listing files with utf8 filenames raises UnicodeDecodeError with certain characters

[shockwave107@…]

How to Reproduce

1. Requires: filesystem with utf8 filename support
2. create mercurial repository
3. create file with utf8 coded character "ö"

(["ö"="0xc3 0xb6"] or others with utf8 byte sequence containing "bytes out of range(128)" )

4. add/commitpush, and navigate with repository browser to containing path

same procedure with subversion does work.

Original Query and Trace

While doing a GET operation on /browser/WuS/uebung, Trac issued an internal error.

Contained File

WuS/uebung/Lös4.doc

Request parameters:

{'path': u'/WuS/uebung'}

User Agent was: Mozilla/5.0 (compatible; Konqueror/3.5; Linux) KHTML/3.5.9 (like Gecko) (Gentoo)

$&

Trac	0.11.1
Python	2.5.2 (r252, Oct 27 2008, 23:33:32) [GCC 4.1.2 (Gentoo 4.1.2 p1.1)]
setuptools	0.6c8
psycopg2	2.0.2
Genshi	0.5.1
mod_python	3.3.1
Pygments	0.10
Mercurial	1.0.2
jQuery:	1.2.6 $&

Traceback (most recent call last):
  File "/usr/lib64/python2.5/site-packages/trac/web/main.py", line 423, in _dispatch_request
    dispatcher.dispatch(req)
  File "/usr/lib64/python2.5/site-packages/trac/web/main.py", line 197, in dispatch
    resp = chosen_handler.process_request(req)
  File "/usr/lib64/python2.5/site-packages/trac/versioncontrol/web_ui/browser.py", line 361, in process_request
    'dir': node.isdir and self._render_dir(req, repos, node, rev),
  File "/usr/lib64/python2.5/site-packages/trac/versioncontrol/web_ui/browser.py", line 407, in _render_dir
    entries = [entry(n) for n in node.get_entries()]
  File "build/bdist.linux-x86_64/egg/tracext/hg/backend.py", line 652, in get_entries
    entry = posixpath.join(self.path, entry)
  File "/usr/lib64/python2.5/posixpath.py", line 65, in join
    path += '/' + b
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 2: ordinal not in range(128)

Additional

TracMercurial from svn

http://svn.edgewall.com/repos/trac/sandbox/mercurial-plugin-0.11 
Revision: 7656

PS: same with Trac 0.11.0/0.10.5 and TracMercurial 0.10 svn version

[Christian Boos]

#8018 closed as duplicate.

Non-ascii file names are only one part of the problem. See also #7160.

[Joseph Tate <jtate+trac@…>]

This still happens in 0.12 svn:

Most recent call last:
File "/usr/lib/python2.4/site-packages/Trac-0.12multirepos_r0-py2.4.egg/trac/web/main.py", line 467, in _dispatch_request
  dispatcher.dispatch(req)
File "/usr/lib/python2.4/site-packages/Trac-0.12multirepos_r0-py2.4.egg/trac/web/main.py", line 212, in dispatch
  resp = chosen_handler.process_request(req)
File "/usr/lib/python2.4/site-packages/Trac-0.12multirepos_r0-py2.4.egg/trac/versioncontrol/web_ui/browser.py", line 376, in process_request
  dir_data = self._render_dir(
File "/usr/lib/python2.4/site-packages/Trac-0.12multirepos_r0-py2.4.egg/trac/versioncontrol/web_ui/browser.py", line 493, in _render_dir
  entries = [entry(n) for n in node.get_entries()]
File "/home/admin/trac_stuff/mercurial-plugin-0.12/tracext/hg/backend.py", line 754, in get_entries
  dirnodes = self.findnode((n_rev > node_rev) and node_rev or n_rev, dirnames)
File "/home/admin/trac_stuff/mercurial-plugin-0.12/tracext/hg/backend.py", line 661, in findnode
  if f.startswith(d):

d=u'assets/help/' f='v_images/affiliate_0/affiliate_logos/BJ\x92s_Wholesale_Club_logo.gif'

Notice that f is a str. I tried wrapping that and a few other things in to_unicode(), but it ended up being a rat hole. I'm not sure where that wrapping needs to occur, but it appears to be partially applied at the moment.

[Igor Dejanović <igor.dejanovic@…>]

Same issue here. Putting

import sys
sys.setdefaultencoding('utf-8')

in sitecustomize.py seems to fix it.

[Christian Boos]

See #8538, which has the beginning of a patch…