0
Files
src/ios/PRESUBMIT.py
Federica Germinario 2ce51a841d [ios] Add presubmit check to limit use of NSUserDefaults
This CL adds a presubmit check trying to limit new usage of
NSUserDefaults.
Note that this check provides a WARNING message.

This check does not cover all scenarios and can produce some false
positives (in case the developer uses the pattern code to get
the content of an existing userDefault key).

Bug: none
Change-Id: Ie78b731ff5517a5c2d331494b2e76c66f630c6dd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6179973
Reviewed-by: Mark Cogan <marq@chromium.org>
Commit-Queue: Federica Germinario <fedegermi@google.com>
Cr-Commit-Position: refs/heads/main@{#1411606}
2025-01-27 06:44:10 -08:00

346 lines
13 KiB
Python

# Copyright 2017 The Chromium Authors
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.
"""Presubmit script for ios.
See http://dev.chromium.org/developers/how-tos/depottools/presubmit-scripts
for more details about the presubmit API built into depot_tools.
"""
import os
import xml.etree.ElementTree as ElementTree
NULLABILITY_PATTERN = r'(nonnull|nullable|_Nullable|_Nonnull)'
TODO_PATTERN = r'TO[D]O\(([^\)]*)\)'
BUG_PATTERN = r'^(crbug\.com|b)/\d+$'
DEPRECATED_BUG_PATTERN = r'^b/\d+$'
INCLUDE_PATTERN = r'^#include'
PIPE_IN_COMMENT_PATTERN = r'//.*[^|]\|(?!\|)'
IOS_PACKAGE_PATTERN = r'^ios'
BOXED_BOOL_PATTERN = r'@\((YES|NO)\)'
USER_DEFAULTS_PATTERN = r'\[NSUserDefaults standardUserDefaults]'
def IsSubListOf(needle, hay):
"""Returns whether there is a slice of |hay| equal to |needle|."""
for i, line in enumerate(hay):
if line == needle[0]:
if needle == hay[i:i + len(needle)]:
return True
return False
def _CheckNullabilityAnnotations(input_api, output_api):
""" Checks whether there are nullability annotations in ios code.
They are accepted in ios/web_view/public since it tries to mimic
the platform library but not anywhere else.
"""
nullability_regex = input_api.re.compile(NULLABILITY_PATTERN)
errors = []
for f in input_api.AffectedFiles():
if f.LocalPath().startswith('ios/web_view/public/'):
# ios/web_view/public tries to mimic an existing API that
# might have nullability in it and that is acceptable.
continue
for line_num, line in f.ChangedContents():
if nullability_regex.search(line):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
plural_suffix = '' if len(errors) == 1 else 's'
warning_message = ('Found Nullability annotation%(plural)s. '
'Prefer DCHECKs in ios code to check for nullness:' % {
'plural': plural_suffix
})
return [output_api.PresubmitPromptWarning(warning_message, items=errors)]
def _CheckBugInToDo(input_api, output_api):
""" Checks whether TODOs in ios code are identified by a bug number."""
errors = []
warnings = []
for f in input_api.AffectedFiles():
for line_num, line in f.ChangedContents():
if _HasToDoWithNoBug(input_api, line):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if _HasToDoWithDeprecatedBug(input_api, line):
warnings.append('%s:%s' % (f.LocalPath(), line_num))
if not errors and not warnings:
return []
output = []
if errors:
singular_article = 'a ' if len(errors) == 1 else ''
plural_suffix = '' if len(errors) == 1 else 's'
error_message = '\n'.join([
'Found TO'
'DO%(plural)s without %(a)sbug number%(plural)s (expected format '
'is \"TO'
'DO(crbug.com/######)\"):' % {
'plural': plural_suffix,
'a' : singular_article
}
] + errors) + '\n'
output.append(output_api.PresubmitError(error_message))
if warnings:
singular_article = 'a ' if len(warnings) == 1 else ''
plural_suffix = '' if len(warnings) == 1 else 's'
warning_message = '\n'.join([
'Found TO'
'DO%(plural)s with %(a)sdeprecated bug link%(plural)s (found '
'"b/#####\", expected format is \"crbug.com/######"):' % {
'plural': plural_suffix,
'a' : singular_article
}
] + warnings) + '\n'
output.append(output_api.PresubmitPromptWarning(warning_message))
return output
def _CheckHasNoIncludeDirectives(input_api, output_api):
""" Checks that #include preprocessor directives are not present."""
