xff\x00\x00\x00'))
255
>>> read_int4(StringIO.StringIO('\x00\x00\x00\x80')) == -(2**31)
True
i s <ii s& not enough data in stream to read int4N( R
R R R R R ( R
R ( ( R t read_int4 s
t int4i s8 Four-byte signed integer, little-endian, 2's complement.c C s | i } | i d p t d n | d } | oq xn d D]R } | i | o< | i | p t d | | f n | d d !} PqD qD Wt d | n | o | i d } n | S(
s
>>> import StringIO
>>> read_stringnl(StringIO.StringIO("'abcd'\nefg\n"))
'abcd'
>>> read_stringnl(StringIO.StringIO("\n"))
Traceback (most recent call last):
...
ValueError: no string quotes around ''
>>> read_stringnl(StringIO.StringIO("\n"), stripquotes=False)
''
>>> read_stringnl(StringIO.StringIO("''\n"))
''
>>> read_stringnl(StringIO.StringIO('"abcd"'))
Traceback (most recent call last):
...
ValueError: no newline found when trying to read stringnl
Embedded escapes are undone in the result.
>>> read_stringnl(StringIO.StringIO(r"'a\n\\b\x00c\td'" + "\n'e'"))
'a\n\\b\x00c\td'
s
s- no newline found when trying to read stringnlis '"s, strinq quote %r not found at both ends of %ri s no string quotes around %rt
string_escapeN( R
t readlineR t endswithR t stripquotest qt
startswitht decode( R
R R R R ( ( R t
read_stringnl s"
t stringnls A newline-terminated string.
This is a repr-style string, with embedded escapes, and
bracketing quotes.
c C s t | d t d t S( NR R ( R! R
t False( R
( ( R t read_stringnl_noescapeC s t stringnl_noescapes A newline-terminated string.
This is a str-style string, without embedded escapes,
or bracketing quotes. It should consist solely of
printable ASCII characters.
c C s d t | t | f S( s|
>>> import StringIO
>>> read_stringnl_noescape_pair(StringIO.StringIO("Queue\nEmpty\njunk"))
'Queue Empty'
s %s %sN( R$ R
( R
( ( R t read_stringnl_noescape_pairQ s t stringnl_noescape_pairs A pair of newline-terminated strings.
These are str-style strings, without embedded
escapes, or bracketing quotes. They should
consist solely of printable ASCII characters.
The pair is returned as a single string, with
a single blank separating the two strings.
c C sw t | } | d j o t d | n | i | } t | | j o | Sn t d | t | f d S( sh
>>> import StringIO
>>> read_string4(StringIO.StringIO("\x00\x00\x00\x00abc"))
''
>>> read_string4(StringIO.StringIO("\x03\x00\x00\x00abcdef"))
'abc'
>>> read_string4(StringIO.StringIO("\x00\x00\x00\x03abcdef"))
Traceback (most recent call last):
...
ValueError: expected 50331648 bytes in a string4, but only 6 remain
i s string4 byte count < 0: %ds2 expected %d bytes in a string4, but only %d remainN( R R
R R R R R ( R
R R ( ( R t read_string4g s
t string4s A counted string.
The first argument is a 4-byte little-endian signed int giving
the number of bytes in the string, and the second argument is
that many bytes.
c C sV t | } | i | } t | | j o | Sn t d | t | f d S( s
>>> import StringIO
>>> read_string1(StringIO.StringIO("\x00"))
''
>>> read_string1(StringIO.StringIO("\x03abcdef"))
'abc'
s2 expected %d bytes in a string1, but only %d remainN( R R
R R R R R ( R
R R ( ( R t read_string1 s t string1s A counted string.
The first argument is a 1-byte unsigned int giving the number
of bytes in the string, and the second argument is that many
bytes.
c C sC | i } | i d p t d n | d } t | d S( sq
>>> import StringIO
>>> read_unicodestringnl(StringIO.StringIO("abc\uabcd\njunk"))
u'abc\uabcd'
s
s4 no newline found when trying to read unicodestringnlis raw-unicode-escapeN( R
R R R R t unicode( R
R ( ( R t read_unicodestringnl s
t unicodestringnls A newline-terminated Unicode string.
This is raw-unicode-escape encoded, so consists of
printable ASCII characters, and may contain embedded
escape sequences.
c C s t | } | d j o t d | n | i | } t | | j o t | d Sn t d | t | f d S( s
>>> import StringIO
>>> s = u'abcd\uabcd'
>>> enc = s.encode('utf-8')
>>> enc
'abcd\xea\xaf\x8d'
>>> n = chr(len(enc)) + chr(0) * 3 # little-endian 4-byte length
>>> t = read_unicodestring4(StringIO.StringIO(n + enc + 'junk'))
>>> s == t
True
>>> read_unicodestring4(StringIO.StringIO(n + enc[:-1]))
Traceback (most recent call last):
...
