mirror of
https://github.com/python/cpython.git
synced 2025-11-04 03:44:55 +00:00
bbaba85402
The new version (attached) is fast enough all the time in every real module
I have <whew!>. You can make it slow by, e.g., creating an open list with
5,000 90-character identifiers (+ trailing comma) each on its own line, then
adding an item to the end -- but that still consumes less than a second on
my P5-166. Response time in real code appears instantaneous.
Fixed some bugs.
New feature: when hitting ENTER and the cursor is beyond the line's leading
indentation, whitespace is removed on both sides of the cursor; before
whitespace was removed only on the left; e.g., assuming the cursor is
between the comma and the space:
def something(arg1, arg2):
^ cursor to the left of here, and hit ENTER
arg2): # new line used to end up here
arg2): # but now lines up the way you expect
New hack: AutoIndent has grown a context_use_ps1 Boolean config option,
defaulting to 0 (false) and set to 1 (only) by PyShell. Reason: handling
the fancy stuff requires looking backward for a parsing synch point; ps1
lines are the only sensible thing to look for in a shell window, but are a
bad thing to look for in a file window (ps1 lines show up in my module
docstrings often). PythonWin's shell should set this true too.
Persistent problem: strings containing def/class can still screw things up
completely. No improvement. Simplest workaround is on the user's head, and
consists of inserting e.g.
def _(): pass
(or any other def/class) after the end of the multiline string that's
screwing them up. This is especially irksome because IDLE's syntax coloring
is *not* confused, so when this happens the colors don't match the
indentation behavior they see.
500 lines
16 KiB
Python
500 lines
16 KiB
Python
import string
|
|
import re
|
|
import sys
|
|
|
|
# Reason last stmt is continued (or C_NONE if it's not).
|
|
C_NONE, C_BACKSLASH, C_STRING, C_BRACKET = range(4)
|
|
|
|
if 0: # for throwaway debugging output
|
|
def dump(*stuff):
|
|
sys.__stdout__.write(string.join(map(str, stuff), " ") + "\n")
|
|
|
|
# Find a def or class stmt.
|
|
|
|
_defclassre = re.compile(r"""
|
|
^
|
|
[ \t]*
|
|
(?:
|
|
def [ \t]+ [a-zA-Z_]\w* [ \t]* \(
|
|
| class [ \t]+ [a-zA-Z_]\w* [ \t]*
|
|
(?: \( .* \) )?
|
|
[ \t]* :
|
|
)
|
|
""", re.VERBOSE | re.MULTILINE).search
|
|
|
|
# Match blank line or non-indenting comment line.
|
|
|
|
_junkre = re.compile(r"""
|
|
[ \t]*
|
|
(?: \# \S .* )?
|
|
\n
|
|
""", re.VERBOSE).match
|
|
|
|
# Match any flavor of string; the terminating quote is optional
|
|
# so that we're robust in the face of incomplete program text.
|
|
|
|
_match_stringre = re.compile(r"""
|
|
\""" [^"\\]* (?:
|
|
(?: \\. | "(?!"") )
|
|
[^"\\]*
|
|
)*
|
|
(?: \""" )?
|
|
|
|
| " [^"\\\n]* (?: \\. [^"\\\n]* )* "?
|
|
|
|
| ''' [^'\\]* (?:
|
|
(?: \\. | '(?!'') )
|
|
[^'\\]*
|
|
)*
|
|
(?: ''' )?
|
|
|
|
| ' [^'\\\n]* (?: \\. [^'\\\n]* )* '?