errors = []
for f in input_api.AffectedFiles():
if not _IsInIosPackage(input_api, f.LocalPath()):
continue
_, ext = os.path.splitext(f.LocalPath())
if ext != '.mm':
continue
for line_num, line in f.ChangedContents():
if _HasIncludeDirective(input_api, line):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
singular_plural = 'it' if len(errors) == 1 else 'them'
plural_suffix = '' if len(errors) == 1 else 's'
error_message = '\n'.join([
'Found usage of `#include` preprocessor directive%(plural)s! Please, '
'replace %(singular_plural)s with `#import` preprocessor '
'directive%(plural)s instead. '
'Consider replacing all existing `#include` with `#import` (if any) in '
'this file for the code clean up. See '
'https://chromium.googlesource.com/chromium/src.git/+/refs/heads/main'
'/styleguide/objective-c/objective-c.md'
'#import-and-include-in-the-directory for more details. '
'\n\nAffected file%(plural)s:' % {
'plural': plural_suffix,
'singular_plural': singular_plural
}
] + errors) + '\n'
return [output_api.PresubmitError(error_message)]
def _CheckHasNoPipeInComment(input_api, output_api):
""" Checks that comments don't contain pipes."""
pipe_regex = input_api.re.compile(PIPE_IN_COMMENT_PATTERN)
errors = []
for f in input_api.AffectedFiles():
if not _IsInIosPackage(input_api, f.LocalPath()):
continue
for line_num, line in f.ChangedContents():
if pipe_regex.search(line):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
warning_message = '\n'.join([
'Please use backticks "`" instead of pipes "|" if you need to quote'
' variable names and symbols in comments.\n'
'Found potential uses of pipes in:'
] + errors) + '\n'
return [output_api.PresubmitPromptWarning(warning_message)]
def _CheckCanImproveTestUsingExpectNSEQ(input_api, output_api):
""" Checks that test files use EXPECT_NSEQ when possible."""
errors = []
# Substrings that should not be used together with EXPECT_TRUE or
# EXPECT_FALSE in tests.
wrong_patterns = ["isEqualToString:", "isEqualToData:", "isEqualToArray:"]
for f in input_api.AffectedFiles():
if not '_unittest.' in f.LocalPath():
continue
for line_num, line in f.ChangedContents():
if line.startswith(("EXPECT_TRUE", "EXPECT_FALSE")):
# Condition is in one line.
if any(x in line for x in wrong_patterns):
errors.append('%s:%s' % (f.LocalPath(), line_num))
# Condition is split on multiple lines.
elif not line.endswith(";"):
# Check this is not the last line.
if line_num < len(f.NewContents()):
next_line = f.NewContents()[line_num]
if any(x in next_line for x in wrong_patterns):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
plural_suffix = '' if len(errors) == 1 else 's'
warning_message = '\n'.join([
'Found possible improvement in unittest. Prefer using'
' EXPECT_NSEQ() or EXPECT_NSNE() when possible.'
'\n\nAffected file%(plural)s:' % {
'plural': plural_suffix,
}
] + errors) + '\n'
return [output_api.PresubmitPromptWarning(warning_message)]
def _IsInIosPackage(input_api, path):
""" Returns True if path is within ios package"""
ios_package_regex = input_api.re.compile(IOS_PACKAGE_PATTERN)
return ios_package_regex.search(path)
def _HasIncludeDirective(input_api, line):
""" Returns True if #include is found in the line"""
include_regex = input_api.re.compile(INCLUDE_PATTERN)
return include_regex.search(line)
def _HasToDoWithNoBug(input_api, line):
""" Returns True if TODO is not identified by a bug number."""
todo_regex = input_api.re.compile(TODO_PATTERN)
bug_regex = input_api.re.compile(BUG_PATTERN)
todo_match = todo_regex.search(line)
if not todo_match:
return False
return not bug_regex.match(todo_match.group(1))
def _HasToDoWithDeprecatedBug(input_api, line):
""" Returns True if TODO is identified by a deprecated bug number format."""