ValueError: expected 7 bytes in a unicodestring4, but only 6 remain
i s! unicodestring4 byte count < 0: %ds utf-8s9 expected %d bytes in a unicodestring4, but only %d remainN( R R
R R R R R R, ( R
R R ( ( R t read_unicodestring4 s
t unicodestring4sA A counted Unicode string.
The first argument is a 4-byte little-endian signed int
giving the number of bytes in the string, and the second
argument-- the UTF-8 encoding of the Unicode string --
contains that many bytes.
c C s t | d t d t } | i d o t d | n | d j o t Sn | d j o t Sn y t | SWn t j
o t | Sn Xd S( s
>>> import StringIO
>>> read_decimalnl_short(StringIO.StringIO("1234\n56"))
1234
>>> read_decimalnl_short(StringIO.StringIO("1234L\n56"))
Traceback (most recent call last):
...
ValueError: trailing 'L' not allowed in '1234L'
R R t Ls trailing 'L' not allowed in %rt 00t 01N(
R! R
R# t sR R t Truet intt
OverflowErrort long( R
R4 ( ( R t read_decimalnl_short s
c C sF t | d t d t } | i d p t d | n t | S( s
>>> import StringIO
>>> read_decimalnl_long(StringIO.StringIO("1234\n56"))
Traceback (most recent call last):
...
ValueError: trailing 'L' required in '1234'
Someday the trailing 'L' will probably go away from this output.
>>> read_decimalnl_long(StringIO.StringIO("1234L\n56"))
1234L
>>> read_decimalnl_long(StringIO.StringIO("123456789012345678901234L\n6"))
123456789012345678901234L
R R R1 s trailing 'L' required in %rN( R! R
R# R4 R R R8 ( R
R4 ( ( R t read_decimalnl_long s
t decimalnl_shorts A newline-terminated decimal integer literal.
This never has a trailing 'L', and the integer fit
in a short Python int on the box where the pickle
was written -- but there's no guarantee it will fit
in a short Python int on the box where the pickle
is read.
t decimalnl_longs A newline-terminated decimal integer literal.
This has a trailing 'L', and can represent integers
of any size.
c C s" t | d t d t } t | S( s[
>>> import StringIO
>>> read_floatnl(StringIO.StringIO("-1.25\n6"))
-1.25
R R N( R! R
R# R4 t float( R
R4 ( ( R t read_floatnl4 s t floatnls A newline-terminated decimal floating literal.
In general this requires 17 significant digits for roundtrip
identity, and pickling then unpickling infinities, NaNs, and
minus zero doesn't work across boxes, or on some boxes even
on itself (e.g., Windows can't read the strings it produces
for infinities or NaNs).
c C sG | i d } t | d j o t d | d Sn t d d S( s
>>> import StringIO, struct
>>> raw = struct.pack(">d", -1.25)
>>> raw
'\xbf\xf4\x00\x00\x00\x00\x00\x00'
>>> read_float8(StringIO.StringIO(raw + "\n"))
-1.25
i s >di s( not enough data in stream to read float8N( R
R R R R R ( R
R ( ( R t read_float8J s
t float8i s An 8-byte binary representation of a float, big-endian.
The format is unique to Python, and shared with the struct
module (format string '>d') "in theory" (the struct and cPickle
implementations don't share the code -- they should). It's
strongly related to the IEEE-754 double format, and, in normal
cases, is in fact identical to the big-endian 754 double format.
On other boxes the dynamic range is limited to that of a 754
double, and "add a half and chop" rounding is used to reduce
the precision to 53 bits. However, even on a 754 box,
infinities, NaNs, and minus zero may not be handled correctly
(may not survive roundtrip pickling intact).
( s decode_longc C sH t | } | i | } t | | j o t d n t | S( sT
>>> import StringIO
>>> read_long1(StringIO.StringIO("\x00"))
0L
>>> read_long1(StringIO.StringIO("\x02\xff\x00"))
255L
>>> read_long1(StringIO.StringIO("\x02\xff\x7f"))
32767L
>>> read_long1(StringIO.StringIO("\x02\x00\xff"))
-256L
>>> read_long1(StringIO.StringIO("\x02\x00\x80"))
-32768L
s' not enough data in stream to read long1N( R R
R R R R R t decode_long( R
R R ( ( R t
read_long1p s
t long1s A binary long, little-endian, using 1-byte size.
This first reads one byte as an unsigned size, then reads that
many bytes and interprets them as a little-endian 2's-complement long.
If the size is 0, that's taken as a shortcut for the long 0L.
c C si t | } | d j o t d | n | i | } t | | j o t d n t | S( s
>>> import StringIO
>>> read_long4(StringIO.StringIO("\x02\x00\x00\x00\xff\x00"))
255L
>>> read_long4(StringIO.StringIO("\x02\x00\x00\x00\xff\x7f"))
32767L
>>> read_long4(StringIO.StringIO("\x02\x00\x00\x00\x00\xff"))
-256L
>>> read_long4(StringIO.StringIO("\x02\x00\x00\x00\x00\x80"))
-32768L
>>> read_long1(StringIO.StringIO("\x00\x00\x00\x00"))
0L
i s long4 byte count < 0: %ds' not enough data in stream to read long4N( R R
R R R R R RB ( R
R R ( ( R t
read_long4 s
t long4s A binary representation of a long, little-endian.
