blob: b3d3162f1d963efe234f5c5e99b6b79c5a9f0367 [file] [log] [blame]
#!/usr/bin/env python
#
# Copyright 2007 The Closure Linter Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS-IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Main class responsible for automatically fixing simple style violations."""
# Allow non-Google copyright
# pylint: disable=g-bad-file-header
__author__ = 'robbyw@google.com (Robert Walker)'
import re
import gflags as flags
from closure_linter import errors
from closure_linter import javascriptstatetracker
from closure_linter import javascripttokens
from closure_linter import requireprovidesorter
from closure_linter import tokenutil
from closure_linter.common import errorhandler
# Shorthand
Token = javascripttokens.JavaScriptToken
Type = javascripttokens.JavaScriptTokenType
END_OF_FLAG_TYPE = re.compile(r'(}?\s*)$')
# Regex to represent common mistake inverting author name and email as
# @author User Name (user@company)
INVERTED_AUTHOR_SPEC = re.compile(r'(?P<leading_whitespace>\s*)'
r'(?P<name>[^(]+)'
r'(?P<whitespace_after_name>\s+)'
r'\('
r'(?P<email>[^\s]+@[^)\s]+)'
r'\)'
r'(?P<trailing_characters>.*)')
FLAGS = flags.FLAGS
flags.DEFINE_boolean('disable_indentation_fixing', False,
'Whether to disable automatic fixing of indentation.')
class ErrorFixer(errorhandler.ErrorHandler):
"""Object that fixes simple style errors."""
def __init__(self, external_file=None):
"""Initialize the error fixer.
Args:
external_file: If included, all output will be directed to this file
instead of overwriting the files the errors are found in.
"""
errorhandler.ErrorHandler.__init__(self)
self._file_name = None
self._file_token = None
self._external_file = external_file
def HandleFile(self, filename, first_token):
"""Notifies this ErrorPrinter that subsequent errors are in filename.
Args:
filename: The name of the file about to be checked.
first_token: The first token in the file.
"""
self._file_name = filename
self._file_is_html = filename.endswith('.html') or filename.endswith('.htm')
self._file_token = first_token
self._file_fix_count = 0
self._file_changed_lines = set()
def _AddFix(self, tokens):
"""Adds the fix to the internal count.
Args:
tokens: The token or sequence of tokens changed to fix an error.
"""
self._file_fix_count += 1
if hasattr(tokens, 'line_number'):
self._file_changed_lines.add(tokens.line_number)
else:
for token in tokens:
self._file_changed_lines.add(token.line_number)
def HandleError(self, error):
"""Attempts to fix the error.
Args:
error: The error object
"""
code = error.code
token = error.token
if code == errors.JSDOC_PREFER_QUESTION_TO_PIPE_NULL:
iterator = token.attached_object.type_start_token
if iterator.type == Type.DOC_START_BRACE or iterator.string.isspace():
iterator = iterator.next
leading_space = len(iterator.string) - len(iterator.string.lstrip())
iterator.string = '%s?%s' % (' ' * leading_space,
iterator.string.lstrip())
# Cover the no outer brace case where the end token is part of the type.
while iterator and iterator != token.attached_object.type_end_token.next:
iterator.string = iterator.string.replace(
'null|', '').replace('|null', '')
iterator = iterator.next
# Create a new flag object with updated type info.
token.attached_object = javascriptstatetracker.JsDocFlag(token)
self._AddFix(token)
elif code == errors.JSDOC_MISSING_OPTIONAL_TYPE:
iterator = token.attached_object.type_end_token
if iterator.type == Type.DOC_END_BRACE or iterator.string.isspace():
iterator = iterator.previous
ending_space = len(iterator.string) - len(iterator.string.rstrip())
iterator.string = '%s=%s' % (iterator.string.rstrip(),
' ' * ending_space)
# Create a new flag object with updated type info.
token.attached_object = javascriptstatetracker.JsDocFlag(token)
self._AddFix(token)
elif code == errors.JSDOC_MISSING_VAR_ARGS_TYPE:
iterator = token.attached_object.type_start_token
if iterator.type == Type.DOC_START_BRACE or iterator.string.isspace():
iterator = iterator.next
starting_space = len(iterator.string) - len(iterator.string.lstrip())
iterator.string = '%s...%s' % (' ' * starting_space,
iterator.string.lstrip())
