Mercurial > templog
diff web/bottle.py @ 547:7689af9f56a8
Update to bottle 0.12.8
author | Matt Johnston <matt@ucc.asn.au> |
---|---|
date | Wed, 27 May 2015 22:17:39 +0800 |
parents | 298e502fdcd4 |
children |
line wrap: on
line diff
--- a/web/bottle.py Sat May 23 10:08:42 2015 +0800 +++ b/web/bottle.py Wed May 27 22:17:39 2015 +0800 @@ -9,14 +9,14 @@ Homepage and documentation: http://bottlepy.org/ -Copyright (c) 2011, Marcel Hellkamp. +Copyright (c) 2013, Marcel Hellkamp. License: MIT (see LICENSE for details) """ from __future__ import with_statement __author__ = 'Marcel Hellkamp' -__version__ = '0.11.dev' +__version__ = '0.12.8' __license__ = 'MIT' # The gevent server adapter needs to patch some modules before they are imported @@ -36,15 +36,18 @@ import gevent.monkey; gevent.monkey.patch_all() import base64, cgi, email.utils, functools, hmac, imp, itertools, mimetypes,\ - os, re, subprocess, sys, tempfile, threading, time, urllib, warnings + os, re, subprocess, sys, tempfile, threading, time, warnings from datetime import date as datedate, datetime, timedelta from tempfile import TemporaryFile from traceback import format_exc, print_exc - -try: from json import dumps as json_dumps, loads as json_lds +from inspect import getargspec +from unicodedata import normalize + + +try: from simplejson import dumps as json_dumps, loads as json_lds except ImportError: # pragma: no cover - try: from simplejson import dumps as json_dumps, loads as json_lds + try: from json import dumps as json_dumps, loads as json_lds except ImportError: try: from django.utils.simplejson import dumps as json_dumps, loads as json_lds except ImportError: @@ -58,8 +61,9 @@ # It ain't pretty but it works... Sorry for the mess. py = sys.version_info -py3k = py >= (3,0,0) -py25 = py < (2,6,0) +py3k = py >= (3, 0, 0) +py25 = py < (2, 6, 0) +py31 = (3, 1, 0) <= py < (3, 2, 0) # Workaround for the missing "as" keyword in py3k. def _e(): return sys.exc_info()[1] @@ -76,17 +80,20 @@ if py3k: import http.client as httplib import _thread as thread - from urllib.parse import urljoin, parse_qsl, SplitResult as UrlSplitResult + from urllib.parse import urljoin, SplitResult as UrlSplitResult from urllib.parse import urlencode, quote as urlquote, unquote as urlunquote + urlunquote = functools.partial(urlunquote, encoding='latin1') from http.cookies import SimpleCookie from collections import MutableMapping as DictMixin import pickle from io import BytesIO + from configparser import ConfigParser basestring = str unicode = str json_loads = lambda s: json_lds(touni(s)) callable = lambda x: hasattr(x, '__call__') imap = map + def _raise(*a): raise a[0](a[1]).with_traceback(a[2]) else: # 2.x import httplib import thread @@ -96,17 +103,18 @@ from itertools import imap import cPickle as pickle from StringIO import StringIO as BytesIO + from ConfigParser import SafeConfigParser as ConfigParser if py25: - msg = "Python 2.5 support may be dropped in future versions of Bottle." + msg = "Python 2.5 support may be dropped in future versions of Bottle." warnings.warn(msg, DeprecationWarning) - from cgi import parse_qsl from UserDict import DictMixin def next(it): return it.next() bytes = str else: # 2.6, 2.7 - from urlparse import parse_qsl from collections import MutableMapping as DictMixin + unicode = unicode json_loads = json_lds + eval(compile('def _raise(*a): raise a[0], a[1], a[2]', '<py3fix>', 'exec')) # Some helpers for string/byte handling def tob(s, enc='utf8'): @@ -117,12 +125,12 @@ # 3.2 fixes cgi.FieldStorage to accept bytes (which makes a lot of sense). # 3.1 needs a workaround. -NCTextIOWrapper = None -if (3,0,0) < py < (3,2,0): +if py31: from io import TextIOWrapper class NCTextIOWrapper(TextIOWrapper): def close(self): pass # Keep wrapped buffer open. + # A bug in functools causes it to break if the wrapper is an instance method def update_wrapper(wrapper, wrapped, *a, **ka): try: functools.update_wrapper(wrapper, wrapped, *a, **ka) @@ -133,7 +141,7 @@ # These helpers are used at module level and need to be defined first. # And yes, I know PEP-8, but sometimes a lower-case classname makes more sense. -def depr(message): +def depr(message, hard=False): warnings.warn(message, DeprecationWarning, stacklevel=3) def makelist(data): # This is just to handy @@ -173,6 +181,7 @@ property. ''' def __init__(self, func): + self.__doc__ = getattr(func, '__doc__') self.func = func def __get__(self, obj, cls): @@ -207,34 +216,6 @@ pass -#TODO: This should subclass BaseRequest -class HTTPResponse(BottleException): - """ Used to break execution and immediately finish the response """ - def __init__(self, output='', status=200, header=None): - super(BottleException, self).__init__("HTTP Response %d" % status) - self.status = int(status) - self.output = output - self.headers = HeaderDict(header) if header else None - - def apply(self, response): - if self.headers: - for key, value in self.headers.allitems(): - response.headers[key] = value - response.status = self.status - - -class HTTPError(HTTPResponse): - """ Used to generate an error page """ - def __init__(self, code=500, output='Unknown Error', exception=None, - traceback=None, header=None): - super(HTTPError, self).__init__(output, code, header) - self.exception = exception - self.traceback = traceback - - def __repr__(self): - return tonat(template(ERROR_PAGE_TEMPLATE, e=self)) - - @@ -256,11 +237,19 @@ class RouteSyntaxError(RouteError): - """ The route parser found something not supported by this router """ + """ The route parser found something not supported by this router. """ class RouteBuildError(RouteError): - """ The route could not been built """ + """ The route could not be built. """ + + +def _re_flatten(p): + ''' Turn all capturing groups in a regular expression pattern into + non-capturing groups. ''' + if '(' not in p: return p + return re.sub(r'(\\*)(\(\?P<[^>]+>|\((?!\?))', + lambda m: m.group(0) if len(m.group(1)) % 2 else m.group(1) + '(?:', p) class Router(object): @@ -276,34 +265,27 @@ ''' default_pattern = '[^/]+' - default_filter = 're' - #: Sorry for the mess. It works. Trust me. - rule_syntax = re.compile('(\\\\*)'\ - '(?:(?::([a-zA-Z_][a-zA-Z_0-9]*)?()(?:#(.*?)#)?)'\ - '|(?:<([a-zA-Z_][a-zA-Z_0-9]*)?(?::([a-zA-Z_]*)'\ - '(?::((?:\\\\.|[^\\\\>]+)+)?)?)?>))') + default_filter = 're' + + #: The current CPython regexp implementation does not allow more + #: than 99 matching groups per regular expression. + _MAX_GROUPS_PER_PATTERN = 99 def __init__(self, strict=False): - self.rules = {} # A {rule: Rule} mapping - self.builder = {} # A rule/name->build_info mapping - self.static = {} # Cache for static routes: {path: {method: target}} - self.dynamic = [] # Cache for dynamic routes. See _compile() + self.rules = [] # All rules in order + self._groups = {} # index of regexes to find them in dyna_routes + self.builder = {} # Data structure for the url builder + self.static = {} # Search structure for static routes + self.dyna_routes = {} + self.dyna_regexes = {} # Search structure for dynamic routes #: If true, static routes are no longer checked first. self.strict_order = strict - self.filters = {'re': self.re_filter, 'int': self.int_filter, - 'float': self.float_filter, 'path': self.path_filter} - - def re_filter(self, conf): - return conf or self.default_pattern, None, None - - def int_filter(self, conf): - return r'-?