|
|
""", re.VERBOSE | re.DOTALL).match
|
|
|
|
# Match a line that starts with something interesting;
|
|
# used to find the first item of a bracket structure.
|
|
|
|
_itemre = re.compile(r"""
|
|
[ \t]*
|
|
[^\s#\\] # if we match, m.end()-1 is the interesting char
|
|
""", re.VERBOSE).match
|
|
|
|
# Match start of stmts that should be followed by a dedent.
|
|
|
|
_closere = re.compile(r"""
|
|
\s*
|
|
(?: return
|
|
| break
|
|
| continue
|
|
| raise
|
|
| pass
|
|
)
|
|
\b
|
|
""", re.VERBOSE).match
|
|
|
|
# Chew up non-special chars as quickly as possible, but retaining
|
|
# enough info to determine the last non-ws char seen; if match is
|
|
# successful, and m.group(1) isn't None, m.end(1) less 1 is the
|
|
# index of the last non-ws char matched.
|
|
|
|
_chew_ordinaryre = re.compile(r"""
|
|
(?: \s+
|
|
| ( [^\s[\](){}#'"\\]+ )
|
|
)+
|
|
""", re.VERBOSE).match
|
|
|
|
# Build translation table to map uninteresting chars to "x", open
|
|
# brackets to "(", and close brackets to ")".
|
|
|
|
_tran = ['x'] * 256
|
|
for ch in "({[":
|
|
_tran[ord(ch)] = '('
|
|
for ch in ")}]":
|
|
_tran[ord(ch)] = ')'
|
|
for ch in "\"'\\\n#":
|
|
_tran[ord(ch)] = ch
|
|
_tran = string.join(_tran, '')
|
|
del ch
|
|
|
|
class Parser:
|
|
|
|
def __init__(self, indentwidth, tabwidth):
|
|
self.indentwidth = indentwidth
|
|
self.tabwidth = tabwidth
|
|
|
|
def set_str(self, str):
|
|
assert len(str) == 0 or str[-1] == '\n'
|
|
self.str = str
|
|
self.study_level = 0
|
|
|
|
# Return index of start of last (probable!) def or class stmt, or
|
|
# None if none found. It's only probable because we can't know
|
|
# whether we're in a string without reparsing from the start of
|
|
# the file -- and that's too slow in large files for routine use.
|
|
#
|
|
# Ack, hack: in the shell window this kills us, because there's
|
|
# no way to tell the differences between output, >>> etc and
|
|
# user input. Indeed, IDLE's first output line makes the rest
|
|
# look like it's in an unclosed paren!:
|
|
# Python 1.5.2 (#0, Apr 13 1999, ...
|
|
|
|
def find_last_def_or_class(self, use_ps1, _defclassre=_defclassre):
|
|
str, pos = self.str, None
|
|
if use_ps1:
|
|
# hack for shell window
|
|
ps1 = '\n' + sys.ps1
|
|
i = string.rfind(str, ps1)
|
|
if i >= 0:
|
|
pos = i + len(ps1)
|
|
self.str = str[:pos-1] + '\n' + str[pos:]
|
|
else:
|
|
i = 0
|
|
while 1:
|
|
m = _defclassre(str, i)
|
|
if m:
|
|
pos, i = m.span()
|
|
else:
|
|
break
|
|
return pos
|
|
|
|
# Throw away the start of the string. Intended to be called with
|
|
# find_last_def_or_class's result.
|
|
|
|
def set_lo(self, lo):
|
|
assert lo == 0 or self.str[lo-1] == '\n'
|
|
if lo > 0:
|
|
self.str = self.str[lo:]
|
|
|
|
# As quickly as humanly possible <wink>, find the line numbers (0-
|
|
# based) of the non-continuation lines.
|
|
# Creates self.{goodlines, continuation}.
|
|
|
|
def _study1(self, _replace=string.replace, _find=string.find):
|
|
if self.study_level >= 1:
