aiden
/
depot_tools
mirror of https://chromium.googlesource.com/chromium/tools/depot_tools
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fix unicode upgrade of patch.

Browse Source 
When a diff contains utf8 character, the filename would be stored as a unicode
object instead of a str. This is even if the buffer is originally sent as a
utf-8 encoded str. Then once the diff is reconstructed to be sent to 'patch', a
unicode object instead of a str would be sent, confusing patch.

R=cmp@chromium.org
BUG=
TEST=


Review URL: http://codereview.chromium.org/9387024

git-svn-id: svn://svn.chromium.org/chrome/trunk/tools/depot_tools@122361 0039d316-1c4b-4281-b951-d872f2087c98
experimental/szager/collated-output
maruel@chromium.org 13 years ago
parent 98e6945846
commit 8fab6b66ff
3 changed files with 60 additions and 22 deletions
Show Stats Download Patch File Download Diff File

55
patch.py
Unescape Escape View File

@ -37,6 +37,15 @@ class FilePatchBase(object):
# Set when the file is copied or moved. # Set when the file is copied or moved.
self.source_filename = None self.source_filename = None
@property
def filename_utf8(self):
return self.filename.encode('utf-8')
@property
def source_filename_utf8(self):
if self.source_filename is not None:
return self.source_filename.encode('utf-8')
@staticmethod @staticmethod
def _process_filename(filename): def _process_filename(filename):
filename = filename.replace('\\', '/') filename = filename.replace('\\', '/')
@ -88,8 +97,8 @@ class FilePatchBase(object):
out += ' ' out += ' '
out += ' ' out += ' '
if self.source_filename: if self.source_filename:
out += '%s->' % self.source_filename out += '%s->' % self.source_filename_utf8
return out + str(self.filename) return out + self.filename_utf8
class FilePatchDelete(FilePatchBase): class FilePatchDelete(FilePatchBase):
@ -153,25 +162,27 @@ class FilePatchDiff(FilePatchBase):
# patch is stupid. It patches the source_filename instead so get rid of # patch is stupid. It patches the source_filename instead so get rid of
# any source_filename reference if needed. # any source_filename reference if needed.
return ( return (
self.diff_header.replace(self.source_filename, self.filename) + self.diff_header.replace(
self.source_filename_utf8, self.filename_utf8) +
self.diff_hunks) self.diff_hunks)
def set_relpath(self, relpath): def set_relpath(self, relpath):
old_filename = self.filename old_filename = self.filename_utf8
old_source_filename = self.source_filename or self.filename old_source_filename = self.source_filename_utf8 or self.filename_utf8
super(FilePatchDiff, self).set_relpath(relpath) super(FilePatchDiff, self).set_relpath(relpath)
# Update the header too. # Update the header too.
source_filename = self.source_filename or self.filename filename = self.filename_utf8
source_filename = self.source_filename_utf8 or self.filename_utf8
lines = self.diff_header.splitlines(True) lines = self.diff_header.splitlines(True)
for i, line in enumerate(lines): for i, line in enumerate(lines):
if line.startswith('diff --git'): if line.startswith('diff --git'):
lines[i] = line.replace( lines[i] = line.replace(
'a/' + old_source_filename, source_filename).replace( 'a/' + old_source_filename, source_filename).replace(
'b/' + old_filename, self.filename) 'b/' + old_filename, filename)
elif re.match(r'^\w+ from .+$', line) or line.startswith('---'): elif re.match(r'^\w+ from .+$', line) or line.startswith('---'):
lines[i] = line.replace(old_source_filename, source_filename) lines[i] = line.replace(old_source_filename, source_filename)
elif re.match(r'^\w+ to .+$', line) or line.startswith('+++'): elif re.match(r'^\w+ to .+$', line) or line.startswith('+++'):
lines[i] = line.replace(old_filename, self.filename) lines[i] = line.replace(old_filename, filename)
self.diff_header = ''.join(lines) self.diff_header = ''.join(lines)
def _split_header(self, diff): def _split_header(self, diff):
@ -197,7 +208,7 @@ class FilePatchDiff(FilePatchBase):
# Mangle any \\ in the header to /. # Mangle any \\ in the header to /.
header_lines = ('Index:', 'diff', 'copy', 'rename', '+++', '---') header_lines = ('Index:', 'diff', 'copy', 'rename', '+++', '---')
basename = os.path.basename(self.filename) basename = os.path.basename(self.filename_utf8)
for i in xrange(len(header)): for i in xrange(len(header)):
if (header[i].split(' ', 1)[0] in header_lines or if (header[i].split(' ', 1)[0] in header_lines or
header[i].endswith(basename)): header[i].endswith(basename)):
@ -314,7 +325,7 @@ class FilePatchDiff(FilePatchBase):
new = self.mangle(match.group(2)) new = self.mangle(match.group(2))
# The rename is about the new file so the old file can be anything. # The rename is about the new file so the old file can be anything.
if new not in (self.filename, 'dev/null'): if new not in (self.filename_utf8, 'dev/null'):
self._fail('Unexpected git diff output name %s.' % new) self._fail('Unexpected git diff output name %s.' % new)
if old == 'dev/null' and new == 'dev/null': if old == 'dev/null' and new == 'dev/null':
self._fail('Unexpected /dev/null git diff.') self._fail('Unexpected /dev/null git diff.')
