Enable recursive globbing for clear_path

This commit enables recursive globbing in clear_path, allowing the
usage of '**' to match an arbitrary number of sub-directories.

Before this commit, clear_path used only non-recursive globbing. This
allowed to use '*' to expand names within one directory, e.g. '/a/*/c'
can expand to '/a/b/c', but not '/a/b/b/c'. With this commit, '/a/**/c'
can be used to expand to '/a/b/c', '/a/b/b/c', '/a/b/b/b/c' etc.

This is motivated by wanting to recursively delete all occurences of a
filename with the 'stage4/rm' entry of a spec file. The '/rm' entries
are processed with 'clear_path' in the existing code.

Additionally, 'glob.glob' is replaced with 'glob.iglob',
which returns the same files as 'glob.glob', but as an iterator
instead of as a list (so it no longer necessary to hold
all matches in memory at once).

Recursive globbing has been added in Python 3.5.

References:
https://docs.python.org/3/library/glob.html#glob.glob
https://docs.python.org/3/library/glob.html#glob.iglob

Signed-off-by: Felix Bier <felix.bier@rohde-schwarz.com>
Signed-off-by: Matt Turner <mattst88@gentoo.org>

1 files changed, 1 insertions, 1 deletions

diff --git a/catalyst/fileops.py b/catalyst/fileops.py
index 5c6f5cd8..59525420 100644
--- a/catalyst/fileops.py
+++ b/catalyst/fileops.py
@@ -99,7 +99,7 @@ def clear_dir(target, mode=0o755, remove=False,
 
 def clear_path(target_path):
     """Nuke |target_path| regardless of it being a dir, file or glob."""
-    targets = glob.glob(target_path)
+    targets = glob.iglob(target_path, recursive=True)
     for path in targets:
         clear_dir(path, remove=True)