possible bug in cherrypy.lib.autoreloader

[infidel]

I may have found the source of my infinite loop when importing kid
modules from my cherrypy server. Here is some code from the
autoreloader module of cherrypy:

def reloader_thread ():
mtimes = {}

def fileattr(m):
return getattr(m, "__file__", None)

while RUN_RELOADER:
for filename in map(fileattr, sys.modules.val ues()) +
reloadFiles:
if filename:
if filename.endswi th(".pyc"):
filename = filename[:-1]
try:
mtime = os.stat(filenam e).st_mtime
except OSError:
sys.exit(3) # force reload
if filename not in mtimes:
mtimes[filename] = mtime
continue
if mtime > mtimes[filename]:
sys.exit(3) # force reload
time.sleep(1)

So what happens if one of my modules is a kid template named
'login.kid'? kid compiles the template to login.pyc, which means the
os.stat call in this function will be passed 'login.py' which does not
exist.

[infidel]

Re: possible bug in cherrypy.lib.au toreload

Ok, so it turns out that the problem the cherrypy.lib.au toreload module
is having, is that kid imports elementtree and on my machine the
elementtree modules are inside a zip file (.egg). So the "path" to the
elementtree __init__.py file is not a valid OS path because everything
after the .egg file is inside the file.

Is there a quick way of determining that a module was imported from a
zip file, or that a module's __file__ attribute is a valid path as far
as Python is concerned, even though it doesn't exist to the OS?