Skip to content

Commit

Permalink
Merge pull request #389 from gratipay/refactor-dispatch
Browse files Browse the repository at this point in the history 
Refactor dispatch
  • Loading branch information
@pjz
pjz committed Oct 3, 2014
2 parents 7dba20a + c24da5f commit 78ec9a3
Show file tree
Hide file tree
Showing 13 changed files with 288 additions and 174 deletions.
56 changes: 38 additions & 18 deletions aspen/algorithms/website.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,7 @@
from aspen.http.response import Response
from aspen import typecasting
from first import first as _first
from aspen.dispatcher import DispatchResult, DispatchStatus


def parse_environ_into_request(environ):
Expand All @@ -61,20 +62,39 @@ def raise_200_for_OPTIONS(request):


def dispatch_request_to_filesystem(website, request):
dispatcher.dispatch(website, request)

if website.list_directories:
directory_default = website.ours_or_theirs('autoindex.html.spt')
assert directory_default is not None # sanity check
else:
directory_default = None

result = dispatcher.dispatch( indices = website.indices
, media_type_default = website.media_type_default
, pathparts = request.line.uri.path.parts
, uripath = request.line.uri.path.raw
, querystring = request.line.uri.querystring.raw
, startdir = website.www_root
, directory_default = directory_default
, favicon_default = website.find_ours('favicon.ico')
)

for k, v in result.wildcards.iteritems():
request.line.uri.path[k] = v
return {'dispatch_result': result}


def apply_typecasters_to_path(website, request):
typecasting.apply_typecasters(website.typecasters, request.line.uri.path)


def get_resource_for_request(website, request):
return {'resource': resources.get(website, request)}
def get_resource_for_request(website, request, dispatch_result):
return {'resource': resources.get(website, dispatch_result.match)}


def get_response_for_resource(request, resource=None):
def get_response_for_resource(request, dispatch_result, resource=None):
if resource is not None:
return {'response': resource.respond(request)}
return {'response': resource.respond(request, dispatch_result)}


def get_response_for_exception(website, exception):
Expand Down Expand Up @@ -103,17 +123,17 @@ def delegate_error_to_simplate(website, request, response, resource=None):

code = str(response.code)
possibles = [code + ".spt", "error.spt"]
fs = _first(website.ours_or_theirs(errpage) for errpage in possibles)
fspath = _first(website.ours_or_theirs(errpage) for errpage in possibles)

if fs is not None:
request.fs = fs
if fspath is not None:
request.original_resource = resource
if resource is not None:
# Try to return an error that matches the type of the original resource.
request.headers['Accept'] = resource.media_type + ', text/plain; q=0.1'
resource = resources.get(website, request)
resource = resources.get(website, fspath)
dispatch_result = DispatchResult(DispatchStatus.okay, fspath, {}, 'Found.', {}, True)
try:
response = resource.respond(request, response)
response = resource.respond(request, dispatch_result, response)
except Response as response:
if response.code != 406:
raise
Expand All @@ -130,7 +150,7 @@ def log_traceback_for_exception(website, exception):
return {'response': response, 'exception': None}


def log_result_of_request(website, request=None, response=None):
def log_result_of_request(website, request=None, dispatch_result=None, response=None):
"""Log access. With our own format (not Apache's).
"""

Expand All @@ -144,14 +164,14 @@ def log_result_of_request(website, request=None, response=None):
if request is None:
msg = "(no request available)"
else:
fs = getattr(request, 'fs', '')
if fs.startswith(website.www_root):
fs = fs[len(website.www_root):]
if fs:
fs = '.'+fs
fspath = getattr(dispatch_result, 'match', '')
if fspath.startswith(website.www_root):
fspath = fspath[len(website.www_root):]
if fspath:
fspath = '.' + fspath
else:
fs = '...' + fs[-21:]
msg = "%-24s %s" % (request.line.uri.path.raw, fs)
fspath = '...' + fspath[-21:]
msg = "%-24s %s" % (request.line.uri.path.raw, fspath)