\d+', int, lambda x: str(int(x)) - - def float_filter(self, conf): - return r'-?[\d.]+', float, lambda x: str(float(x)) - - def path_filter(self, conf): - return r'.+?', None, None + self.filters = { + 're': lambda conf: + (_re_flatten(conf or self.default_pattern), None, None), + 'int': lambda conf: (r'-?\d+', int, lambda x: str(int(x))), + 'float': lambda conf: (r'-?[\d.]+', float, lambda x: str(float(x))), + 'path': lambda conf: (r'.+?', None, None)} def add_filter(self, name, func): ''' Add a filter. The provided function is called with the configuration @@ -311,9 +293,12 @@ The first element is a string, the last two are callables or None. ''' self.filters[name] = func - def parse_rule(self, rule): - ''' Parses a rule into a (name, filter, conf) token stream. If mode is - None, name contains a static rule part. ''' + rule_syntax = re.compile('(\\\\*)'\ + '(?:(?::([a-zA-Z_][a-zA-Z_0-9]*)?()(?:#(.*?)#)?)'\ + '|(?:<([a-zA-Z_][a-zA-Z_0-9]*)?(?::([a-zA-Z_]*)'\ + '(?::((?:\\\\.|[^\\\\>]+)+)?)?)?>))') + + def _itertokens(self, rule): offset, prefix = 0, '' for match in self.rule_syntax.finditer(rule): prefix += rule[offset:match.start()] @@ -322,77 +307,95 @@ prefix += match.group(0)[len(g[0]):] offset = match.end() continue - if prefix: yield prefix, None, None - name, filtr, conf = g[1:4] if not g[2] is None else g[4:7] - if not filtr: filtr = self.default_filter - yield name, filtr, conf or None + if prefix: + yield prefix, None, None + name, filtr, conf = g[4:7] if g[2] is None else g[1:4] + yield name, filtr or 'default', conf or None offset, prefix = match.end(), '' if offset <= len(rule) or prefix: yield prefix+rule[offset:], None, None def add(self, rule, method, target, name=None): - ''' Add a new route or replace the target for an existing route. ''' - if rule in self.rules: - self.rules[rule][method] = target - if name: self.builder[name] = self.builder[rule] - return - - target = self.rules[rule] = {method: target} - - # Build pattern and other structures for dynamic routes - anons = 0 # Number of anonymous wildcards - pattern = '' # Regular expression pattern - filters = [] # Lists of wildcard input filters - builder = [] # Data structure for the URL builder + ''' Add a new rule or replace the target for an existing rule. ''' + anons = 0 # Number of anonymous wildcards found + keys = [] # Names of keys + pattern = '' # Regular expression pattern with named groups + filters = [] # Lists of wildcard input filters + builder = [] # Data structure for the URL builder is_static = True - for key, mode, conf in self.parse_rule(rule): + + for key, mode, conf in self._itertokens(rule): if mode: is_static = False + if mode == 'default': mode = self.default_filter mask, in_filter, out_filter = self.filters[mode](conf) - if key: + if not key: + pattern += '(?:%s)' % mask + key = 'anon%d' % anons + anons += 1 + else: pattern += '(?P<%s>%s)' % (key, mask) - else: - pattern += '(?:%s)' % mask - key = 'anon%d' % anons; anons += 1 + keys.append(key) if in_filter: filters.append((key, in_filter)) builder.append((key, out_filter or str)) elif key: pattern += re.escape(key) builder.append((None, key)) + self.builder[rule] = builder if name: self.builder[name] = builder if is_static and not self.strict_order: - self.static[self.build(rule)] = target + self.static.setdefault(method, {}) + self.static[method][self.build(rule)] = (target, None) return - def fpat_sub(m): - return m.group(0) if len(m.group(1)) % 2 else m.group(1) + '(?:' - flat_pattern = re.sub(r'(\\*)(\(\?P<[^>]*>|\((?!\?))', fpat_sub, pattern) - try: - re_match = re.compile('^(%s)$' % pattern).match + re_pattern = re.compile('^(%s)$' % pattern) + re_match = re_pattern.match except re.error: raise RouteSyntaxError("Could not add Route: %s (%s)" % (rule, _e())) - def match(path): - """ Return an url-argument dictionary. """ - url_args = re_match(path).groupdict() - for name, wildcard_filter in filters: - try: - url_args[name] = wildcard_filter(url_args[name]) - except ValueError: - raise HTTPError(400, 'Path has wrong format.') - return url_args - - try: - combined = '%s|(^%s$)' % (self.dynamic[-1][0].pattern, flat_pattern) - self.dynamic[-1] = (re.compile(combined), self.dynamic[-1][1]) - self.dynamic[-1][1].append((match, target)) - except (AssertionError, IndexError): # AssertionError: Too many groups - self.dynamic.append((re.compile('(^%s$)' % flat_pattern), - [(match, target)])) - return match + if filters: + def getargs(path): + url_args = re_match(path).groupdict() + for name, wildcard_filter in filters: + try: + url_args[name] = wildcard_filter(url_args[name]) + except ValueError: + raise HTTPError(400, 'Path has wrong format.') + return url_args + elif re_pattern.groupindex: + def getargs(path): + return re_match(path).groupdict() + else: + getargs = None + + flatpat = _re_flatten(pattern) + whole_rule = (rule, flatpat, target, getargs) + + if (flatpat, method) in self._groups: + if DEBUG: + msg = 'Route <%s %s> overwrites a previously defined route' + warnings.warn(msg % (method, rule), RuntimeWarning) + self.dyna_routes[method][self._groups[flatpat, method]] = whole_rule + else: + self.dyna_routes.setdefault(method, []).append(whole_rule) + self._groups[flatpat, method] = len(self.dyna_routes[method]) - 1 + + self._compile(method) + + def _compile(self, method): + all_rules = self.dyna_routes[method] + comborules = self.dyna_regexes[method] = [] + maxgroups = self._MAX_GROUPS_PER_PATTERN + for x in range(0, len(all_rules), maxgroups): + some = all_rules[x:x+maxgroups] + combined = (flatpat for (_, flatpat, _, _) in some) + combined = '|'.join('(^%s$)' % flatpat for flatpat in combined) + combined = re.compile(combined).match + rules = [(target, getargs) for (_, _, target, getargs) in some] + comborules.append((combined, rules)) def build(self, _name, *anons, **query): ''' Build an URL by filling the wildcards in a rule. ''' @@ -407,31 +410,46 @@ def match(self, environ): ''' Return a (target, url_agrs) tuple or raise HTTPError(400/404/405). ''' - path, targets, urlargs = environ['PATH_INFO'] or '/', None, {} - if path in self.static: - targets = self.static[path] + verb = environ['REQUEST_METHOD'].upper() + path = environ['PATH_INFO'] or '/' + target = None + if verb == 'HEAD': + methods = ['PROXY', verb, 'GET', 'ANY'] else: - for combined, rules in self.dynamic: - match = combined.match(path) - if