# Create a new flag object with updated type info.
token.attached_object = javascriptstatetracker.JsDocFlag(token)
self._AddFix(token)
elif code in (errors.MISSING_SEMICOLON_AFTER_FUNCTION,
errors.MISSING_SEMICOLON):
semicolon_token = Token(';', Type.SEMICOLON, token.line,
token.line_number)
tokenutil.InsertTokenAfter(semicolon_token, token)
token.metadata.is_implied_semicolon = False
semicolon_token.metadata.is_implied_semicolon = False
self._AddFix(token)
elif code in (errors.ILLEGAL_SEMICOLON_AFTER_FUNCTION,
errors.REDUNDANT_SEMICOLON,
errors.COMMA_AT_END_OF_LITERAL):
self._DeleteToken(token)
self._AddFix(token)
elif code == errors.INVALID_JSDOC_TAG:
if token.string == '@returns':
token.string = '@return'
self._AddFix(token)
elif code == errors.FILE_MISSING_NEWLINE:
# This error is fixed implicitly by the way we restore the file
self._AddFix(token)
elif code == errors.MISSING_SPACE:
if error.fix_data:
token.string = error.fix_data
self._AddFix(token)
elif error.position:
if error.position.IsAtBeginning():
tokenutil.InsertSpaceTokenAfter(token.previous)
elif error.position.IsAtEnd(token.string):
tokenutil.InsertSpaceTokenAfter(token)
else:
token.string = error.position.Set(token.string, ' ')
self._AddFix(token)
elif code == errors.EXTRA_SPACE:
if error.position:
token.string = error.position.Set(token.string, '')
self._AddFix(token)
elif code == errors.MISSING_LINE:
if error.position.IsAtBeginning():
tokenutil.InsertBlankLineAfter(token.previous)
else:
tokenutil.InsertBlankLineAfter(token)
self._AddFix(token)
elif code == errors.EXTRA_LINE:
self._DeleteToken(token)
self._AddFix(token)
elif code == errors.WRONG_BLANK_LINE_COUNT:
if not token.previous:
# TODO(user): Add an insertBefore method to tokenutil.
return
num_lines = error.fix_data
should_delete = False
if num_lines < 0:
num_lines *= -1
should_delete = True
for unused_i in xrange(1, num_lines + 1):
if should_delete:
# TODO(user): DeleteToken should update line numbers.
self._DeleteToken(token.previous)
else:
tokenutil.InsertBlankLineAfter(token.previous)
self._AddFix(token)
elif code == errors.UNNECESSARY_DOUBLE_QUOTED_STRING:
end_quote = tokenutil.Search(token, Type.DOUBLE_QUOTE_STRING_END)
if end_quote:
single_quote_start = Token(
"'", Type.SINGLE_QUOTE_STRING_START, token.line, token.line_number)
single_quote_end = Token(
"'", Type.SINGLE_QUOTE_STRING_START, end_quote.line,
token.line_number)
tokenutil.InsertTokenAfter(single_quote_start, token)
tokenutil.InsertTokenAfter(single_quote_end, end_quote)
self._DeleteToken(token)
self._DeleteToken(end_quote)
self._AddFix([token, end_quote])
elif code == errors.MISSING_BRACES_AROUND_TYPE:
fixed_tokens = []
start_token = token.attached_object.type_start_token
if start_token.type != Type.DOC_START_BRACE:
leading_space = (
len(start_token.string) - len(start_token.string.lstrip()))
if leading_space:
start_token = tokenutil.SplitToken(start_token, leading_space)
# Fix case where start and end token were the same.
if token.attached_object.type_end_token == start_token.previous:
token.attached_object.type_end_token = start_token
new_token = Token('{', Type.DOC_START_BRACE, start_token.line,
start_token.line_number)
tokenutil.InsertTokenAfter(new_token, start_token.previous)
token.attached_object.type_start_token = new_token
fixed_tokens.append(new_token)
end_token = token.attached_object.type_end_token
if end_token.type != Type.DOC_END_BRACE:
# If the start token was a brace, the end token will be a
# FLAG_ENDING_TYPE token, if there wasn't a starting brace then
# the end token is the last token of the actual type.
last_type = end_token
if not fixed_tokens:
last_type = end_token.previous
while last_type.string.isspace():
