[PATCH] Fix handling of symlinks pointing to a directory. BZ 969754

[Zdenek Pavlas]

1) make _os_path_walk() compatible with os.path.walk()

http://docs.python.org/2/library/os.path.html

Symbolic links to directories are not treated as subdirectories,
and that walk() therefore will not visit them. To visit linked
directories you must identify them with os.path.islink(file)
and os.path.isdir(file), and invoke walk() as necessary.

2) process each symlink only once. This breaks directory loops.
Since we may have processed the target dir already, we process
it at most twice.
---
createrepo/__init__.py | 9 +++++++++
1 file changed, 9 insertions(+)

diff --git a/createrepo/__init__.py b/createrepo/__init__.py
index 1b18a9f..8be48c5 100644
--- a/createrepo/__init__.py
+++ b/createrepo/__init__.py
@@ -298,17 +298,26 @@ class MetaDataGenerator:
for name in names:
name = os.path.join(top, name)
if os.path.isdir(name):
+ if os.path.islink(name):
+ continue
self._os_path_walk(name, func, arg)
def getFileList(self, directory, ext):
"""Return all files in path matching ext, store them in filelist,
recurse dirs. Returns a list object"""

extlen = len(ext)
+ seen = set()

def extension_visitor(filelist, dirname, names):
for fn in names:
fn = os.path.join(dirname, fn)
if os.path.isdir(fn):
+ if os.path.islink(fn):
+ st = os.stat(fn)
+ key = st.st_dev, st.st_ino
+ if key not in seen:
+ seen.add(key)
+ self._os_path_walk(fn, extension_visitor, filelist)
continue
if self.conf.skip_symlinks and os.path.islink(fn):
continue

[Zdenek Pavlas]

Post by Zdenek Pavlas
2) process each symlink only once. This breaks directory loops.
Since we may have processed the target dir already, we process
it at most twice.

Maybe we should not try to detect loops, but ANY means of processing
the same directory twice.. and err out instead:

--- a/createrepo/__init__.py
+++ b/createrepo/__init__.py
@@ -304,8 +304,15 @@ class MetaDataGenerator:
recurse dirs. Returns a list object"""

extlen = len(ext)
+ seen = set()

def extension_visitor(filelist, dirname, names):
+ st = os.stat(dirname)
+ key = st.st_dev, st.st_ino
+ if key in seen:
+ msg = _('Directory \'%s\' listed more than once')
+ raise MDError, msg % dirname[len(startdir):]
+ seen.add(key)
for fn in names:
fn = os.path.join(dirname, fn)
if os.path.isdir(fn):