todo_regex = input_api.re.compile(TODO_PATTERN)
deprecated_bug_regex = input_api.re.compile(DEPRECATED_BUG_PATTERN)
todo_match = todo_regex.search(line)
if not todo_match:
return False
return deprecated_bug_regex.match(todo_match.group(1))
def _CheckHasNoBoxedBOOL(input_api, output_api):
""" Checks that there are no @(YES) or @(NO)."""
boxed_BOOL_regex = input_api.re.compile(BOXED_BOOL_PATTERN)
errors = []
for f in input_api.AffectedFiles():
for line_num, line in f.ChangedContents():
if boxed_BOOL_regex.search(line):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
plural_suffix = '' if len(errors) == 1 else 's'
warning_message = ('Found boxed BOOL%(plural)s. '
'Prefer @YES or @NO in ios code:' % {
'plural': plural_suffix
})
return [output_api.PresubmitPromptWarning(warning_message, items=errors)]
def _CheckNoTearDownEGTest(input_api, output_api):
""" Checks that `- (void)tearDown {` is not present in an egtest.mm"""
errors = []
for f in input_api.AffectedFiles():
if not '_egtest.' in f.LocalPath():
continue
for line_num, line in f.ChangedContents():
if line.startswith("- (void)tearDown {"):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
warning_message = '\n'.join([
'To support hermetic EarlGrey test cases, tearDown has been renamed '
'to tearDownHelper, and will soon be removed. If tearDown is really '
'necessary for this test, please use addTeardownBlock'
] + errors) + '\n'
return [output_api.PresubmitError(warning_message)]
def _IsAlphabeticallySortedXML(file):
"""Check that the `file` is alphabetically sorted"""
parser = ElementTree.XMLParser(target=ElementTree.TreeBuilder(
insert_comments=True))
with open(file, 'r') as xml_file:
tree = ElementTree.parse(xml_file, parser)
root = tree.getroot()
original_tree_string = ElementTree.tostring(root, encoding='utf8')
messages_element = tree.findall('.//messages')[0]
messages = messages_element.findall('message')
messages.sort(key=lambda message: message.attrib["name"])
for message in messages:
messages_element.remove(message)
for message in messages:
messages_element.append(message)
ordered_tree_string = ElementTree.tostring(root, encoding='utf8')
return ordered_tree_string == original_tree_string
def _CheckOrderedStringFile(input_api, output_api):
""" Checks that the string files are alphabetically ordered"""
errors = []
for f in input_api.AffectedFiles():
if not f.LocalPath().endswith("_strings.grd"):
continue
if not _IsAlphabeticallySortedXML(f.AbsoluteLocalPath()):
errors.append(' python3 ios/tools/order_string_file.py ' +
f.LocalPath())
if not errors:
return []
warning_message = '\n'.join(
['Files not alphabetically sorted, try running:'] + errors) + '\n'
return [output_api.PresubmitPromptWarning(warning_message)]
def _CheckNotUsingNSUserDefaults(input_api, output_api):
""" Checks the added code to limit new usage of NSUserDefaults """
user_defaults_regex = input_api.re.compile(USER_DEFAULTS_PATTERN)
errors = []
for f in input_api.AffectedFiles():
if (not f.LocalPath().endswith('.mm')):
continue
for line_num, line in f.ChangedContents():
if user_defaults_regex.search(line):
errors.append('%s:%s' % (f.LocalPath(), line_num))
if not errors:
return []
warning_message = '\n'.join([
'A new use of NSUserDefaults was added. If this is a newly added key '
'consider storing it to PrefService instead.'
] + errors) + '\n'
return [output_api.PresubmitPromptWarning(warning_message)]
def CheckChangeOnUpload(input_api, output_api):
results = []
results.extend(_CheckBugInToDo(input_api, output_api))
results.extend(_CheckNullabilityAnnotations(input_api, output_api))
results.extend(_CheckHasNoIncludeDirectives(input_api, output_api))
results.extend(_CheckHasNoPipeInComment(input_api, output_api))
results.extend(_CheckHasNoBoxedBOOL(input_api, output_api))
results.extend(_CheckNoTearDownEGTest(input_api, output_api))
results.extend(_CheckCanImproveTestUsingExpectNSEQ(input_api, output_api))
results.extend(_CheckOrderedStringFile(input_api, output_api))
results.extend(_CheckNotUsingNSUserDefaults(input_api, output_api))
return results