@ -323,9 +334,9 @@ class FilePatchDiff(FilePatchBase):
if not old or not new: if not old or not new:
self._fail('Unexpected git diff; couldn\'t find git header.') self._fail('Unexpected git diff; couldn\'t find git header.')
if old not in (self.filename, 'dev/null'): if old not in (self.filename_utf8, 'dev/null'):
# Copy or rename. # Copy or rename.
self.source_filename = old self.source_filename = old.decode('utf-8')
self.is_new = True self.is_new = True
last_line = '' last_line = ''
@ -337,7 +348,7 @@ class FilePatchDiff(FilePatchBase):
# Cheap check to make sure the file name is at least mentioned in the # Cheap check to make sure the file name is at least mentioned in the
# 'diff' header. That the only remaining invariant. # 'diff' header. That the only remaining invariant.
if not self.filename in self.diff_header: if not self.filename_utf8 in self.diff_header:
self._fail('Diff seems corrupted.') self._fail('Diff seems corrupted.')
def _verify_git_header_process_line(self, lines, line, last_line): def _verify_git_header_process_line(self, lines, line, last_line):
@ -349,7 +360,7 @@ class FilePatchDiff(FilePatchBase):
http://www.kernel.org/pub/software/scm/git/docs/git-diff.html http://www.kernel.org/pub/software/scm/git/docs/git-diff.html
""" """
match = re.match(r'^(rename|copy) from (.+)$', line) match = re.match(r'^(rename|copy) from (.+)$', line)
old = self.source_filename or self.filename old = self.source_filename_utf8 or self.filename_utf8
if match: if match:
if old != match.group(2): if old != match.group(2):
self._fail('Unexpected git diff input name for line %s.' % line) self._fail('Unexpected git diff input name for line %s.' % line)
@ -361,7 +372,7 @@ class FilePatchDiff(FilePatchBase):
match = re.match(r'^(rename|copy) to (.+)$', line) match = re.match(r'^(rename|copy) to (.+)$', line)
if match: if match:
if self.filename != match.group(2): if self.filename_utf8 != match.group(2):
self._fail('Unexpected git diff output name for line %s.' % line) self._fail('Unexpected git diff output name for line %s.' % line)
if not last_line.startswith('%s from ' % match.group(1)): if not last_line.startswith('%s from ' % match.group(1)):
self._fail( self._fail(
@ -404,7 +415,7 @@ class FilePatchDiff(FilePatchBase):
self._fail('Unexpected git diff: --- not following +++.') self._fail('Unexpected git diff: --- not following +++.')
if '/dev/null' == match.group(1): if '/dev/null' == match.group(1):
self.is_delete = True self.is_delete = True
elif self.filename != self.mangle(match.group(1)): elif self.filename_utf8 != self.mangle(match.group(1)):
self._fail( self._fail(
'Unexpected git diff: %s != %s.' % (self.filename, match.group(1))) 'Unexpected git diff: %s != %s.' % (self.filename, match.group(1)))
if lines: if lines:
@ -429,7 +440,7 @@ class FilePatchDiff(FilePatchBase):
# Cheap check to make sure the file name is at least mentioned in the # Cheap check to make sure the file name is at least mentioned in the
# 'diff' header. That the only remaining invariant. # 'diff' header. That the only remaining invariant.
if not self.filename in self.diff_header: if not self.filename_utf8 in self.diff_header:
self._fail('Diff seems corrupted.') self._fail('Diff seems corrupted.')
def _verify_svn_header_process_line(self, lines, line, last_line): def _verify_svn_header_process_line(self, lines, line, last_line):
@ -443,9 +454,9 @@ class FilePatchDiff(FilePatchBase):
self._fail('--- and +++ are reversed') self._fail('--- and +++ are reversed')
if match.group(1) == '/dev/null': if match.group(1) == '/dev/null':
self.is_new = True self.is_new = True
elif self.mangle(match.group(1)) != self.filename: elif self.mangle(match.group(1)) != self.filename_utf8:
# guess the source filename. # guess the source filename.
self.source_filename = match.group(1) self.source_filename = match.group(1).decode('utf-8')
self.is_new = True self.is_new = True
if not lines or not lines[0].startswith('+++'): if not lines or not lines[0].startswith('+++'):
self._fail('Nothing after header.') self._fail('Nothing after header.')
@ -457,7 +468,7 @@ class FilePatchDiff(FilePatchBase):
self._fail('Unexpected diff: --- not following +++.') self._fail('Unexpected diff: --- not following +++.')
if match.group(1) == '/dev/null': if match.group(1) == '/dev/null':
self.is_delete = True self.is_delete = True
elif self.mangle(match.group(1)) != self.filename: elif self.mangle(match.group(1)) != self.filename_utf8:
self._fail('Unexpected diff: %s.' % match.group(1)) self._fail('Unexpected diff: %s.' % match.group(1))
if lines: if lines:
self._fail('Crap after +++') self._fail('Crap after +++')
@ -479,10 +490,10 @@ class PatchSet(object):
Deletes are last. Deletes are last.
""" """
if p.source_filename: if p.source_filename:
return (p.is_delete, p.source_filename, p.filename) return (p.is_delete, p.source_filename_utf8, p.filename_utf8)
else: else:
# tuple are always greater than string, abuse that fact. # tuple are always greater than string, abuse that fact.
return (p.is_delete, (p.filename,), p.filename) return (p.is_delete, (p.filename_utf8,), p.filename_utf8)
self.patches = sorted(patches, key=key) self.patches = sorted(patches, key=key)