|
|
return
|
|
self.study_level = 1
|
|
|
|
# Map all uninteresting characters to "x", all open brackets
|
|
# to "(", all close brackets to ")", then collapse runs of
|
|
# uninteresting characters. This can cut the number of chars
|
|
# by a factor of 10-40, and so greatly speed the following loop.
|
|
str = self.str
|
|
str = string.translate(str, _tran)
|
|
str = _replace(str, 'xxxxxxxx', 'x')
|
|
str = _replace(str, 'xxxx', 'x')
|
|
str = _replace(str, 'xx', 'x')
|
|
str = _replace(str, 'xx', 'x')
|
|
str = _replace(str, '\nx', '\n')
|
|
# note that replacing x\n with \n would be incorrect, because
|
|
# x may be preceded by a backslash
|
|
|
|
# March over the squashed version of the program, accumulating
|
|
# the line numbers of non-continued stmts, and determining
|
|
# whether & why the last stmt is a continuation.
|
|
continuation = C_NONE
|
|
level = lno = 0 # level is nesting level; lno is line number
|
|
self.goodlines = goodlines = [0]
|
|
push_good = goodlines.append
|
|
i, n = 0, len(str)
|
|
while i < n:
|
|
ch = str[i]
|
|
i = i+1
|
|
|
|
# cases are checked in decreasing order of frequency
|
|
if ch == 'x':
|
|
continue
|
|
|
|
if ch == '\n':
|
|
lno = lno + 1
|
|
if level == 0:
|
|
push_good(lno)
|
|
# else we're in an unclosed bracket structure
|
|
continue
|
|
|
|
if ch == '(':
|
|
level = level + 1
|
|
continue
|
|
|
|
if ch == ')':
|
|
if level:
|
|
level = level - 1
|
|
# else the program is invalid, but we can't complain
|
|
continue
|
|
|
|
if ch == '"' or ch == "'":
|
|
# consume the string
|
|
quote = ch
|
|
if str[i-1:i+2] == quote * 3:
|
|
quote = quote * 3
|
|
w = len(quote) - 1
|
|
i = i+w
|
|
while i < n:
|
|
ch = str[i]
|
|
i = i+1
|
|
|
|
if ch == 'x':
|
|
continue
|
|
|
|
if str[i-1:i+w] == quote:
|
|
i = i+w
|
|
break
|
|
|
|
if ch == '\n':
|
|
lno = lno + 1
|
|
if w == 0:
|
|
# unterminated single-quoted string
|
|
if level == 0:
|
|
push_good(lno)
|
|
break
|
|
continue
|
|
|
|
if ch == '\\':
|
|
assert i < n
|
|
if str[i] == '\n':
|
|
lno = lno + 1
|
|
i = i+1
|
|
continue
|
|
|
|
# else comment char or paren inside string
|
|
|
|
else:
|
|
# didn't break out of the loop, so we're still
|
|
# inside a string
|
|
continuation = C_STRING
|
|
continue # with outer loop
|
|
|
|
if ch == '#':
|
|
# consume the comment
|
|
i = _find(str, '\n', i)
|
|
assert i >= 0
|
|
continue
|
|
|
|
assert ch == '\\'
|
|
assert i < n
|
|
if str[i] == '\n':
|
|
lno = lno + 1
|
|
if i+1 == n:
|
|
continuation = C_BACKSLASH
|
|
i = i+1
|
|
|
|
# The last stmt may be continued for all 3 reasons.
|
|
# String continuation takes precedence over bracket
|
|
# continuation, which beats backslash continuation.
|
|
if continuation != C_STRING and level > 0:
|
|
continuation = C_BRACKET
|
|
self.continuation = continuation
|
|
|
|
# Push the final line number as a sentinel value, regardless of
|
|
# whether it's continued.
|
|
assert (continuation == C_NONE) == (goodlines[-1] == lno)
|
|
if goodlines[-1] != lno:
|
|
push_good(lno)
|
|
|
|
def get_continuation_type(self):
|
|
self._study1()