not match: continue - getargs, targets = rules[match.lastindex - 1] - urlargs = getargs(path) if getargs else {} - break - - if not targets: - raise HTTPError(404, "Not found: " + repr(environ['PATH_INFO'])) - method = environ['REQUEST_METHOD'].upper() - if method in targets: - return targets[method], urlargs - if method == 'HEAD' and 'GET' in targets: - return targets['GET'], urlargs - if 'ANY' in targets: - return targets['ANY'], urlargs - allowed = [verb for verb in targets if verb != 'ANY'] - if 'GET' in allowed and 'HEAD' not in allowed: - allowed.append('HEAD') - raise HTTPError(405, "Method not allowed.", - header=[('Allow',",".join(allowed))]) + methods = ['PROXY', verb, 'ANY'] + + for method in methods: + if method in self.static and path in self.static[method]: + target, getargs = self.static[method][path] + return target, getargs(path) if getargs else {} + elif method in self.dyna_regexes: + for combined, rules in self.dyna_regexes[method]: + match = combined(path) + if match: + target, getargs = rules[match.lastindex - 1] + return target, getargs(path) if getargs else {} + + # No matching route found. Collect alternative methods for 405 response + allowed = set([]) + nocheck = set(methods) + for method in set(self.static) - nocheck: + if path in self.static[method]: + allowed.add(verb) + for method in set(self.dyna_regexes) - allowed - nocheck: + for combined, rules in self.dyna_regexes[method]: + match = combined(path) + if match: + allowed.add(method) + if allowed: + allow_header = ",".join(sorted(allowed)) + raise HTTPError(405, "Method not allowed.", Allow=allow_header) + + # No matching route and no alternative method found. We give up + raise HTTPError(404, "Not found: " + repr(path)) + + + + class Route(object): @@ -459,12 +477,12 @@ #: Additional keyword arguments passed to the :meth:`Bottle.route` #: decorator are stored in this dictionary. Used for route-specific #: plugin configuration and meta-data. - self.config = ConfigDict(config) + self.config = ConfigDict().load_dict(config, make_namespaces=True) def __call__(self, *a, **ka): depr("Some APIs changed to return Route() instances instead of"\ " callables. Make sure to use the Route.call method and not to"\ - " call Route instances directly.") + " call Route instances directly.") #0.12 return self.call(*a, **ka) @cached_property @@ -484,7 +502,7 @@ @property def _context(self): - depr('Switch to Plugin API v2 and access the Route object directly.') + depr('Switch to Plugin API v2 and access the Route object directly.') #0.12 return dict(rule=self.rule, method=self.method, callback=self.callback, name=self.name, app=self.app, config=self.config, apply=self.plugins, skip=self.skiplist) @@ -516,8 +534,32 @@ update_wrapper(callback, self.callback) return callback + def get_undecorated_callback(self): + ''' Return the callback. If the callback is a decorated function, try to + recover the original function. ''' + func = self.callback + func = getattr(func, '__func__' if py3k else 'im_func', func) + closure_attr = '__closure__' if py3k else 'func_closure' + while hasattr(func, closure_attr) and getattr(func, closure_attr): + func = getattr(func, closure_attr)[0].cell_contents + return func + + def get_callback_args(self): + ''' Return a list of argument names the callback (most likely) accepts + as keyword arguments. If the callback is a decorated function, try + to recover the original function before inspection. ''' + return getargspec(self.get_undecorated_callback())[0] + + def get_config(self, key, default=None): + ''' Lookup a config field and return its value, first checking the + route.config, then route.app.config.''' + for conf in (self.config, self.app.conifg): + if key in conf: return conf[key] + return default + def __repr__(self): - return '<%s %r %r>' % (self.method, self.rule, self.callback) + cb = self.get_undecorated_callback() + return '<%s %r %r>' % (self.method, self.rule, cb) @@ -539,28 +581,71 @@ """ def __init__(self, catchall=True, autojson=True): - #: If true, most exceptions are caught and returned as :exc:`HTTPError` - self.catchall = catchall - - #: A :cls:`ResourceManager` for application files + + #: A :class:`ConfigDict` for app specific configuration. + self.config = ConfigDict() + self.config._on_change = functools.partial(self.trigger_hook, 'config') + self.config.meta_set('autojson', 'validate', bool) + self.config.meta_set('catchall', 'validate', bool) + self.config['catchall'] = catchall + self.config['autojson'] = autojson + + #: A :class:`ResourceManager` for application files self.resources = ResourceManager() - #: A :cls:`ConfigDict` for app specific configuration. - self.config = ConfigDict() - self.config.autojson = autojson - self.routes = [] # List of installed :class:`Route` instances. self.router = Router() # Maps requests to :class:`Route` instances. self.error_handler = {} # Core plugins self.plugins = [] # List of installed plugins. - self.hooks = HooksPlugin() - self.install(self.hooks) - if self.config.autojson: + if self.config['autojson']: self.install(JSONPlugin()) self.install(TemplatePlugin()) + #: If true, most exceptions are caught and returned as :exc:`HTTPError` + catchall = DictProperty('config', 'catchall') + + __hook_names = 'before_request', 'after_request', 'app_reset', 'config' + __hook_reversed = 'after_request' + + @cached_property + def _hooks(self): + return dict((name, []) for name in self.__hook_names) + + def add_hook(self, name, func): + ''' Attach a callback to a hook. Three hooks are currently implemented: + + before_request + Executed once before each request. The request context is + available, but no routing has happened yet. + after_request + Executed once after each request regardless of its outcome. + app_reset + Called whenever :meth:`Bottle.reset` is called. + ''' + if name in self.