cascache: move list refs method to utils

This is more generic such that it can be used for listing files in other
paths.

Part #974

2 files changed, 32 insertions, 26 deletions

diff --git a/buildstream/_cas/cascache.py b/buildstream/_cas/cascache.py
index 5f67dc0c1..d7ab869b7 100644
--- a/buildstream/_cas/cascache.py
+++ b/buildstream/_cas/cascache.py
@@ -24,7 +24,6 @@ import stat
 import errno
 import uuid
 import contextlib
-from fnmatch import fnmatch
 
 import grpc
 
@@ -514,31 +513,8 @@ class CASCache():
     #
     def list_refs(self, *, glob=None):
         # string of: /path/to/repo/refs/heads
-        ref_heads = os.path.join(self.casdir, 'refs', 'heads')
-        path = ref_heads
-
-        if glob is not None:
-            globdir = os.path.dirname(glob)
-            if not any(c in "*?[" for c in globdir):
-                # path prefix contains no globbing characters so
-                # append the glob to optimise the os.walk()
-                path = os.path.join(ref_heads, globdir)
-
-        refs = []
-        mtimes = []
-
-        for root, _, files in os.walk(path):
-            for filename in files:
-                ref_path = os.path.join(root, filename)
-                relative_path = os.path.relpath(ref_path, ref_heads)  # Relative to refs head
-                if not glob or fnmatch(relative_path, glob):
-                    refs.append(relative_path)
-                    # Obtain the mtime (the time a file was last modified)
-                    mtimes.append(os.path.getmtime(ref_path))
-
-        # NOTE: Sorted will sort from earliest to latest, thus the
-        # first ref of this list will be the file modified earliest.
-        return [ref for _, ref in sorted(zip(mtimes, refs))]
+        return [ref for _, ref in sorted(list(utils._list_directory(
+            os.path.join(self.casdir, 'refs', 'heads'), glob_expr=glob)))]
 
     # list_objects():
     #
diff --git a/buildstream/utils.py b/buildstream/utils.py
index ade593750..7d6db0b67 100644
--- a/buildstream/utils.py
+++ b/buildstream/utils.py
@@ -23,6 +23,7 @@ Utilities
 
 import calendar
 import errno
+from fnmatch import fnmatch
 import hashlib
 import os
 import re
@@ -1291,3 +1292,32 @@ def _deterministic_umask():
         yield
     finally:
         os.umask(old_umask)
+
+
+# _list_directory()
+#
+# List files in a directory, given a base path
+#
+# Args:
+#    base_path (str): Base path to traverse over
+#    glob_expr (str|None): Optional glob expression to match against files
+#
+# Returns:
+#     (iter (mtime, filename)]): iterator of tuples of mtime and filenames
+#
+def _list_directory(base_path, *, glob_expr=None):
+    path = base_path
+    if glob_expr is not None:
+        globdir = os.path.dirname(glob_expr)
+        if not any(c in "*?[" for c in globdir):
+            # path prefix contains no globbing characters so
+            # append the glob to optimise the os.walk()
+            path = os.path.join(base_path, globdir)
+
+    for root, _, files in os.walk(path):
+        for filename in files:
+            ref_path = os.path.join(root, filename)
+            relative_path = os.path.relpath(ref_path, base_path)  # Relative to refs head
+            if not glob_expr or fnmatch(relative_path, glob_expr):
+                # Obtain the mtime (the time a file was last modified)
+                yield (os.path.getmtime(ref_path), relative_path)