|
|
return self.continuation
|
|
|
|
# study1 was sufficient to determine the continuation status,
|
|
# but doing more requires looking at every character. study2
|
|
# does this for the last interesting statement in the block.
|
|
# Creates:
|
|
# self.stmt_start, stmt_end
|
|
# slice indices of last interesting stmt
|
|
# self.lastch
|
|
# last non-whitespace character before optional trailing
|
|
# comment
|
|
# self.lastopenbracketpos
|
|
# if continuation is C_BRACKET, index of last open bracket
|
|
|
|
def _study2(self, _rfind=string.rfind, _find=string.find,
|
|
_ws=string.whitespace):
|
|
if self.study_level >= 2:
|
|
return
|
|
self._study1()
|
|
self.study_level = 2
|
|
|
|
# Set p and q to slice indices of last interesting stmt.
|
|
str, goodlines = self.str, self.goodlines
|
|
i = len(goodlines) - 1
|
|
p = len(str) # index of newest line
|
|
while i:
|
|
assert p
|
|
# p is the index of the stmt at line number goodlines[i].
|
|
# Move p back to the stmt at line number goodlines[i-1].
|
|
q = p
|
|
for nothing in range(goodlines[i-1], goodlines[i]):
|
|
# tricky: sets p to 0 if no preceding newline
|
|
p = _rfind(str, '\n', 0, p-1) + 1
|
|
# The stmt str[p:q] isn't a continuation, but may be blank
|
|
# or a non-indenting comment line.
|
|
if _junkre(str, p):
|
|
i = i-1
|
|
else:
|
|
break
|
|
if i == 0:
|
|
# nothing but junk!
|
|
assert p == 0
|
|
q = p
|
|
self.stmt_start, self.stmt_end = p, q
|
|
|
|
# Analyze this stmt, to find the last open bracket (if any)
|
|
# and last interesting character (if any).
|
|
lastch = ""
|
|
stack = [] # stack of open bracket indices
|
|
push_stack = stack.append
|
|
while p < q:
|
|
# suck up all except ()[]{}'"#\\
|
|
m = _chew_ordinaryre(str, p, q)
|
|
if m:
|
|
i = m.end(1) - 1 # last non-ws (if any)
|
|
if i >= 0:
|
|
lastch = str[i]
|
|
p = m.end()
|
|
if p >= q:
|
|
break
|
|
|
|
ch = str[p]
|
|
|
|
if ch in "([{":
|
|
push_stack(p)
|
|
lastch = ch
|
|
p = p+1
|
|
continue
|
|
|
|
if ch in ")]}":
|
|
if stack:
|
|
del stack[-1]
|
|
lastch = ch
|
|
p = p+1
|
|
continue
|
|
|
|
if ch == '"' or ch == "'":
|
|
# consume string
|
|
# Note that study1 did this with a Python loop, but
|
|
# we use a regexp here; the reason is speed in both
|
|
# cases; the string may be huge, but study1 pre-squashed
|
|
# strings to a couple of characters per line. study1
|
|
# also needed to keep track of newlines, and we don't
|
|
# have to.
|
|
lastch = ch
|
|
p = _match_stringre(str, p, q).end()
|
|
continue
|
|
|
|
if ch == '#':
|
|
# consume comment and trailing newline
|
|
p = _find(str, '\n', p, q) + 1
|
|
assert p > 0
|
|
continue
|
|
|
|
assert ch == '\\'
|
|
p = p+1 # beyond backslash
|
|
assert p < q
|
|
if str[p] != '\n':
|
|
# the program is invalid, but can't complain
|
|
lastch = ch + str[p]
|
|
p = p+1 # beyond escaped char
|
|
|
|
# end while p < q:
|
|
|
|
self.lastch = lastch
|
|
if stack:
|
|
self.lastopenbracketpos = stack[-1]