This first reads four bytes as a signed size (but requires the
size to be >= 0), then reads that many bytes and interprets them
as a little-endian 2's-complement long. If the size is 0, that's taken
as a shortcut for the long 0L, although LONG1 should really be used
then instead (and in any case where # of bytes < 256).
t StackObjectc B s t Z d Z d Z d Z RS( NR t obtypeR c C sD | | _ t | t o x | D] } q Wn | | _ | | _ d S( N( R R t
isinstanceRH t tuplet containedR ( R R RH R RK ( ( R R s c C s | i S( N( R R ( R ( ( R t __repr__ s ( s names obtypes doc( R
R R R RL ( ( ( R RG s
R6 RH s3 A short (as opposed to long) Python integer object.R8 s3 A long (as opposed to short) Python integer object.t int_or_bools: A Python integer object (short or long), or a Python bool.t bools A Python bool object.R= s A Python float object.t strs A Python string object.R, s A Python Unicode string object.t Nones The Python None object.RJ s A Python tuple object.t lists A Python list object.t dicts A Python dict object.t anys Any kind of object whatsoever.t marks 'The mark' is a unique object.
Opcodes that operate on a variable number of objects
generally don't embed the count of objects in the opcode,
or pull it off the stack. Instead the MARK opcode is used
to push a special marker object on the stack, and then
some other opcodes grab all the objects from the top of
the stack down to (but not including) the topmost marker
object.
t
stackslicesK An object representing a contiguous slice of the stack.
This is used in conjuction with markobject, to represent all
of the stack following the topmost markobject. For example,
the POP_MARK opcode changes the stack from
[..., markobject, stackslice]
to
[...]
No matter how many object are on the stack after the topmost
markobject, POP_MARK gets rid of all of them (including the
topmost markobject too).
t
OpcodeInfoc B s t Z d Z d Z RS(
NR t codet argt stack_beforet stack_aftert protoR c C se | | _ | | _ | | _ x | D] } q" W| | _ x | D] } q< W| | _ | | _ | | _ d S( N( R R RW RX RY t xRZ R[ R ( R R RW RX RY RZ R[ R R\ ( ( R R X s* ( s names codes args stack_befores stack_afters protos doc( R
R R R ( ( ( R RV 9 s t INTRW t IRX RY RZ R[ i s Push an integer or bool.
The argument is a newline-terminated decimal literal string.
The intent may have been that this always fit in a short Python int,
but INT can be generated in pickles written on a 64-bit box that
require a Python long on a 32-bit box. The difference between this
and LONG then is that INT skips a trailing 'L', and produces a short
int whenever possible.
Another difference is due to that, when bool was introduced as a
distinct type in 2.3, builtin names True and False were also added to
2.2.2, mapping to ints 1 and 0. For compatibility in both directions,
True gets pickled as INT + "I01\n", and False as INT + "I00\n".
Leading zeroes are never produced for a genuine integer. The 2.3
(and later) unpicklers special-case these and return bool instead;
earlier unpicklers ignore the leading "0" and return the int.
t BININTt Js1 Push a four-byte signed integer.
This handles the full range of Python (short) integers on a 32-bit
box, directly as binary bytes (1 for the opcode and 4 for the integer).
If the integer is non-negative and fits in 1 or 2 bytes, pickling via
BININT1 or BININT2 saves space.
t BININT1t Ks Push a one-byte unsigned integer.
This is a space optimization for pickling very small non-negative ints,
in range(256).
t BININT2t Ms Push a two-byte unsigned integer.
This is a space optimization for pickling small positive ints, in
range(256, 2**16). Integers in range(256) can also be pickled via
BININT2, but BININT1 instead saves a byte.
t LONGR1 s Push a long integer.
The same as INT, except that the literal ends with 'L', and always
unpickles to a Python long. There doesn't seem a real purpose to the
trailing 'L'.
Note that LONG takes time quadratic in the number of digits when
unpickling (this is simply due to the nature of decimal->binary
conversion). Proto 2 added linear-time (in C; still quadratic-time
in Python) LONG1 and LONG4 opcodes.
t LONG1s s| Long integer using one-byte length.
A more efficient encoding of a Python long; the long1 encoding
says it all.t LONG4s s~ Long integer using found-byte length.
A more efficient encoding of a Python long; the long4 encoding
says it all.t STRINGt Ss Push a Python string object.
The argument is a repr-style string, with bracketing quote characters,
and perhaps embedded escapes. The argument extends until the next
newline character.
t BINSTRINGt Ts Push a Python string object.
There are two arguments: the first is a 4-byte little-endian signed int
giving the number of bytes in the string, and the second is that many
bytes, which are taken literally as the string content.
t SHORT_BINSTRINGt Us Push a Python string object.
There are two arguments: the first is a 1-byte unsigned int giving
the number of bytes in the string, and the second is that many bytes,
which are taken literally as the string content.
t NONEt Ns Push None on the stack.t NEWTRUEs s& True.
