Dump the cache file to a temporary file and rename it the expected

2008-12-08  Johan Dahlin  <johan@async.com.br>

	* giscanner/cachestore.py (CacheStore.store): Dump the cache
	file to a temporary file and rename it the expected filename
	only when it's completely written.
	This should make the cache more roboust when run in parallel,
	and hopefully avoid triggering bugs in Python.


svn path=/trunk/; revision=985

2 files changed, 27 insertions, 7 deletions

diff --git a/ChangeLog b/ChangeLog
index b5d6ba3c..ba52604a 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+2008-12-08  Johan Dahlin  <johan@async.com.br>
+
+	* giscanner/cachestore.py (CacheStore.store): Dump the cache
+	file to a temporary file and rename it the expected filename
+	only when it's completely written.
+	This should make the cache more roboust when run in parallel,
+	and hopefully avoid triggering bugs in Python.
+
 2008-12-03  Christophe Fergeau  <teuf@gnome.org>
 
 	Bug 562971 – g-ir-scanner failure on libgpod headers
diff --git a/giscanner/cachestore.py b/giscanner/cachestore.py
index 4629e622..d9163464 100644
--- a/giscanner/cachestore.py
+++ b/giscanner/cachestore.py
@@ -18,11 +18,11 @@
 # Boston, MA 02111-1307, USA.
 #
 
+import errno
 import cPickle
 import hashlib
 import os
-import errno
-
+import tempfile
 
 def _get_cachedir():
     cachedir = os.path.join(os.environ['HOME'], '.cache')
@@ -61,7 +61,7 @@ class CacheStore(object):
         return (os.stat(store_filename).st_mtime >=
                 os.stat(filename).st_mtime)
 
-    def _purge_cache(self, filename):
+    def _remove_filename(filename):
         try:
             os.unlink(filename)
         except IOError, e:
@@ -81,19 +81,31 @@ class CacheStore(object):
         store_filename = self._get_filename(filename)
         if store_filename is None:
             return
+
         if (os.path.exists(store_filename) and
             self._cache_is_valid(store_filename, filename)):
             return None
-        fd = open(store_filename, 'w')
+
+        tmp_fd, tmp_filename = tempfile.mkstemp(prefix='g-ir-scanner-cache-')
         try:
-            cPickle.dump(data, fd)
+            cPickle.dump(data, os.fdopen(tmp_fd, 'w'))
         except IOError, e:
             # No space left on device
-            if e.errno == e.ENOSPC:
+            if e.errno == errno.ENOSPC:
+                self._remove_filename(tmp_filename)
                 return
             else:
                 raise
 
+        try:
+            os.rename(tmp_filename, store_filename)
+        except OSError, e:
+            # Permission denied
+            if e.errno == errno.EACCES:
+                self._remove_filename(tmp_filename)
+            else:
+                raise
+
     def load(self, filename):
         store_filename = self._get_filename(filename)
         if store_filename is None:
@@ -111,6 +123,6 @@ class CacheStore(object):
             data = cPickle.load(fd)
         except (EOFError, ValueError, cPickle.BadPickleGet):
             # Broken cache entry, remove it
-            self._purge_cache(store_filename)
+            self._remove_filename(store_filename)
             data = None
         return data