|
|
|
|
# Assuming continuation is C_BRACKET, return the number
|
|
# of spaces the next line should be indented.
|
|
|
|
def compute_bracket_indent(self, _find=string.find):
|
|
self._study2()
|
|
assert self.continuation == C_BRACKET
|
|
j = self.lastopenbracketpos
|
|
str = self.str
|
|
n = len(str)
|
|
origi = i = string.rfind(str, '\n', 0, j) + 1
|
|
j = j+1 # one beyond open bracket
|
|
# find first list item; set i to start of its line
|
|
while j < n:
|
|
m = _itemre(str, j)
|
|
if m:
|
|
j = m.end() - 1 # index of first interesting char
|
|
extra = 0
|
|
break
|
|
else:
|
|
# this line is junk; advance to next line
|
|
i = j = _find(str, '\n', j) + 1
|
|
else:
|
|
# nothing interesting follows the bracket;
|
|
# reproduce the bracket line's indentation + a level
|
|
j = i = origi
|
|
while str[j] in " \t":
|
|
j = j+1
|
|
extra = self.indentwidth
|
|
return len(string.expandtabs(str[i:j],
|
|
self.tabwidth)) + extra
|
|
|
|
# Return number of physical lines in last stmt (whether or not
|
|
# it's an interesting stmt! this is intended to be called when
|
|
# continuation is C_BACKSLASH).
|
|
|
|
def get_num_lines_in_stmt(self):
|
|
self._study1()
|
|
goodlines = self.goodlines
|
|
return goodlines[-1] - goodlines[-2]
|
|
|
|
# Assuming continuation is C_BACKSLASH, return the number of spaces
|
|
# the next line should be indented. Also assuming the new line is
|
|
# the first one following the initial line of the stmt.
|
|
|
|
def compute_backslash_indent(self):
|
|
self._study2()
|
|
assert self.continuation == C_BACKSLASH
|
|
str = self.str
|
|
i = self.stmt_start
|
|
while str[i] in " \t":
|
|
i = i+1
|
|
startpos = i
|
|
|
|
# See whether the initial line starts an assignment stmt; i.e.,
|
|
# look for an = operator
|
|
endpos = string.find(str, '\n', startpos) + 1
|
|
found = level = 0
|
|
while i < endpos:
|
|
ch = str[i]
|
|
if ch in "([{":
|
|
level = level + 1
|
|
i = i+1
|
|
elif ch in ")]}":
|
|
if level:
|
|
level = level - 1
|
|
i = i+1
|
|
elif ch == '"' or ch == "'":
|
|
i = _match_stringre(str, i, endpos).end()
|
|
elif ch == '#':
|
|
break
|
|
elif level == 0 and ch == '=' and \
|
|
(i == 0 or str[i-1] not in "=<>!") and \
|
|
str[i+1] != '=':
|
|
found = 1
|
|
break
|
|
else:
|
|
i = i+1
|
|
|
|
if found:
|
|
# found a legit =, but it may be the last interesting
|
|
# thing on the line
|
|
i = i+1 # move beyond the =
|
|
found = re.match(r"\s*\\", str[i:endpos]) is None
|
|
|
|
if not found:
|
|
# oh well ... settle for moving beyond the first chunk
|
|
# of non-whitespace chars
|
|
i = startpos
|
|
while str[i] not in " \t\n":
|
|
i = i+1
|
|
|
|
return len(string.expandtabs(str[self.stmt_start :
|
|
i],
|
|
self.tabwidth)) + 1
|
|
|
|
# Return the leading whitespace on the initial line of the last
|
|
# interesting stmt.
|
|
|
|
def get_base_indent_string(self):
|
|
self._study2()
|
|
i, n = self.stmt_start, self.stmt_end
|
|
j = i
|
|
str = self.str
|
|
while j < n and str[j] in " \t":
|
|
j = j + 1
|
|
return str[i:j]
|
|
|
|
# Did the last interesting stmt open a block?
|
|
|
|
def is_block_opener(self):
|
|
self._study2()
|
|
return self.lastch == ':'
|
|
|
|
# Did the last interesting stmt close a block?
|
|
|
|
def is_block_closer(self):
|
|
self._study2()
|
|
return _closere(self.str, self.stmt_start) is not None
|