Push True onto the stack.t NEWFALSEs s' True.
Push False onto the stack.t UNICODEt Vs Push a Python Unicode string object.
The argument is a raw-unicode-escape encoding of a Unicode string,
and so may contain embedded escape sequences. The argument extends
until the next newline character.
t
BINUNICODEt Xs Push a Python Unicode string object.
There are two arguments: the first is a 4-byte little-endian signed int
giving the number of bytes in the string. The second is that many
bytes, and is the UTF-8 encoding of the Unicode string.
t FLOATt Fs Newline-terminated decimal float literal.
The argument is repr(a_float), and in general requires 17 significant
digits for roundtrip conversion to be an identity (this is so for
IEEE-754 double precision values, which is what Python float maps to
on most boxes).
In general, FLOAT cannot be used to transport infinities, NaNs, or
minus zero across boxes (or even on a single box, if the platform C
library can't read the strings it produces for such things -- Windows
is like that), but may do less damage than BINFLOAT on boxes with
greater precision or dynamic range than IEEE-754 double.
t BINFLOATt Gs Float stored in binary form, with 8 bytes of data.
This generally requires less than half the space of FLOAT encoding.
In general, BINFLOAT cannot be used to transport infinities, NaNs, or
minus zero, raises an exception if the exponent exceeds the range of
an IEEE-754 double, and retains no more than 53 bits of precision (if
there are more than that, "add a half and chop" rounding is used to
cut it back to 53 significant bits).
t
EMPTY_LISTt ]s Push an empty list.t APPENDt as Append an object to a list.
Stack before: ... pylist anyobject
Stack after: ... pylist+[anyobject]
although pylist is really extended in-place.
t APPENDSt es Extend a list by a slice of stack objects.
Stack before: ... pylist markobject stackslice
Stack after: ... pylist+stackslice
although pylist is really extended in-place.
t LISTt lss Build a list out of the topmost stack slice, after markobject.
All the stack entries following the topmost markobject are placed into
a single Python list, which single list object replaces all of the
stack from the topmost markobject onward. For example,
Stack before: ... markobject 1 2 3 'abc'
Stack after: ... [1, 2, 3, 'abc']
t EMPTY_TUPLEt )s Push an empty tuple.t TUPLEt tsv Build a tuple out of the topmost stack slice, after markobject.
All the stack entries following the topmost markobject are placed into
a single Python tuple, which single tuple object replaces all of the
stack from the topmost markobject onward. For example,
Stack before: ... markobject 1 2 3 'abc'
Stack after: ... (1, 2, 3, 'abc')
t TUPLE1s s One-tuple.
This code pops one value off the stack and pushes a tuple of
length 1 whose one item is that value back onto it. IOW:
stack[-1] = tuple(stack[-1:])
t TUPLE2s s One-tuple.
This code pops two values off the stack and pushes a tuple
of length 2 whose items are those values back onto it. IOW:
stack[-2:] = [tuple(stack[-2:])]
t TUPLE3s s One-tuple.
This code pops three values off the stack and pushes a tuple
of length 3 whose items are those values back onto it. IOW:
stack[-3:] = [tuple(stack[-3:])]
t
EMPTY_DICTt }s Push an empty dict.t DICTt ds Build a dict out of the topmost stack slice, after markobject.
All the stack entries following the topmost markobject are placed into
a single Python dict, which single dict object replaces all of the
stack from the topmost markobject onward. The stack slice alternates
key, value, key, value, .... For example,
Stack before: ... markobject 1 2 3 'abc'
Stack after: ... {1: 2, 3: 'abc'}
t SETITEMR4 s Add a key+value pair to an existing dict.
Stack before: ... pydict key value
Stack after: ... pydict
where pydict has been modified via pydict[key] = value.
t SETITEMSt us\ Add an arbitrary number of key+value pairs to an existing dict.
The slice of the stack following the topmost markobject is taken as
an alternating sequence of keys and values, added to the dict
immediately under the topmost markobject. Everything at and after the
topmost markobject is popped, leaving the mutated dict at the top
of the stack.
Stack before: ... pydict markobject key_1 value_1 ... key_n value_n
Stack after: ... pydict
where pydict has been modified via pydict[key_i] = value_i for i in
1, 2, ..., n, and in that order.
t POPt 0s< Discard the top stack item, shrinking the stack by one item.t DUPt 2s= Push the top stack item onto the stack again, duplicating it.t MARKt (s Push markobject onto the stack.
markobject is a unique object, used by other opcodes to identify a
region of the stack containing a variable number of objects for them
to work on. See markobject.doc for more detail.
t POP_MARKt 1s Pop all the stack objects at and above the topmost markobject.
When an opcode using a variable number of stack objects is done,
POP_MARK is used to remove those objects, and to remove the markobject
that delimited their starting position on the stack.
t GETt gs Read an object from the memo and push it on the stack.