__hook_reversed: + self._hooks[name].insert(0, func) + else: + self._hooks[name].append(func) + + def remove_hook(self, name, func): + ''' Remove a callback from a hook. ''' + if name in self._hooks and func in self._hooks[name]: + self._hooks[name].remove(func) + return True + + def trigger_hook(self, __name, *args, **kwargs): + ''' Trigger a hook and return a list of results. ''' + return [hook(*args, **kwargs) for hook in self._hooks[__name][:]] + + def hook(self, name): + """ Return a decorator that attaches a callback to a hook. See + :meth:`add_hook` for details.""" + def decorator(func): + self.add_hook(name, func) + return func + return decorator def mount(self, prefix, app, **options): ''' Mount an application (:class:`Bottle` or plain WSGI) to a specific @@ -575,8 +660,7 @@ All other parameters are passed to the underlying :meth:`route` call. ''' if isinstance(app, basestring): - prefix, app = app, prefix - depr('Parameter order of Bottle.mount() changed.') # 0.10 + depr('Parameter order of Bottle.mount() changed.', True) # 0.10 segments = [p for p in prefix.split('/') if p] if not segments: raise ValueError('Empty path prefix.') @@ -585,19 +669,25 @@ def mountpoint_wrapper(): try: request.path_shift(path_depth) - rs = BaseResponse([], 200) - def start_response(status, header): + rs = HTTPResponse([]) + def start_response(status, headerlist, exc_info=None): + if exc_info: + try: + _raise(*exc_info) + finally: + exc_info = None rs.status = status - for name, value in header: rs.add_header(name, value) + for name, value in headerlist: rs.add_header(name, value) return rs.body.append body = app(request.environ, start_response) - body = itertools.chain(rs.body, body) - return HTTPResponse(body, rs.status_code, rs.headers) + if body and rs.body: body = itertools.chain(rs.body, body) + rs.body = body or rs.body + return rs finally: request.path_shift(-path_depth) options.setdefault('skip', True) - options.setdefault('method', 'ANY') + options.setdefault('method', 'PROXY') options.setdefault('mountpoint', {'prefix': prefix, 'target': app}) options['callback'] = mountpoint_wrapper @@ -642,10 +732,6 @@ if removed: self.reset() return removed - def run(self, **kwargs): - ''' Calls :func:`run` with the same parameters. ''' - run(self, **kwargs) - def reset(self, route=None): ''' Reset all routes (force plugins to be re-applied) and clear all caches. If an ID or route object is given, only that specific route @@ -656,7 +742,7 @@ for route in routes: route.reset() if DEBUG: for route in routes: route.prepare() - self.hooks.trigger('app_reset') + self.trigger_hook('app_reset') def close(self): ''' Close the application and all installed plugins. ''' @@ -664,6 +750,10 @@ if hasattr(plugin, 'close'): plugin.close() self.stopped = True + def run(self, **kwargs): + ''' Calls :func:`run` with the same parameters. ''' + run(self, **kwargs) + def match(self, environ): """ Search for a matching route and return a (:class:`Route` , urlargs) tuple. The second value is a dictionary with parameters extracted @@ -748,39 +838,31 @@ return handler return wrapper - def hook(self, name): - """ Return a decorator that attaches a callback to a hook. Three hooks - are currently implemented: - - - before_request: Executed once before each request - - after_request: Executed once after each request - - app_reset: Called whenever :meth:`reset` is called. - """ - def wrapper(func): - self.hooks.add(name, func) - return func - return wrapper - - def handle(self, path, method='GET'): - """ (deprecated) Execute the first matching route callback and return - the result. :exc:`HTTPResponse` exceptions are caught and returned. - If :attr:`Bottle.catchall` is true, other exceptions are caught as - well and returned as :exc:`HTTPError` instances (500). - """ - depr("This method will change semantics in 0.10. Try to avoid it.") - if isinstance(path, dict): - return self._handle(path) - return self._handle({'PATH_INFO': path, 'REQUEST_METHOD': method.upper()}) + def default_error_handler(self, res): + return tob(template(ERROR_PAGE_TEMPLATE, e=res)) def _handle(self, environ): + path = environ['bottle.raw_path'] = environ['PATH_INFO'] + if py3k: + try: + environ['PATH_INFO'] = path.encode('latin1').decode('utf8') + except UnicodeError: + return HTTPError(400, 'Invalid path string. Expected UTF-8') + try: environ['bottle.app'] = self request.bind(environ) response.bind() - route, args = self.router.match(environ) - environ['route.handle'] = environ['bottle.route'] = route - environ['route.url_args'] = args - return route.call(**args) + try: + self.trigger_hook('before_request') + route, args = self.router.match(environ) + environ['route.handle'] = route + environ['bottle.route'] = route + environ['route.url_args'] = args + return route.call(**args) + finally: + self.trigger_hook('after_request') + except HTTPResponse: return _e() except RouteReset: @@ -803,7 +885,8 @@ # Empty output is done here if not out: - response['Content-Length'] = 0 + if 'Content-Length' not in response: + response['Content-Length'] = 0 return [] # Join lists of byte or unicode strings. Mixed lists are NOT supported if isinstance(out, (tuple, list))\ @@ -814,19 +897,18 @@ out = out.encode(response.charset) # Byte Strings are just returned if isinstance(out, bytes): - response['Content-Length'] = len(out) + if 'Content-Length' not in response: + response['Content-Length'] = len(out) return [out] # HTTPError or HTTPException (recursive, because they may wrap anything) # TODO: Handle these explicitly in handle() or make them iterable. if isinstance(out, HTTPError): out.apply(response) - out = self.error_handler.get(out.status, repr)(out) - if isinstance(out, HTTPResponse): - depr('Error handlers must not return :exc:`HTTPResponse`.') #0.9 + out = self.error_handler.get(out.status_code, self.default_error_handler)(out) return self._cast(out) if isinstance(out, HTTPResponse): out.apply(response) - return self._cast(out.output) + return self._cast(out.body) # File-like objects. if hasattr(out, 'read'): @@ -837,10 +919,10 @@ # Handle Iterables. We peek into them to detect their inner type. try: - out = iter(out) - first = next(out) + iout = iter(out) + first = next(iout) while not first: - first = next(out) + first = next(iout) except StopIteration: return self._cast('') except HTTPResponse: @@ -854,13 +936,17 @@ # These are the inner types allowed in iterator or generator objects. if isinstance(first, HTTPResponse): return self._cast(first) - if isinstance(first, bytes): - return itertools.chain([first], out) - if isinstance(first, unicode): - return imap(lambda x: x.encode(response.charset), - itertools.chain([first], out)) - return self._cast(HTTPError(500, 'Unsupported response type: %s'\ - % type(first))) + elif isinstance(first, bytes): + new_iter = itertools.chain([first], iout) + elif isinstance(first, unicode): + encoder = lambda x: x.encode(response.charset) + new_iter = imap(encoder, itertools.chain([first], iout)) + else: + msg = 'Unsupported response type: %s' % type(first) + return self._cast(HTTPError(500, msg)) + if hasattr(out, 'close'): + new_iter = _closeiter(new_iter, out.close) + return new_iter def wsgi(self, environ, start_response): """ The bottle WSGI-interface. """ @@ -868,12 +954,10 @@ out = self._cast(self._handle(environ)) # rfc2616 section 4.3 if response._status_code in (100, 101, 204, 304)\ - or request.method == 'HEAD': + or environ['REQUEST_METHOD'] == 'HEAD': if hasattr(out, 'close'): out.close() out = [] - if isinstance(response._status_line, unicode): - response._status_line = str(response._status_line) - start_response(response._status_line, list(response.iter_headers())) + start_response(response._status_line, response.headerlist) return out except (KeyboardInterrupt, SystemExit, MemoryError): raise @@ -887,7 +971,7 @@ % (html_escape(repr(_e())), html_escape(format_exc())) environ['wsgi.errors'].write(err) headers = [('Content-Type', 'text/html; charset=UTF-8')] - start_response('500 INTERNAL SERVER ERROR', headers) + start_response('500 INTERNAL SERVER ERROR', headers, sys.exc_info()) return [tob(err)] def __call__(self, environ, start_response): @@ -903,7 +987,6 @@ # HTTP and WSGI Tools ########################################################## ############################################################################### - class BaseRequest(object): """ A wrapper for WSGI environment dictionaries that adds a lot of convenient access methods and properties. Most of them are read-only. @@ -917,8 +1000,6 @@ #: Maximum size of memory buffer for :attr:`body` in bytes. MEMFILE_MAX = 102400 - #: Maximum number pr GET or POST parameters per request - MAX_PARAMS = 100 def __init__(self, environ=None): """ Wrap a WSGI environ dictionary. """ @@ -932,6 +1013,16 @@ ''' Bottle application handling this request. ''' raise RuntimeError('This request is not connected to an application.') + @DictProperty('environ', 'bottle.route', read_only=True) + def route(self): + """ The bottle :class:`Route` object that matches this request. """ + raise RuntimeError('This request is not connected to a route.') + + @DictProperty('environ', 'route.url_args', read_only=True) + def url_args(self): + """ The arguments extracted from the URL. """ + raise RuntimeError('This request is not connected to a route.') + @property def path(self): ''' The value of ``PATH_INFO`` with exactly one prefixed slash (to fix @@ -957,8 +1048,7 @@ def cookies(self): """ Cookies parsed into a :class:`FormsDict`. Signed cookies are NOT decoded. Use :meth:`get_cookie` if you expect signed cookies. """ - cookies = SimpleCookie(self.environ.get('HTTP_COOKIE','')) - cookies = list(cookies.values())[:self.MAX_PARAMS] + cookies = SimpleCookie(self.environ.get('HTTP_COOKIE','')).values() return FormsDict((c.key, c.value) for c in cookies) def get_cookie(self, key, default=None, secret=None): @@ -978,21 +1068,21 @@ values are sometimes called "URL arguments" or "GET parameters", but not to be confused with "URL wildcards" as they are provided by the :class:`Router`. ''' - pairs = parse_qsl(self.query_string, keep_blank_values=True) get = self.environ['bottle.get'] = FormsDict() - for key, value in pairs[:self.MAX_PARAMS]: + pairs = _parse_qsl(self.environ.get('QUERY_STRING', '')) + for key, value in pairs: get[key] = value return get @DictProperty('environ', 'bottle.request.forms', read_only=True) def forms(self): """ Form values parsed from an `url-encoded` or `multipart/form-data` - encoded POST or PUT request body. The result is retuned as a + encoded POST or PUT request body. The result is returned as a :class:`FormsDict`. All keys and values are strings. File uploads are stored separately in :attr:`files`. """ forms = FormsDict() for name, item in self.POST.allitems(): - if not hasattr(item, 'filename'): + if not isinstance(item, FileUpload): forms[name] = item return forms @@ -1009,24 +1099,13 @@ @DictProperty('environ', 'bottle.request.files', read_only=True) def files(self): - """ File uploads parsed from an `url-encoded` or `multipart/form-data` - encoded POST or PUT request body. The values are instances of - :class:`cgi.FieldStorage`. The most important attributes are: - - filename - The filename, if specified; otherwise None; this is the client - side filename, *not* the file name on which it is stored (that's - a temporary file you don't deal with) - file - The file(-like) object from which you can read the data. - value - The value as a *string*; for file uploads, this transparently - reads the file every time you request the value. Do not do this - on big files. + """ File uploads parsed from `multipart/form-data` encoded POST or PUT + request body. The values are instances of :class:`FileUpload`. + """ files = FormsDict() for name, item in self.POST.allitems(): - if hasattr(item, 'filename'): + if isinstance(item, FileUpload): files[name] = item return files @@ -1036,25 +1115,78 @@ property holds the parsed content of the request body. Only requests smaller than :attr:`MEMFILE_MAX` are processed to avoid memory exhaustion. ''' - if 'application/json' in self.environ.get('CONTENT_TYPE', '') \ - and 0 < self.content_length < self.MEMFILE_MAX: - return json_loads(self.body.read(self.MEMFILE_MAX)) + ctype = self.environ.get('CONTENT_TYPE', '').lower().split(';')[0] + if ctype == 'application/json': + b = self._get_body_string() + if not b: + return None + return json_loads(b) return None + def _iter_body(self, read, bufsize): + maxread = max(0, self.content_length) + while maxread: + part = read(min(maxread, bufsize)) + if not part: break + yield part + maxread -= len(part) + + def _iter_chunked(self, read, bufsize): + err = HTTPError(400, 'Error while parsing chunked transfer body.') + rn, sem, bs = tob('\r\n'), tob(';'), tob('') + while True: + header = read(1) + while header[-2:] != rn: + c = read(1) + header += c + if not c: raise err + if len(header) > bufsize: raise err + size, _, _ = header.partition(sem) + try: + maxread = int(tonat(size.strip()), 16) + except ValueError: + raise err + if maxread == 0: break + buff = bs + while maxread > 0: + if not buff: + buff = read(min(maxread, bufsize)) + part, buff = buff[:maxread], buff[maxread:] + if not part: raise err + yield part + maxread -= len(part) + if read(2) != rn: + raise err + @DictProperty('environ', 'bottle.request.body', read_only=True) def _body(self): - maxread = max(0, self.content_length) - stream = self.environ['wsgi.input'] - body = BytesIO() if maxread < self.MEMFILE_MAX else TemporaryFile(mode='w+b') - while maxread > 0: - part = stream.read(min(maxread, self.MEMFILE_MAX)) - if not part: break + body_iter = self._iter_chunked if self.chunked else self._iter_body + read_func = self.environ['wsgi.input'].read + body, body_size, is_temp_file = BytesIO(), 0, False + for part in body_iter(read_func, self.MEMFILE_MAX): body.write(part) - maxread -= len(part) + body_size += len(part) + if not is_temp_file and body_size > self.MEMFILE_MAX: + body, tmp = TemporaryFile(mode='w+b'), body + body.write(tmp.getvalue()) + del tmp + is_temp_file = True self.environ['wsgi.input'] = body body.seek(0) return body + def _get_body_string(self): + ''' read body until content-length or MEMFILE_MAX into a string. Raise + HTTPError(413) on requests that are to large. ''' + clen = self.content_length + if clen > self.MEMFILE_MAX: + raise HTTPError(413, 'Request to large') + if clen < 0: clen = self.MEMFILE_MAX + 1 + data = self.body.read(clen) + if len(data) > self.MEMFILE_MAX: # Fail fast + raise HTTPError(413, 'Request to large') + return data + @property def body(self): """ The HTTP request body as a seek-able file-like object. Depending on @@ -1065,6 +1197,11 @@ self._body.seek(0) return self._body + @property + def