13
testing_support/patches_data.py
Unescape Escape View File

@ -1,3 +1,4 @@
# coding: utf-8
# Copyright (c) 2012 The Chromium Authors. All rights reserved. # Copyright (c) 2012 The Chromium Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be # Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file. # found in the LICENSE file.
@ -122,6 +123,18 @@ class RAW(object):
'-{"traces": {"1t_proc": ["2.0", "0.0"], "1t_proc_ref": ["1.0", ...\n' '-{"traces": {"1t_proc": ["2.0", "0.0"], "1t_proc_ref": ["1.0", ...\n'
'+{"traces": {"1t_proc": ["2.0", "0.0"], "1t_proc_ref": ["1.0", ...\n') '+{"traces": {"1t_proc": ["2.0", "0.0"], "1t_proc_ref": ["1.0", ...\n')
RENAME_UTF8 = (
u'--- file_à\n'
u'+++ filé_b\n'
u'@@ -3,6 +3,7 @@ bb\n'
u' ccc\n'
u' ddé\n'
u' e\n'
u'+FÔÒ!\n'
u' ff\n'
u' ggg\n'
u' hh\n').encode('utf-8')
class GIT(object): class GIT(object):
"""Sample patches generated by git diff.""" """Sample patches generated by git diff."""

14
tests/patch_test.py
Unescape Escape View File

@ -1,4 +1,5 @@
#!/usr/bin/env python #!/usr/bin/env python
# coding: utf-8
# Copyright (c) 2012 The Chromium Authors. All rights reserved. # Copyright (c) 2012 The Chromium Authors. All rights reserved.
# Use of this source code is governed by a BSD-style license that can be # Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file. # found in the LICENSE file.
@ -275,6 +276,19 @@ class PatchTest(unittest.TestCase):
p, 'file_b', RAW.MINIMAL_RENAME, source_filename='file_a', is_new=True, p, 'file_b', RAW.MINIMAL_RENAME, source_filename='file_a', is_new=True,
nb_hunks=0) nb_hunks=0)
def testUnicodeFilenameGet(self):
p = patch.FilePatchDiff(u'filé_b', RAW.RENAME_UTF8, [])
self._check_patch(
p, u'filé_b', RAW.RENAME_UTF8, source_filename=u'file_à', is_new=True,
nb_hunks=1)
self.assertTrue(isinstance(p.get(False), str))
p.set_relpath('foo')
self.assertTrue(isinstance(p.get(False), str))
self.assertEquals(u'foo/file_à'.encode('utf-8'), p.source_filename_utf8)
self.assertEquals(u'foo/file_à', p.source_filename)
self.assertEquals(u'foo/filé_b'.encode('utf-8'), p.filename_utf8)
self.assertEquals(u'foo/filé_b', p.filename)
def testGitCopyPartial(self): def testGitCopyPartial(self):
p = patch.FilePatchDiff('wtf2', GIT.COPY_PARTIAL, []) p = patch.FilePatchDiff('wtf2', GIT.COPY_PARTIAL, [])
self._check_patch( self._check_patch(

Write Preview
Loading…
Cancel
Save