The index of the memo object to push is given by the newline-teriminated
decimal string following. BINGET and LONG_BINGET are space-optimized
versions.
t BINGETt hs Read an object from the memo and push it on the stack.
The index of the memo object to push is given by the 1-byte unsigned
integer following.
t LONG_BINGETt js Read an object from the memo and push it on the stack.
The index of the memo object to push is given by the 4-byte signed
little-endian integer following.
t PUTt ps Store the stack top into the memo. The stack is not popped.
The index of the memo location to write into is given by the newline-
terminated decimal string following. BINPUT and LONG_BINPUT are
space-optimized versions.
t BINPUTR s Store the stack top into the memo. The stack is not popped.
The index of the memo location to write into is given by the 1-byte
unsigned integer following.
t LONG_BINPUTt rs Store the stack top into the memo. The stack is not popped.
The index of the memo location to write into is given by the 4-byte
signed little-endian integer following.
t EXT1s s Extension code.
This code and the similar EXT2 and EXT4 allow using a registry
of popular objects that are pickled by name, typically classes.
It is envisioned that through a global negotiation and
registration process, third parties can set up a mapping between
ints and object names.
In order to guarantee pickle interchangeability, the extension
code registry ought to be global, although a range of codes may
be reserved for private use.
EXT1 has a 1-byte integer argument. This is used to index into the
extension registry, and the object at that index is pushed on the stack.
t EXT2s sN Extension code.
See EXT1. EXT2 has a two-byte integer argument.
t EXT4s sO Extension code.
See EXT1. EXT4 has a four-byte integer argument.
t GLOBALt cs Push a global object (module.attr) on the stack.
Two newline-terminated strings follow the GLOBAL opcode. The first is
taken as a module name, and the second as a class name. The class
object module.class is pushed on the stack. More accurately, the
object returned by self.find_class(module, class) is pushed on the
stack, so unpickling subclasses can override this form of lookup.
t REDUCEt Rsg Push an object built from a callable and an argument tuple.
The opcode is named to remind of the __reduce__() method.
Stack before: ... callable pytuple
Stack after: ... callable(*pytuple)
The callable and the argument tuple are the first two items returned
by a __reduce__ method. Applying the callable to the argtuple is
supposed to reproduce the original object, or at least get it started.
If the __reduce__ method returns a 3-tuple, the last component is an
argument to be passed to the object's __setstate__, and then the REDUCE
opcode is followed by code to create setstate's argument, and then a
BUILD opcode to apply __setstate__ to that argument.
There are lots of special cases here. The argtuple can be None, in
which case callable.__basicnew__() is called instead to produce the
object to be pushed on the stack. This appears to be a trick unique
to ExtensionClasses, and is deprecated regardless.
If type(callable) is not ClassType, REDUCE complains unless the
callable has been registered with the copy_reg module's
safe_constructors dict, or the callable has a magic
'__safe_for_unpickling__' attribute with a true value. I'm not sure
why it does this, but I've sure seen this complaint often enough when
I didn't want to <wink>.
t BUILDt bs Finish building an object, via __setstate__ or dict update.
Stack before: ... anyobject argument
Stack after: ... anyobject
where anyobject may have been mutated, as follows:
If the object has a __setstate__ method,
anyobject.__setstate__(argument)
is called.
Else the argument must be a dict, the object must have a __dict__, and
the object is updated via
anyobject.__dict__.update(argument)
This may raise RuntimeError in restricted execution mode (which
disallows access to __dict__ directly); in that case, the object
is updated instead via
for k, v in argument.items():
anyobject[k] = v
t INSTt is Build a class instance.
This is the protocol 0 version of protocol 1's OBJ opcode.
INST is followed by two newline-terminated strings, giving a
module and class name, just as for the GLOBAL opcode (and see
GLOBAL for more details about that). self.find_class(module, name)
is used to get a class object.
In addition, all the objects on the stack following the topmost
markobject are gathered into a tuple and popped (along with the
topmost markobject), just as for the TUPLE opcode.
Now it gets complicated. If all of these are true:
+ The argtuple is empty (markobject was at the top of the stack
at the start).
+ It's an old-style class object (the type of the class object is
ClassType).
+ The class object does not have a __getinitargs__ attribute.
then we want to create an old-style class instance without invoking
its __init__() method (pickle has waffled on this over the years; not
calling __init__() is current wisdom). In this case, an instance of
an old-style dummy class is created, and then we try to rebind its
__class__ attribute to the desired class object. If this succeeds,
the new instance object is pushed on the stack, and we're done. In
restricted execution mode it can fail (assignment to __class__ is
disallowed), and I'm not really sure what happens then -- it looks
like the code ends up calling the class object's __init__ anyway,
via falling into the next case.
Else (the argtuple is not empty, it's not an old-style class object,
or the class object does have a __getinitargs__ attribute), the code
first insists that the class object have a __safe_for_unpickling__
attribute. Unlike as for the __safe_for_unpickling__ check in REDUCE,
it doesn't matter whether this attribute has a true or false value, it
only matters whether it exists (XXX this is a bug; cPickle
requires the attribute to be true). If __safe_for_unpickling__
doesn't exist, UnpicklingError is raised.