last_type = last_type.previous
# If there was no starting brace then a lone end brace wouldn't have
# been type end token. Now that we've added any missing start brace,
# see if the last effective type token was an end brace.
if last_type.type != Type.DOC_END_BRACE:
trailing_space = (len(last_type.string) -
len(last_type.string.rstrip()))
if trailing_space:
tokenutil.SplitToken(last_type,
len(last_type.string) - trailing_space)
new_token = Token('}', Type.DOC_END_BRACE, last_type.line,
last_type.line_number)
tokenutil.InsertTokenAfter(new_token, last_type)
token.attached_object.type_end_token = new_token
fixed_tokens.append(new_token)
self._AddFix(fixed_tokens)
elif code == errors.GOOG_REQUIRES_NOT_ALPHABETIZED:
require_start_token = error.fix_data
sorter = requireprovidesorter.RequireProvideSorter()
sorter.FixRequires(require_start_token)
self._AddFix(require_start_token)
elif code == errors.GOOG_PROVIDES_NOT_ALPHABETIZED:
provide_start_token = error.fix_data
sorter = requireprovidesorter.RequireProvideSorter()
sorter.FixProvides(provide_start_token)
self._AddFix(provide_start_token)
elif code == errors.UNNECESSARY_BRACES_AROUND_INHERIT_DOC:
if token.previous.string == '{' and token.next.string == '}':
self._DeleteToken(token.previous)
self._DeleteToken(token.next)
self._AddFix([token])
elif code == errors.INVALID_AUTHOR_TAG_DESCRIPTION:
match = INVERTED_AUTHOR_SPEC.match(token.string)
if match:
token.string = '%s%s%s(%s)%s' % (match.group('leading_whitespace'),
match.group('email'),
match.group('whitespace_after_name'),
match.group('name'),
match.group('trailing_characters'))
self._AddFix(token)
elif (code == errors.WRONG_INDENTATION and
not FLAGS.disable_indentation_fixing):
token = tokenutil.GetFirstTokenInSameLine(token)
actual = error.position.start
expected = error.position.length
# Cases where first token is param but with leading spaces.
if (len(token.string.lstrip()) == len(token.string) - actual and
token.string.lstrip()):
token.string = token.string.lstrip()
actual = 0
if token.type in (Type.WHITESPACE, Type.PARAMETERS) and actual != 0:
token.string = token.string.lstrip() + (' ' * expected)
self._AddFix([token])
else:
# We need to add indentation.
new_token = Token(' ' * expected, Type.WHITESPACE,
token.line, token.line_number)
# Note that we'll never need to add indentation at the first line,
# since it will always not be indented. Therefore it's safe to assume
# token.previous exists.
tokenutil.InsertTokenAfter(new_token, token.previous)
self._AddFix([token])
elif code in [errors.MALFORMED_END_OF_SCOPE_COMMENT,
errors.MISSING_END_OF_SCOPE_COMMENT]:
# Only fix cases where }); is found with no trailing content on the line
# other than a comment. Value of 'token' is set to } for this error.
if (token.type == Type.END_BLOCK and
token.next.type == Type.END_PAREN and
token.next.next.type == Type.SEMICOLON):
current_token = token.next.next.next
removed_tokens = []
while current_token and current_token.line_number == token.line_number:
if current_token.IsAnyType(Type.WHITESPACE,
Type.START_SINGLE_LINE_COMMENT,
Type.COMMENT):
removed_tokens.append(current_token)
current_token = current_token.next
else:
return
if removed_tokens:
self._DeleteTokens(removed_tokens[0], len(removed_tokens))
whitespace_token = Token(' ', Type.WHITESPACE, token.line,
token.line_number)
start_comment_token = Token('//', Type.START_SINGLE_LINE_COMMENT,
token.line, token.line_number)
comment_token = Token(' goog.scope', Type.COMMENT, token.line,
token.line_number)
insertion_tokens = [whitespace_token, start_comment_token,
comment_token]
tokenutil.InsertTokensAfter(insertion_tokens, token.next.next)
self._AddFix(removed_tokens + insertion_tokens)
elif code in [errors.EXTRA_GOOG_PROVIDE, errors.EXTRA_GOOG_REQUIRE]:
tokens_in_line = tokenutil.GetAllTokensInSameLine(token)
self._DeleteTokens(tokens_in_line[0], len(tokens_in_line))
self._AddFix(tokens_in_line)
elif code in [errors.MISSING_GOOG_PROVIDE, errors.MISSING_GOOG_REQUIRE]:
is_provide = code == errors.MISSING_GOOG_PROVIDE
is_require = code == errors.MISSING_GOOG_REQUIRE
missing_namespaces = error.fix_data[0]
need_blank_line = error.fix_data[1]
if need_blank_line is None:
