jinx/python-uncompyle6
1
Fork 0
mirror of https://github.com/rocky/python-uncompyle6.git synced 2025-08-03 08:49:51 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'python-3.0-to-3.2' into python-2.4-to-2.7

Browse Source
This commit is contained in:
rocky
2024-07-18 10:25:03 -04:00
parent d26a8163c0 5cf1315805
commit dc765d7bbb
2 changed files with 54 additions and 133 deletions
Download Patch File Download Diff File Expand all files Collapse all files

122
uncompyle6/scanners/scanner3.py
View File

@@ -286,6 +286,7 @@ class Scanner3(Scanner):
) )
return new_tokens return new_tokens
# Move to scanner35?
def bound_map_from_inst_35(self, insts, next_tokens, t, i): def bound_map_from_inst_35(self, insts, next_tokens, t, i):
""" """
Try to a sequence of instruction that ends with a BUILD_MAP into Try to a sequence of instruction that ends with a BUILD_MAP into
@@ -375,92 +376,6 @@ class Scanner3(Scanner):
) )
return new_tokens return new_tokens
def bound_map_from_inst_pre35(
self, insts, next_tokens, t, i
):
"""
Try to a sequence of instruction that ends with a BUILD_MAP into
a sequence that can be parsed much faster, but inserting the
token boundary at the beginning of the sequence.
"""
count = t.attr
assert isinstance(count, int)
# For small lists don't bother
if count < 10:
return None
# Older Python BUILD_MAP argument's count is a
# key and value pair and STORE_MAP. So it is multiplied by three.
collection_end = i + 1 + count * 3
for j in range(i + 1, collection_end, 3):
if insts[j].opname not in ("LOAD_CONST",):
return None
if insts[j + 1].opname not in ("LOAD_CONST",):
return None
if insts[j + 2].opname not in ("STORE_MAP",):
return None
collection_enum = CONST_COLLECTIONS.index("CONST_MAP")
new_tokens = next_tokens[:i]
start_offset = insts[i].offset
new_tokens.append(
Token(
opname="COLLECTION_START",
attr=collection_enum,
pattr="CONST_MAP",
offset="%s_0" % start_offset,
linestart=insts[i].starts_line,
has_arg=True,
has_extended_arg=False,
opc=self.opc,
optype="pseudo",
)
)
for j in range(i + 1, collection_end, 3):
new_tokens.append(
Token(
opname="ADD_KEY",
attr=insts[j + 1].argval,
pattr=insts[j + 1].argrepr,
offset=insts[j + 1].offset,
linestart=insts[j + 1].starts_line,
has_arg=True,
has_extended_arg=False,
opc=self.opc,
optype="pseudo",
)
)
new_tokens.append(
Token(
opname="ADD_VALUE",
attr=insts[j].argval,
pattr=insts[j].argrepr,
offset=insts[j].offset,
linestart=insts[j].starts_line,
has_arg=True,
has_extended_arg=False,
opc=self.opc,
optype="pseudo",
)
)
new_tokens.append(
Token(
opname="BUILD_DICT_OLDER",
attr=t.attr,
pattr=t.pattr,
offset=t.offset,
linestart=t.linestart,
has_arg=t.has_arg,
has_extended_arg=False,
opc=t.opc,
optype="pseudo",
)
)
return new_tokens
def ingest(self, co, classname=None, code_objects={}, show_asm=None): def ingest(self, co, classname=None, code_objects={}, show_asm=None):
""" """
Create "tokens" the bytecode of an Python code object. Largely these Create "tokens" the bytecode of an Python code object. Largely these
@@ -565,6 +480,7 @@ class Scanner3(Scanner):
last_op_was_break = False last_op_was_break = False
new_tokens = [] new_tokens = []
skip_end_offset = None skip_end_offset = None
for i, inst in enumerate(self.insts): for i, inst in enumerate(self.insts):
@@ -611,33 +527,17 @@ class Scanner3(Scanner):
elif opname in ("BUILD_MAP",): elif opname in ("BUILD_MAP",):
if self.version >= (3, 5): if self.version >= (3, 5):
bound_map_from_insts_fn = self.bound_map_from_inst_35 try_tokens = self.bound_map_from_inst_35(
else: self.insts,
bound_map_from_insts_fn = self.bound_map_from_inst_pre35 new_tokens,
try_tokens = bound_map_from_insts_fn( t,
self.insts, i,
new_tokens, )
t, if try_tokens is not None:
i,
)
if try_tokens is not None:
if self.version < (3, 5):
assert try_tokens[-1] == "BUILD_DICT_OLDER"
prev_offset = inst.offset
for j in range(i, len(self.insts)):
if self.insts[j].opname == "STORE_NAME":
new_tokens = try_tokens
skip_end_offset = prev_offset
# Set a hacky sentinal to indicate skipping to the
# next instruction
opname = "EXTENDED_ARG"
break
prev_offset = self.insts[j].offset
pass
pass
else:
new_tokens = try_tokens new_tokens = try_tokens
continue continue
pass
pass
pass pass
argval = inst.argval argval = inst.argval