Else (the class object does have a __safe_for_unpickling__ attr),
the class object obtained from INST's arguments is applied to the
argtuple obtained from the stack, and the resulting instance object
is pushed on the stack.
NOTE: checks for __safe_for_unpickling__ went away in Python 2.3.
t OBJt os Build a class instance.
This is the protocol 1 version of protocol 0's INST opcode, and is
very much like it. The major difference is that the class object
is taken off the stack, allowing it to be retrieved from the memo
repeatedly if several instances of the same class are created. This
can be much more efficient (in both time and space) than repeatedly
embedding the module and class names in INST opcodes.
Unlike INST, OBJ takes no arguments from the opcode stream. Instead
the class object is taken off the stack, immediately above the
topmost markobject:
Stack before: ... markobject classobject stackslice
Stack after: ... new_instance_object
As for INST, the remainder of the stack above the markobject is
gathered into an argument tuple, and then the logic seems identical,
except that no __safe_for_unpickling__ check is done (XXX this is
a bug; cPickle does test __safe_for_unpickling__). See INST for
the gory details.
NOTE: In Python 2.3, INST and OBJ are identical except for how they
get the class object. That was always the intent; the implementations
had diverged for accidental reasons.
t NEWOBJs sL Build an object instance.
The stack before should be thought of as containing a class
object followed by an argument tuple (the tuple being the stack
top). Call these cls and args. They are popped off the stack,
and the value returned by cls.__new__(cls, *args) is pushed back
onto the stack.
t PROTOs s Protocol version indicator.
For protocol 2 and above, a pickle must start with this opcode.
The argument is the protocol version, an int in range(2, 256).
t STOPt .s Stop the unpickling machine.
Every pickle ends with this opcode. The object at the top of the stack
is popped, and that's the result of unpickling. The stack should be
empty then.
t PERSIDt Ps Push an object identified by a persistent ID.
The pickle module doesn't define what a persistent ID means. PERSID's
argument is a newline-terminated str-style (no embedded escapes, no
bracketing quote characters) string, which *is* "the persistent ID".
The unpickler passes this string to self.persistent_load(). Whatever
object that returns is pushed on the stack. There is no implementation
of persistent_load() in Python's unpickler: it must be supplied by an
unpickler subclass.
t BINPERSIDt QsX Push an object identified by a persistent ID.
Like PERSID, except the persistent ID is popped off the stack (instead
of being a string embedded in the opcode bytestream). The persistent
ID is passed to self.persistent_load(), and whatever object that
returns is pushed on the stack. See PERSID for more detail.
s% repeated name %r at indices %d and %ds% repeated code %r at indices %d and %dc C s d k } d k } t i } x| i D]
} | i d | p | o
d | GHq( q( n t | | } t
| t p t | d j o! | o d | | f GHq( q( n | | j o_ | o d | | f GHn | | } | i | j o t d | | | i f n | | =q( t d | | f q( W| oY d g } x4 | i D]& \ } } | i d | i | f qSWt d
i | n d S( Ns [A-Z][A-Z0-9_]+$s0 skipping %r: it doesn't look like an opcode namei s5 skipping %r: value %r doesn't look like a pickle codes+ checking name %r w/ code %r for consistencysB for pickle code %r, pickle.py uses name %r but we're using name %rsP pickle.py appears to have a pickle opcode with name %r and code %r, but we don'ts= we appear to have pickle opcodes that pickle.py doesn't have:s name %r with code %rs
( t picklet ret code2opt copyt __all__R t matcht verboset getattrt
picklecodeRI RO R R R t msgt itemsRW t appendt join( R R R R RW R R R R ( ( R t assure_pickle_consistency s6
$
c c s/ d k } t | t o | i | } n t | d o
| i } n
d } x t o | } | i
d } t i
| } | d j oK | d j o t d q t d | d j o d p | | f n | i d j o
d } n | i i | } | | | f V| d j o PqU qU Wd S(
sz Generate all the opcodes in a pickle.
'pickle' is a file-like object, or string, containing the pickle.
Each opcode in the pickle is generated, from the current pickle position,
stopping after a STOP opcode is delivered. A triple is generated for
each opcode:
opcode, arg, pos
opcode is an OpcodeInfo record, describing the current opcode.
If the opcode has an argument embedded in the pickle, arg is its decoded
value, as a Python object. If the opcode doesn't have an argument, arg
is None.
If the pickle has a tell() method, pos was the value of pickle.tell()
before reading the current opcode. If the pickle is a string object,
it's wrapped in a StringIO object, and the latter's tell() result is
used. Else (the pickle doesn't have a tell(), and it's not obvious how
to query its current position) pos is None.
Nt tellc C s d S( N( RP ( ( ( R t <lambda>3 s i t s# pickle exhausted before seeing STOPs! at position %s, opcode %r unknowns <unknown>R ( t cStringIOt StringIORI R RO t hasattrR t getposR5 t posR RW R t gett opcodeRP R RX R ( R RW R R R R RX ( ( R R s.