# TODO(user): This happens when there are no existing
# goog.provide or goog.require statements to position new statements
# relative to. Consider handling this case with a heuristic.
return
insert_location = token.previous
# If inserting a missing require with no existing requires, insert a
# blank line first.
if need_blank_line and is_require:
tokenutil.InsertBlankLineAfter(insert_location)
insert_location = insert_location.next
for missing_namespace in missing_namespaces:
new_tokens = self._GetNewRequireOrProvideTokens(
is_provide, missing_namespace, insert_location.line_number + 1)
tokenutil.InsertLineAfter(insert_location, new_tokens)
insert_location = new_tokens[-1]
self._AddFix(new_tokens)
# If inserting a missing provide with no existing provides, insert a
# blank line after.
if need_blank_line and is_provide:
tokenutil.InsertBlankLineAfter(insert_location)
def _GetNewRequireOrProvideTokens(self, is_provide, namespace, line_number):
"""Returns a list of tokens to create a goog.require/provide statement.
Args:
is_provide: True if getting tokens for a provide, False for require.
namespace: The required or provided namespaces to get tokens for.
line_number: The line number the new require or provide statement will be
on.
Returns:
Tokens to create a new goog.require or goog.provide statement.
"""
string = 'goog.require'
if is_provide:
string = 'goog.provide'
line_text = string + '(\'' + namespace + '\');\n'
return [
Token(string, Type.IDENTIFIER, line_text, line_number),
Token('(', Type.START_PAREN, line_text, line_number),
Token('\'', Type.SINGLE_QUOTE_STRING_START, line_text, line_number),
Token(namespace, Type.STRING_TEXT, line_text, line_number),
Token('\'', Type.SINGLE_QUOTE_STRING_END, line_text, line_number),
Token(')', Type.END_PAREN, line_text, line_number),
Token(';', Type.SEMICOLON, line_text, line_number)
]
def _DeleteToken(self, token):
"""Deletes the specified token from the linked list of tokens.
Updates instance variables pointing to tokens such as _file_token if
they reference the deleted token.
Args:
token: The token to delete.
"""
if token == self._file_token:
self._file_token = token.next
tokenutil.DeleteToken(token)
def _DeleteTokens(self, token, token_count):
"""Deletes the given number of tokens starting with the given token.
Updates instance variables pointing to tokens such as _file_token if
they reference the deleted token.
Args:
token: The first token to delete.
token_count: The total number of tokens to delete.
"""
if token == self._file_token:
for unused_i in xrange(token_count):
self._file_token = self._file_token.next
tokenutil.DeleteTokens(token, token_count)
def FinishFile(self):
"""Called when the current file has finished style checking.
Used to go back and fix any errors in the file. It currently supports both
js and html files. For js files it does a simple dump of all tokens, but in
order to support html file, we need to merge the original file with the new
token set back together. This works because the tokenized html file is the
original html file with all non js lines kept but blanked out with one blank
line token per line of html.
"""
if self._file_fix_count:
# Get the original file content for html.
if self._file_is_html:
f = open(self._file_name, 'r')
original_lines = f.readlines()
f.close()
f = self._external_file
if not f:
error_noun = 'error' if self._file_fix_count == 1 else 'errors'
print 'Fixed %d %s in %s' % (
self._file_fix_count, error_noun, self._file_name)
f = open(self._file_name, 'w')
token = self._file_token
# Finding the first not deleted token.
while token.is_deleted:
token = token.next
# If something got inserted before first token (e.g. due to sorting)
# then move to start. Bug 8398202.
while token.previous:
token = token.previous
char_count = 0
line = ''
while token:
line += token.string
char_count += len(token.string)
if token.IsLastInLine():
# We distinguish if a blank line in html was from stripped original
# file or newly added error fix by looking at the "org_line_number"
# field on the token. It is only set in the tokenizer, so for all
# error fixes, the value should be None.
if (line or not self._file_is_html or
token.orig_line_number is None):
f.write(line)
f.write('\n')
else:
f.write(original_lines[token.orig_line_number - 1])
line = ''
if char_count > 80 and token.line_number in self._file_changed_lines:
print 'WARNING: Line %d of %s is now longer than 80 characters.' % (
token.line_number, self._file_name)
char_count = 0
token = token.next
if not self._external_file:
# Close the file if we created it
f.close()