# Where was response raised from?
Expand Down
148 changes: 75 additions & 73 deletions aspen/dispatcher.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -51,13 +51,11 @@ def debug_ext():
return a, b


class DispatchStatus:
class DispatchStatus(object):
okay, missing, non_leaf = range(3)


DispatchResult = namedtuple( 'DispatchResult'
, 'status match wildcards detail'.split()
)
DispatchResult = namedtuple('DispatchResult', 'status match wildcards detail extra constrain_path')


def dispatch_abstract(listnodes, is_leaf, traverse, find_index, noext_matched,
Expand Down Expand Up @@ -100,7 +98,7 @@ def get_wildleaf_fallback():
ext = lastnode_ext if lastnode_ext in wildleafs else None
curnode, wildvals = wildleafs[ext]
debug(lambda: "Wildcard leaf match %r and ext %r" % (curnode, ext))
return DispatchResult(DispatchStatus.okay, curnode, wildvals, "Found.")
return DispatchResult(DispatchStatus.okay, curnode, wildvals, "Found.", {}, True)
return None

for depth, node in enumerate(nodepath):
Expand Down Expand Up @@ -147,14 +145,22 @@ def get_wildleaf_fallback():
curnode = traverse(curnode, found_n)
node_name = found_n[1:-4] # strip leading % and trailing .spt
wildvals[node_name] = node
return DispatchResult(DispatchStatus.okay, curnode, wildvals, "Found.")
return DispatchResult( DispatchStatus.okay
, curnode
, wildvals
, "Found."
, {}
, True
)
elif node in subnodes and is_leaf_node(node):
debug(lambda: "...found exact file, must be static")
if is_spt(node):
return DispatchResult( DispatchStatus.missing
, None
, None
, "Node %r Not Found" % node
, {}
, True
)
else:
found_n = node
Expand Down Expand Up @@ -182,6 +188,8 @@ def get_wildleaf_fallback():
, curnode
, None
, "Tried to access non-leaf node as leaf."
, {}
, True
)
return result
elif node in subnodes:
Expand All @@ -190,6 +198,8 @@ def get_wildleaf_fallback():
, curnode
, None
, "Tried to access non-leaf node as leaf."
, {}
, True
)
else:
debug(lambda: "fallthrough")
Expand All @@ -199,6 +209,8 @@ def get_wildleaf_fallback():
, None
, None
, "Node %r Not Found" % node
, {}
, True
)
return result

Expand All @@ -223,10 +235,12 @@ def get_wildleaf_fallback():
, None
, None
, "Node %r Not Found" % node
, {}
, True
)
return result

return DispatchResult(DispatchStatus.okay, curnode, wildvals, "Found.")
return DispatchResult(DispatchStatus.okay, curnode, wildvals, "Found.", {}, True)


def match_index(indices, indir):
Expand All @@ -248,35 +262,29 @@ def is_first_index(indices, basedir, name):
return False


def update_neg_type(website, request, filename):
def update_neg_type(media_type_default, capture_accept, filename):
media_type = mimetypes.guess_type(filename, strict=False)[0]
if media_type is None:
media_type = website.media_type_default
request.headers['X-Aspen-Accept'] = media_type
debug(lambda: "set x-aspen-accept to %r" % media_type)
media_type = media_type_default
capture_accept['accept'] = media_type
debug(lambda: "set result.extra['accept'] to %r" % media_type)


def dispatch(website, request, pure_dispatch=False):
def dispatch(indices, media_type_default, pathparts, uripath, querystring, startdir,
directory_default, favicon_default):
"""Concretize dispatch_abstract.
This is all side-effecty on the request object, setting, at the least,
request.fs, and at worst other random contents including but not limited
to: request.line.uri.path, request.headers.
"""

# Handle URI path parts
pathparts = request.line.uri.path.parts

# Set up the real environment for the dispatcher.
# ===============================================

capture_accept = {}
listnodes = os.listdir
is_leaf = os.path.isfile
traverse = os.path.join
find_index = lambda x: match_index(website.indices, x)
noext_matched = lambda x: update_neg_type(website, request, x)
startdir = website.www_root
find_index = lambda x: match_index(indices, x)
noext_matched = lambda x: update_neg_type(media_type_default, capture_accept, x)