.
c C s g } | d j o
h } n d } g } d | } d } x!t | D]\ } } } | d j o | d | In d t | i d d !| t | | i f } t | | i } | i } | i }
t | } d } t | j p( | i d j o | o | d t j o | o | i }
|
d j o
d } n d |
} x | d t j o | i qFW| i y | i t } Wqt j
o
d } qXqd
} } n | i d j oY | | j o d | } qV| p
d } qV| d t j o
d } qV| d | | <n= | i d j o, | | j o | | g }
qVd | } n | d j p | o] | d d t | i 7} | d j o | d t | 7} n | o | d | 7} qn | | IJ| o t | n t | | j o t d | t | f n | o | | 3n t |
j o | i | n | i! |
qF W| d I| IJ| o t d | n d S( s Produce a symbolic disassembly of a pickle.
'pickle' is a file-like object, or string, containing a (at least one)
pickle. The pickle is disassembled from the current position, through
the first STOP opcode encountered.
Optional arg 'out' is a file-like object to which the disassembly is
printed. It defaults to sys.stdout.
Optional arg 'memo' is a Python dict, used as the pickle's memo. It
may be mutated by dis(), if the pickle contains PUT or BINPUT opcodes.
Passing the same memo object to another dis() call then allows disassembly
to proceed across multiple pickles that were all created by the same
pickler with the same memo. Ordinarily you don't need to worry about this.
Optional arg indentlevel is the number of blanks by which to indent
a new MARK level. It defaults to 4.
In addition to printing the disassembly, some sanity checks are made:
+ All embedded opcode arguments "make sense".
+ Explicit and implicit pop operations have enough items on the stack.
+ When an opcode implicitly refers to a markobject, a markobject is
actually on the stack.
+ A memo entry isn't referenced before it's defined.
+ The markobject isn't stored in the memo.
+ A memo entry isn't redefined.
it s %5d:s %-4s %s%si R s (MARK at unknown opcode offset)s (MARK at %d)i s no MARK exists on stackR R R s memo key %r already defineds' stack is empty -- can't store into memos" can't store markobject in the memoR R R s& memo key %r has never been stored intoi
s3 tries to pop %d items from stack with only %d itemss highest protocol among opcodes =s stack not empty after STOP: %rN( R R R ( R R R (" t stackt memoRP t maxprotot markstackt indentlevelt indentchunkt errormsgR R R RX R t outt reprRW R R t linet maxR[ RY t beforeRZ t aftert numtopopt markmsgt
markobjectt popt markpost indexR R t extend( R R R R R R R RX R R R R R R R R R R ( ( R R L s !
0 5
s
>>> import pickle
>>> x = [1, 2, (3, 4), {'abc': u"def"}]
>>> pkl = pickle.dumps(x, 0)
>>> dis(pkl)
0: ( MARK
1: l LIST (MARK at 0)
2: p PUT 0
5: I INT 1
8: a APPEND
9: I INT 2
12: a APPEND
13: ( MARK
14: I INT 3
17: I INT 4
20: t TUPLE (MARK at 13)
21: p PUT 1
24: a APPEND
25: ( MARK
26: d DICT (MARK at 25)
27: p PUT 2
30: S STRING 'abc'
37: p PUT 3
40: V UNICODE u'def'
45: p PUT 4
48: s SETITEM
49: a APPEND
50: . STOP
highest protocol among opcodes = 0
Try again with a "binary" pickle.
>>> pkl = pickle.dumps(x, 1)
>>> dis(pkl)
0: ] EMPTY_LIST
1: q BINPUT 0
3: ( MARK
4: K BININT1 1
6: K BININT1 2
8: ( MARK
9: K BININT1 3
11: K BININT1 4
13: t TUPLE (MARK at 8)
14: q BINPUT 1
16: } EMPTY_DICT
17: q BINPUT 2
19: U SHORT_BINSTRING 'abc'
24: q BINPUT 3
26: X BINUNICODE u'def'
34: q BINPUT 4
36: s SETITEM
37: e APPENDS (MARK at 3)
38: . STOP
highest protocol among opcodes = 1
Exercise the INST/OBJ/BUILD family.
>>> import random
>>> dis(pickle.dumps(random.random, 0))
0: c GLOBAL 'random random'
15: p PUT 0
18: . STOP
highest protocol among opcodes = 0
>>> x = [pickle.PicklingError()] * 2
>>> dis(pickle.dumps(x, 0))
0: ( MARK
1: l LIST (MARK at 0)
2: p PUT 0
5: ( MARK
6: i INST 'pickle PicklingError' (MARK at 5)
28: p PUT 1
31: ( MARK
32: d DICT (MARK at 31)
33: p PUT 2
36: S STRING 'args'
44: p PUT 3
47: ( MARK
48: t TUPLE (MARK at 47)
49: s SETITEM
50: b BUILD
51: a APPEND
52: g GET 1
55: a APPEND
56: . STOP
highest protocol among opcodes = 0
>>> dis(pickle.dumps(x, 1))
0: ] EMPTY_LIST
1: q BINPUT 0
3: ( MARK
4: ( MARK
5: c GLOBAL 'pickle PicklingError'
27: q BINPUT 1
29: o OBJ (MARK at 4)
30: q BINPUT 2
32: } EMPTY_DICT
33: q BINPUT 3
35: U SHORT_BINSTRING 'args'
41: q BINPUT 4
43: ) EMPTY_TUPLE
44: s SETITEM
45: b BUILD
46: h BINGET 2
48: e APPENDS (MARK at 3)
49: . STOP
highest protocol among opcodes = 1
Try "the canonical" recursive-object test.