65
uncompyle6/semantics/fragments.py
View File

@@ -1538,9 +1538,9 @@ class FragmentsWalker(pysource.SourceWalker, object):
self.write("(") self.write("(")
if kwargs: if kwargs:
# Last arg is tuple of keyword values: omit # Last arg is tuple of keyword values: omit
l = n - 1 m = n - 1
else: else:
l = n m = n
if kwargs: if kwargs:
# 3.6+ does this # 3.6+ does this
@@ -1552,7 +1552,7 @@ class FragmentsWalker(pysource.SourceWalker, object):
j += 1 j += 1
j = 0 j = 0
while i < l: while i < m:
self.write(sep) self.write(sep)
value = self.traverse(node[i]) value = self.traverse(node[i])
self.write("%s=%s" % (kwargs[j], value)) self.write("%s=%s" % (kwargs[j], value))
@@ -1560,7 +1560,7 @@ class FragmentsWalker(pysource.SourceWalker, object):
j += 1 j += 1
i += 1 i += 1
else: else:
while i < l: while i < m:
value = self.traverse(node[i]) value = self.traverse(node[i])
i += 1 i += 1
self.write(sep, value) self.write(sep, value)
@@ -1812,12 +1812,12 @@ class FragmentsWalker(pysource.SourceWalker, object):
def template_engine(self, entry, startnode): def template_engine(self, entry, startnode):
"""The format template interpretation engine. See the comment at the """The format template interpretation engine. See the comment at the
beginning of this module for the how we interpret format beginning of this module for how we interpret format
specifications such as %c, %C, and so on. specifications such as %c, %C, and so on.
""" """
# print("-----") # print("-----")
# print(startnode) # print(startnode.kind)
# print(entry[0]) # print(entry[0])
# print('======') # print('======')
@@ -1872,14 +1872,27 @@ class FragmentsWalker(pysource.SourceWalker, object):
index = entry[arg] index = entry[arg]
if isinstance(index, tuple): if isinstance(index, tuple):
assert ( if isinstance(index[1], str):
node[index[0]] == index[1] # if node[index[0]] != index[1]:
), "at %s[%d], expected %s node; got %s" % ( # from trepan.api import debug; debug()
node.kind, assert (
arg, node[index[0]] == index[1]
node[index[0]].kind, ), "at %s[%d], expected '%s' node; got '%s'" % (
index[1], node.kind,
) arg,
index[1],
node[index[0]].kind,
)
else:
assert (
node[index[0]] in index[1]
), "at %s[%d], expected to be in '%s' node; got '%s'" % (
node.kind,
arg,
index[1],
node[index[0]].kind,
)
index = index[0] index = index[0]
assert isinstance( assert isinstance(
index, int index, int
@@ -1899,14 +1912,21 @@ class FragmentsWalker(pysource.SourceWalker, object):
assert isinstance(tup, tuple) assert isinstance(tup, tuple)
if len(tup) == 3: if len(tup) == 3:
(index, nonterm_name, self.prec) = tup (index, nonterm_name, self.prec) = tup
assert ( if isinstance(tup[1], str):
node[index] == nonterm_name assert (
), "at %s[%d], expected '%s' node; got '%s'" % ( node[index] == nonterm_name
node.kind, ), "at %s[%d], expected '%s' node; got '%s'" % (
arg, node.kind,
nonterm_name, arg,
node[index].kind, nonterm_name,
) node[index].kind,
)
else:
assert node[tup[0]] in tup[1], (
f"at {node.kind}[{tup[0]}], expected to be in '{tup[1]}' "
f"node; got '{node[tup[0]].kind}'"
)
else: else:
assert len(tup) == 2 assert len(tup) == 2
(index, self.prec) = entry[arg] (index, self.prec) = entry[arg]
@@ -2117,6 +2137,7 @@ def code_deparse(
# Build Syntax Tree from tokenized and massaged disassembly. # Build Syntax Tree from tokenized and massaged disassembly.
# deparsed = pysource.FragmentsWalker(out, scanner, showast=showast) # deparsed = pysource.FragmentsWalker(out, scanner, showast=showast)
show_tree = debug_opts.get("tree", False) show_tree = debug_opts.get("tree", False)
linestarts = dict(scanner.opc.findlinestarts(co))
deparsed = walker( deparsed = walker(
version, version,
scanner, scanner,