# Dispatch!
# =========
Expand All @@ -292,71 +300,64 @@ def dispatch(website, request, pure_dispatch=False):

debug(lambda: "dispatch_abstract returned: " + repr(result))

if 'accept' in capture_accept:
result.extra['accept'] = capture_accept['accept']

if result.match:
debug(lambda: "result.match is true" )
matchbase, matchname = result.match.rsplit(os.path.sep,1)
if pathparts[-1] != '' and matchname in website.indices and \
is_first_index(website.indices, matchbase, matchname):
if pathparts[-1] != '' and matchname in indices and \
is_first_index(indices, matchbase, matchname):
# asked for something that maps to a default index file; redirect to / per issue #175
debug( lambda: "found default index '%s' maps into %r"
% (pathparts[-1], website.indices)
% (pathparts[-1], indices)
)
uri = request.line.uri
location = uri.path.raw[:-len(pathparts[-1])]
if uri.querystring.raw:
location += '?' + uri.querystring.raw
location = uripath[:-len(pathparts[-1])]
if querystring:
location += '?' + querystring
raise Response(302, headers={'Location': location})

if not pure_dispatch:

# favicon.ico
# ===========
# Serve Aspen's favicon if there's not one.

if request.line.uri.path.raw == '/favicon.ico':
if result.status != DispatchStatus.okay:
path = request.line.uri.path.raw[1:]
request.fs = website.find_ours(path)
return


# robots.txt
# ==========
# Don't let robots.txt be handled by anything other than an actual
# robots.txt file

if request.line.uri.path.raw == '/robots.txt':
if result.status != DispatchStatus.missing:
if not result.match.endswith('robots.txt'):
raise Response(404)


# Handle returned states.
# =======================

if result.status != DispatchStatus.missing:
if uripath == '/robots.txt' and not result.match.endswith('robots.txt'): # robots.txt
# Don't let robots.txt be handled by anything other than an actual robots.txt file,
# because if you don't have a robots.txt but you do have a wildcard, then you end
# up with logspam.
raise Response(404)

if result.status == DispatchStatus.okay:
if result.match.endswith('/'): # autoindex
if not website.list_directories:
if result.match.endswith('/'):
if directory_default: # autoindex
result = DispatchResult( result.status
, directory_default
, {}
, 'Directory default.'
, {'autoindexdir': result.match}
, False
)
else:
raise Response(404)
autoindex = website.ours_or_theirs('autoindex.html.spt')
assert autoindex is not None # sanity check
request.headers['X-Aspen-AutoIndexDir'] = result.match
request.fs = autoindex
return # return so we skip the no-escape check
else: # normal match
request.fs = result.match
for k, v in result.wildcards.iteritems():
request.line.uri.path[k] = v

elif result.status == DispatchStatus.non_leaf: # trailing-slash redirect
uri = request.line.uri
location = uri.path.raw + '/'
if uri.querystring.raw:
location += '?' + uri.querystring.raw

elif result.status == DispatchStatus.non_leaf: # trailing slash
location = uripath + '/'
if querystring:
location += '?' + querystring
raise Response(302, headers={'Location': location})

elif result.status == DispatchStatus.missing: # 404
raise Response(404)
elif result.status == DispatchStatus.missing: # 404, but ...
if uripath == '/favicon.ico' and favicon_default: # favicon.ico
result = DispatchResult( DispatchStatus.okay
, favicon_default
, {}
, 'Favicon default.'
, {}
, False
)
else:
raise Response(404)

else:
raise Response(500, "Unknown result status.")
Expand All @@ -365,6 +366,7 @@ def dispatch(website, request, pure_dispatch=False):
# Protect against escaping the www_root.
# ======================================

if not request.fs.startswith(startdir):
if result.constrain_path and not result.match.startswith(startdir):
raise Response(404)

return result
Loading

0 comments on commit 78ec9a3

Please sign in to comment.