>>> L = []
>>> T = L,
>>> L.append(T)
>>> L[0] is T
True
>>> T[0] is L
True
>>> L[0][0] is L
True
>>> T[0][0] is T
True
>>> dis(pickle.dumps(L, 0))
0: ( MARK
1: l LIST (MARK at 0)
2: p PUT 0
5: ( MARK
6: g GET 0
9: t TUPLE (MARK at 5)
10: p PUT 1
13: a APPEND
14: . STOP
highest protocol among opcodes = 0
>>> dis(pickle.dumps(L, 1))
0: ] EMPTY_LIST
1: q BINPUT 0
3: ( MARK
4: h BINGET 0
6: t TUPLE (MARK at 3)
7: q BINPUT 1
9: a APPEND
10: . STOP
highest protocol among opcodes = 1
Note that, in the protocol 0 pickle of the recursive tuple, the disassembler
has to emulate the stack in order to realize that the POP opcode at 16 gets
rid of the MARK at 0.
>>> dis(pickle.dumps(T, 0))
0: ( MARK
1: ( MARK
2: l LIST (MARK at 1)
3: p PUT 0
6: ( MARK
7: g GET 0
10: t TUPLE (MARK at 6)
11: p PUT 1
14: a APPEND
15: 0 POP
16: 0 POP (MARK at 0)
17: g GET 1
20: . STOP
highest protocol among opcodes = 0
>>> dis(pickle.dumps(T, 1))
0: ( MARK
1: ] EMPTY_LIST
2: q BINPUT 0
4: ( MARK
5: h BINGET 0
7: t TUPLE (MARK at 4)
8: q BINPUT 1
10: a APPEND
11: 1 POP_MARK (MARK at 0)
12: h BINGET 1
14: . STOP
highest protocol among opcodes = 1
Try protocol 2.
>>> dis(pickle.dumps(L, 2))
0: \x80 PROTO 2
2: ] EMPTY_LIST
3: q BINPUT 0
5: h BINGET 0
7: \x85 TUPLE1
8: q BINPUT 1
10: a APPEND
11: . STOP
highest protocol among opcodes = 2
>>> dis(pickle.dumps(T, 2))
0: \x80 PROTO 2
2: ] EMPTY_LIST
3: q BINPUT 0
5: h BINGET 0
7: \x85 TUPLE1
8: q BINPUT 1
10: a APPEND
11: 0 POP
12: h BINGET 1
14: . STOP
highest protocol among opcodes = 2
sM
>>> import pickle
>>> from StringIO import StringIO
>>> f = StringIO()
>>> p = pickle.Pickler(f, 2)
>>> x = [1, 2, 3]
>>> p.dump(x)
>>> p.dump(x)
>>> f.seek(0)
>>> memo = {}
>>> dis(f, memo=memo)
0: \x80 PROTO 2
2: ] EMPTY_LIST
3: q BINPUT 0
5: ( MARK
6: K BININT1 1
8: K BININT1 2
10: K BININT1 3
12: e APPENDS (MARK at 5)
13: . STOP
highest protocol among opcodes = 2
>>> dis(f, memo=memo)
14: \x80 PROTO 2
16: h BINGET 0
18: . STOP
highest protocol among opcodes = 2
t disassembler_testt disassembler_memo_testc C s d k } | i S( N( t doctestt testmod( R ( ( R t _test s t __main__N(\ t __doc__R t
UP_TO_NEWLINEt TAKEN_FROM_ARGUMENT1t TAKEN_FROM_ARGUMENT4t objectR t structt unpackR R R R R R R R5 R! R" R$ R% R& R' R( R) R* R+ R- R. R/ R0 R9 R: R; R< R> R? R@ RA R RB RC RD RE RF RG R6 t pyintR8 t pylongRN t pyinteger_or_boolt pyboolR= t pyfloatRO t pystringR, t pyunicodet typeRP t pynoneRJ t pytupleRQ t pylistRR t pydictt anyobjectR RU RV R^ t opcodest name2it code2it enumerateR R R R RW R R# R R R t _dis_testt
_memo_testt __test__R R
(E R R. RE RC R R/ R R R R R R RA R RU R R R R- R R R R< R R R R R R RV R R! R R$ R" RG R R R R R? R0 R: R R R R; R% R> R R RB R9 R R R@ R R R R) R+ R' R^ RD RF R& R R* R( ( ( R t ? s "
1
;
3
''
& 9