Fix long-standing bugs with MANIFEST.in parsing on Windows (#6884).

These regex changes fix a number of issues for distutils on Windows:
- #6884: impossible to include a file starting with 'build'
- #9691 and #14004: sdist includes too many files
- #13193: test_filelist failures

This commit replaces the incorrect changes done in 0a94e2f807c7 and
90b30d62caf2 to fix #13193; we were too eager to fix the test failures
and I did not study the code enough before greenlighting patches.  This
time we have unit tests from the problems reported by users to be sure
we have the right fix.

Thanks to Nadeem Vawda for his help.
This commit is contained in:
Éric Araujo 2012-02-25 16:28:05 +01:00
parent 3d4809f464
commit 2e0a0e1640
4 changed files with 133 additions and 47 deletions

View file

@ -201,6 +201,7 @@ class FileList:
Return True if files are found, False otherwise.
"""
# XXX docstring lying about what the special chars are?
files_found = False
pattern_re = translate_pattern(pattern, anchor, prefix, is_regex)
self.debug_print("include_pattern: applying regex r'%s'" %
@ -284,11 +285,14 @@ def glob_to_re(pattern):
# IMHO is wrong -- '?' and '*' aren't supposed to match slash in Unix,
# and by extension they shouldn't match such "special characters" under
# any OS. So change all non-escaped dots in the RE to match any
# character except the special characters.
# XXX currently the "special characters" are just slash -- i.e. this is
# Unix-only.
pattern_re = re.sub(r'((?<!\\)(\\\\)*)\.', r'\1[^/]', pattern_re)
# character except the special characters (currently: just os.sep).
sep = os.sep
if os.sep == '\\':
# we're using a regex to manipulate a regex, so we need
# to escape the backslash twice
sep = r'\\\\'
escaped = r'\1[^%s]' % sep
pattern_re = re.sub(r'((?<!\\)(\\\\)*)\.', escaped, pattern_re)
return pattern_re
@ -312,9 +316,11 @@ def translate_pattern(pattern, anchor=1, prefix=None, is_regex=0):
if prefix is not None:
# ditch end of pattern character
empty_pattern = glob_to_re('')
prefix_re = (glob_to_re(prefix))[:-len(empty_pattern)]
# paths should always use / in manifest templates
pattern_re = "^%s/.*%s" % (prefix_re, pattern_re)
prefix_re = glob_to_re(prefix)[:-len(empty_pattern)]
sep = os.sep
if os.sep == '\\':
sep = r'\\'
pattern_re = "^" + sep.join((prefix_re, ".*" + pattern_re))
else: # no prefix -- respect anchor flag
if anchor:
pattern_re = "^" + pattern_re