chunked(self): + ''' True if Chunked transfer encoding was. ''' + return 'chunked' in self.environ.get('HTTP_TRANSFER_ENCODING', '').lower() + #: An alias for :attr:`query`. GET = query @@ -1075,25 +1212,35 @@ instances of :class:`cgi.FieldStorage` (file uploads). """ post = FormsDict() + # We default to application/x-www-form-urlencoded for everything that + # is not multipart and take the fast path (also: 3.1 workaround) + if not self.content_type.startswith('multipart/'): + pairs = _parse_qsl(tonat(self._get_body_string(), 'latin1')) + for key, value in pairs: + post[key] = value + return post + safe_env = {'QUERY_STRING':''} # Build a safe environment for cgi for key in ('REQUEST_METHOD', 'CONTENT_TYPE', 'CONTENT_LENGTH'): if key in self.environ: safe_env[key] = self.environ[key] - if NCTextIOWrapper: - fb = NCTextIOWrapper(self.body, encoding='ISO-8859-1', newline='\n') - else: - fb = self.body - data = cgi.FieldStorage(fp=fb, environ=safe_env, keep_blank_values=True) - for item in (data.list or [])[:self.MAX_PARAMS]: - post[item.name] = item if item.filename else item.value + args = dict(fp=self.body, environ=safe_env, keep_blank_values=True) + if py31: + args['fp'] = NCTextIOWrapper(args['fp'], encoding='utf8', + newline='\n') + elif py3k: + args['encoding'] = 'utf8' + data = cgi.FieldStorage(**args) + self['_cgi.FieldStorage'] = data #http://bugs.python.org/issue18394#msg207958 + data = data.list or [] + for item in data: + if item.filename: + post[item.name] = FileUpload(item.file, item.name, + item.filename, item.headers) + else: + post[item.name] = item.value return post @property - def COOKIES(self): - ''' Alias for :attr:`cookies` (deprecated). ''' - depr('BaseRequest.COOKIES was renamed to BaseRequest.cookies (lowercase).') - return self.cookies - - @property def url(self): """ The full request URI including hostname and scheme. If your app lives behind a reverse proxy or load balancer and you get confusing @@ -1108,7 +1255,7 @@ but the fragment is always empty because it is not visible to the server. ''' env = self.environ - http = env.get('wsgi.url_scheme', 'http') + http = env.get('HTTP_X_FORWARDED_PROTO') or env.get('wsgi.url_scheme', 'http') host = env.get('HTTP_X_FORWARDED_HOST') or env.get('HTTP_HOST') if not host: # HTTP 1.1 requires a Host-header. This is for HTTP/1.0 clients. @@ -1157,6 +1304,11 @@ return int(self.environ.get('CONTENT_LENGTH') or -1) @property + def content_type(self): + ''' The Content-Type header as a lowercase-string (default: empty). ''' + return self.environ.get('CONTENT_TYPE', '').lower() + + @property def is_xhr(self): ''' True if the request was triggered by a XMLHttpRequest. This only works with JavaScript libraries that support the `X-Requested-With` @@ -1239,7 +1391,7 @@ var = self.environ['bottle.request.ext.%s'%name] return var.__get__(self) if hasattr(var, '__get__') else var except KeyError: - raise AttributeError('Attribute %r not defined.' % name) + raise AttributeError('Attribute %r not defined.' % name) def __setattr__(self, name, value): if name == 'environ': return object.__setattr__(self, name, value) @@ -1276,6 +1428,14 @@ This class does support dict-like case-insensitive item-access to headers, but is NOT a dict. Most notably, iterating over a response yields parts of the body and not the headers. + + :param body: The response body as one of the supported types. + :param status: Either an HTTP status code (e.g. 200) or a status line + including the reason phrase (e.g. '200 OK'). + :param headers: A dictionary or a list of name-value pairs. + + Additional keyword arguments are added to the list of headers. + Underscores in the header name are replaced with dashes. """ default_status = 200 @@ -1289,22 +1449,30 @@ 'Content-Length', 'Content-Range', 'Content-Type', 'Content-Md5', 'Last-Modified'))} - def __init__(self, body='', status=None, **headers): - self._status_line = None - self._status_code = None + def __init__(self, body='', status=None, headers=None, **more_headers): self._cookies = None - self._headers = {'Content-Type': [self.default_content_type]} + self._headers = {} self.body = body self.status = status or self.default_status if headers: - for name, value in headers.items(): - self[name] = value - - def copy(self): + if isinstance(headers, dict): + headers = headers.items() + for name, value in headers: + self.add_header(name, value) + if more_headers: + for name, value in more_headers.items(): + self.add_header(name, value) + + def copy(self, cls=None): ''' Returns a copy of self. ''' - copy = Response() + cls = cls or BaseResponse + assert issubclass(cls, BaseResponse) + copy = cls() copy.status = self.status copy._headers = dict((k, v[:]) for (k, v) in self._headers.items()) + if self._cookies: + copy._cookies = SimpleCookie() + copy._cookies.load(self._cookies.output(header='')) return copy def __iter__(self): @@ -1334,7 +1502,7 @@ raise ValueError('String status line without a reason phrase.') if not 100 <= code <= 999: raise ValueError('Status code out of range.') self._status_code = code - self._status_line = status or ('%d Unknown' % code) + self._status_line = str(status or ('%d Unknown' % code)) def _get_status(self): return self._status_line @@ -1351,7 +1519,7 @@ def headers(self): ''' An instance of :class:`HeaderDict`, a case-insensitive dict-like view on the response headers. ''' - self.__dict__['headers'] = hdict = HeaderDict() + hdict = HeaderDict() hdict.dict = self._headers return hdict @@ -1365,13 +1533,10 @@ header with that name, return a default value. ''' return self._headers.get(_hkey(name), [default])[-1] - def set_header(self, name, value, append=False): + def set_header(self, name, value): ''' Create a new response header, replacing any previously defined headers with the same name. ''' - if append: - self.add_header(name, value) - else: - self._headers[_hkey(name)] = [str(value)] + self._headers[_hkey(name)] = [str(value)] def add_header(self, name, value): ''' Add an additional response header, not removing duplicates. ''' @@ -1380,44 +1545,36 @@ def iter_headers(self): ''' Yield (header, value) tuples, skipping headers that are not allowed with the current response status code. ''' - headers = self._headers.items() - bad_headers = self.bad_headers.get(self._status_code) - if bad_headers: - headers = [h for h in headers if h[0] not in bad_headers] - for name, values in headers: - for value in values: - yield name, value - if self._cookies: - for c in self._cookies.values(): - yield 'Set-Cookie', c.OutputString() - - def wsgiheader(self): - depr('The wsgiheader method is deprecated. See headerlist.') #0.10 return self.headerlist @property def headerlist(self): ''' WSGI conform list of (header, value) tuples. ''' - return list(self.iter_headers()) + out = [] + headers = list(self._headers.items()) + if 'Content-Type' not in self._headers: + headers.append(('Content-Type', [self.default_content_type])) + if self._status_code in self.bad_headers: + bad_headers = self.bad_headers[self._status_code] + headers = [h for h in headers if h[0] not in bad_headers] + out += [(name, val) for name, vals in headers for val in vals] + if self._cookies: + for c in self._cookies.values(): + out.append(('Set-Cookie', c.OutputString())) + return out content_type = HeaderProperty('Content-Type') content_length = HeaderProperty('Content-Length', reader=int) + expires = HeaderProperty('Expires', + reader=lambda x: datetime.utcfromtimestamp(parse_date(x)), + writer=lambda x: http_date(x)) @property - def charset(self): + def charset(self, default='UTF-8'): """ Return the charset specified in the content-type header (default: utf8). """ if 'charset=' in self.content_type: return self.content_type.split('charset=')[-1].split(';')[0].strip() - return 'UTF-8' - - @property - def COOKIES(self): - """ A dict-like SimpleCookie instance. This should not be used directly. - See :meth:`set_cookie`. """ - depr('The COOKIES dict is deprecated. Use `set_cookie()` instead.') # 0.10 - if not self._cookies: - self._cookies = SimpleCookie() - return self._cookies + return default def set_cookie(self, name, value, secret=None, **options): ''' Create a new cookie or replace an old one. If the `secret` parameter is @@ -1488,48 +1645,66 @@ out += '%s: %s\n' % (name.title(), value.strip()) return out -#: Thread-local storage for :class:`LocalRequest` and :class:`LocalResponse` -#: attributes. -_lctx = threading.local() - -def local_property(name): + +def local_property(name=None): + if name: depr('local_property() is deprecated and will be removed.') #0.12 + ls = threading.local() def fget(self): - try: - return getattr(_lctx, name) + try: return ls.var except AttributeError: raise RuntimeError("Request context not initialized.") - def fset(self, value): setattr(_lctx, name, value) - def fdel(self): delattr(_lctx, name) - return property(fget, fset, fdel, - 'Thread-local property stored in :data:`_lctx.%s`' % name) + def fset(self, value): ls.var = value + def fdel(self): del ls.var + return property(fget, fset, fdel, 'Thread-local property') class LocalRequest(BaseRequest): ''' A thread-local subclass of :class:`BaseRequest` with a different - set of attribues for each thread. There is usually only one global + set of attributes for each thread. There is usually only one global instance of this class (:data:`request`). If accessed during a request/response cycle, this instance always refers to the *current* request (even on a multithreaded server). ''' bind = BaseRequest.__init__ - environ = local_property('request_environ') + environ = local_property() class LocalResponse(BaseResponse): ''' A thread-local subclass of :class:`BaseResponse` with a different - set of attribues for each thread. There is usually only one global + set of attributes for each thread. There is usually only one global instance of this class (:data:`response`). Its attributes are used to build the HTTP response at the end of the request/response cycle. ''' bind = BaseResponse.__init__ - _status_line = local_property('response_status_line') - _status_code = local_property('response_status_code') - _cookies = local_property('response_cookies') - _headers = local_property('response_headers') - body = local_property('response_body') - -Response = LocalResponse # BC 0.9 -Request = LocalRequest # BC 0.9 - + _status_line = local_property() + _status_code = local_property() + _cookies = local_property() + _headers = local_property() + body = local_property() + + +Request = BaseRequest +Response = BaseResponse + + +class HTTPResponse(Response, BottleException): + def __init__(self, body='', status=None, headers=None, **more_headers): + super(HTTPResponse, self).__init__(body, status, headers, **more_headers) + + def apply(self, response): + response._status_code = self._status_code + response._status_line = self._status_line + response._headers = self._headers + response._cookies = self._cookies + response.body = self.body + + +class HTTPError(HTTPResponse): + default_status = 500 + def __init__(self, status=None, body=None, exception=None, traceback=None, + **options): + self.exception = exception + self.traceback = traceback + super(HTTPError, self).__init__(body, status, **options) @@ -1541,6 +1716,7 @@ class PluginError(BottleException): pass + class JSONPlugin(object): name = 'json' api = 2 @@ -1548,63 +1724,26 @@ def __init__(self, json_dumps=json_dumps): self.json_dumps = json_dumps - def apply(self, callback, context): + def apply(self, callback, route): dumps = self.json_dumps if not dumps: return callback def wrapper(*a, **ka): - rv = callback(*a, **ka) + try: + rv = callback(*a, **ka) + except HTTPError: + rv = _e() + if isinstance(rv, dict): #Attempt to serialize, raises exception on failure json_response = dumps(rv) #Set content type only if serialization succesful response.content_type = 'application/json' return json_response + elif isinstance(rv, HTTPResponse) and isinstance(rv.body, dict): + rv.body = dumps(rv.body) + rv.content_type = 'application/json' return rv - return wrapper - - -class HooksPlugin(object): - name = 'hooks' - api = 2 - - _names = 'before_request', 'after_request', 'app_reset' - - def __init__(self): - self.hooks = dict((name, []) for name in self._names) - self.app = None - - def _empty(self): - return not (self.hooks['before_request'] or self.hooks['after_request']) - - def setup(self, app): - self.app = app - - def add(self, name, func): - ''' Attach a callback to a hook. ''' - was_empty = self._empty() - self.hooks.setdefault(name, []).append(func) - if self.app and was_empty and not self._empty(): self.app.reset() - - def remove(self, name, func): - ''' Remove a callback from a hook. ''' - was_empty = self._empty() - if name in self.hooks and func in self.hooks[name]: - self.hooks[name].remove(func) - if self.app and not was_empty and self._empty(): self.app.reset() - - def trigger(self, name, *a, **ka): - ''' Trigger a hook and return a list of results. ''' - hooks = self.hooks[name] - if ka.pop('reversed', False): hooks = hooks[::-1] - return [hook(*a, **ka) for hook in hooks] - - def apply(self, callback, context): - if self._empty(): return callback - def wrapper(*a, **ka): - self.trigger('before_request') - rv = callback(*a, **ka) - self.trigger('after_request', reversed=True) - return rv + return wrapper @@ -1620,9 +1759,6 @@ conf = route.config.get('template') if isinstance(conf, (tuple, list)) and len(conf) == 2: return view(conf[0], **conf[1])(callback) - elif isinstance(conf, str) and 'template_opts' in route.config: - depr('The `template_opts` parameter is deprecated.') #0.9 - return view(conf, **route.config['template_opts'])(callback) elif isinstance(conf, str): return view(conf)(callback) else: @@ -1642,13 +1778,13 @@ def find_module(self, fullname, path=None): if '.' not in fullname: return - packname, modname = fullname.rsplit('.', 1) + packname = fullname.rsplit('.', 1)[0] if packname != self.name: return return self def load_module(self, fullname): if fullname in sys.modules: return sys.modules[fullname] - packname, modname = fullname.rsplit('.', 1) + modname = fullname.rsplit('.', 1)[1] realname = self.impmask % modname __import__(realname) module = sys.modules[fullname] = sys.modules[realname] @@ -1739,7 +1875,6 @@ getlist = getall - class FormsDict(MultiDict): ''' This :class:`MultiDict` subclass is used to store request form data. Additionally to the normal dict-like item access methods (which return @@ -1756,10 +1891,11 @@ def _fix(self, s, encoding=None): if isinstance(s, unicode) and self.recode_unicode: # Python 3 WSGI - s = s.encode('latin1') - if isinstance(s, bytes): # Python 2 WSGI + return s.encode('latin1').decode(encoding or self.input_encoding) + elif isinstance(s, bytes): # Python 2 WSGI return s.decode(encoding or self.input_encoding) - return s + else: + return s def decode(self, encoding=None): ''' Returns a copy with all keys and values de- or recoded to match @@ -1773,6 +1909,7 @@ return copy def getunicode(self, name, default=None, encoding=None): + ''' Return the value as a unicode string, or the default. ''' try: return self._fix(self[name], encoding) except (UnicodeError, KeyError): @@ -1820,7 +1957,7 @@ Currently PEP 333, 444 and 3333 are supported. (PEP 444 is the only one that uses non-native strings.) ''' - #: List of keys that do not have a 'HTTP_' prefix. + #: List of keys that do not have a ``HTTP_`` prefix. cgikeys = ('CONTENT_TYPE', 'CONTENT_LENGTH') def __init__(self, environ): @@ -1858,39 +1995,212 @@ def __contains__(self, key): return self._ekey(key) in self.environ + class ConfigDict(dict): - ''' A dict-subclass with some extras: You can access keys like attributes. - Uppercase attributes create new ConfigDicts and act as name-spaces. - Other missing attributes return None. Calling a ConfigDict updates its - values and returns itself. - - >>> cfg = ConfigDict() - >>> cfg.Namespace.value = 5 - >>> cfg.OtherNamespace(a=1, b=2) - >>> cfg - {'Namespace': {'value': 5}, 'OtherNamespace': {'a': 1, 'b': 2}} + ''' A dict-like configuration storage with additional support for + namespaces, validators, meta-data, on_change listeners and more. + + This storage is optimized for fast read access. Retrieving a key + or using non-altering dict methods (e.g. `dict.get()`) has no overhead + compared to a native dict. ''' - + __slots__ = ('_meta', '_on_change') + + class Namespace(DictMixin): + + def __init__(self, config, namespace): + self._config = config + self._prefix = namespace + + def __getitem__(self, key): + depr('Accessing namespaces as dicts is discouraged. ' + 'Only use flat item access: ' + 'cfg["names"]["pace"]["key"] -> cfg["name.space.key"]') #0.12 + return self._config[self._prefix + '.' + key] + + def __setitem__(self, key, value): + self._config[self._prefix + '.' + key] = value + + def __delitem__(self, key): + del self._config[self._prefix + '.' + key] + + def __iter__(self): + ns_prefix = self._prefix + '.' + for key in self._config: + ns, dot, name = key.rpartition('.') + if ns == self._prefix and name: + yield name + + def keys(self): return [x for x in self] + def __len__(self): return len(self.keys()) + def __contains__(self, key): return self._prefix + '.' + key in self._config + def __repr__(self): return '<Config.Namespace %s.*>' % self._prefix + def __str__(self): return '<Config.Namespace %s.*>' % self._prefix + + # Deprecated ConfigDict features + def __getattr__(self, key): + depr('Attribute access is deprecated.') #0.12 + if key not in self and key[0].isupper(): + self[key] = ConfigDict.Namespace(self._config, self._prefix + '.' + key) + if key not in self and key.startswith('__'): + raise AttributeError(key) + return self.get(key) + + def __setattr__(self, key, value): + if key in ('_config', '_prefix'): + self.__dict__[key] = value + return + depr('Attribute assignment is deprecated.') #0.12 + if hasattr(DictMixin, key): + raise AttributeError('Read-only attribute.') + if key in self and self[key] and isinstance(self[key], self.__class__): + raise AttributeError('Non-empty namespace attribute.') + self[key] = value + + def __delattr__(self, key): + if key in self: + val = self.pop(key) + if isinstance(val, self.__class__): + prefix = key + '.' + for key in self: + if key.startswith(prefix): + del self[prefix+key] + + def __call__(self, *a, **ka): + depr('Calling ConfDict is deprecated. Use the update() method.') #0.12 + self.update(*a, **ka) + return self + + def __init__(self, *a, **ka): + self._meta = {} + self._on_change = lambda name, value: None + if a or ka: + depr('Constructor does no longer accept parameters.') #0.12 + self.update(*a, **ka) + + def load_config(self, filename): + ''' Load values from an *.ini style config file. + + If the config file contains sections, their names are used as + namespaces for the values within. The two special sections + ``DEFAULT`` and ``bottle`` refer to the root namespace (no prefix). + ''' + conf = ConfigParser() + conf.read(filename) + for section in conf.sections(): + for key, value in conf.items(section): + if section not in ('DEFAULT', 'bottle'): + key = section + '.' + key + self[key] = value + return self + + def load_dict(self, source, namespace='', make_namespaces=False): + ''' Import values from a dictionary structure. Nesting can be used to + represent namespaces. + + >>> ConfigDict().load_dict({'name': {'space': {'key': 'value'}}}) + {'name.space.key': 'value'} + ''' + stack = [(namespace, source)] + while stack: + prefix, source = stack.pop() + if not isinstance(source, dict): + raise TypeError('Source is not a dict (r)' % type(key)) + for key, value in source.items(): + if not isinstance(key, str): + raise TypeError('Key is not a string (%r)' % type(key)) + full_key = prefix + '.' + key if prefix else key + if isinstance(value, dict): + stack.append((full_key, value)) + if make_namespaces: + self[full_key] = self.Namespace(self, full_key) + else: + self[full_key] = value + return self + + def update(self, *a, **ka): + ''' If the first parameter is a string, all keys are prefixed with this + namespace. Apart from that it works just as the usual dict.update(). + Example: ``update('some.namespace', key='value')`` ''' + prefix = '' + if a and isinstance(a[0], str): + prefix = a[0].strip('.') + '.' + a = a[1:] + for key, value in dict(*a, **ka).items(): + self[prefix+key] = value + + def setdefault(self, key, value): + if key not in self: + self[key] = value + return self[key] + + def __setitem__(self, key, value): + if not isinstance(key, str): + raise TypeError('Key has type %r (not a string)' % type(key)) + + value = self.meta_get(key, 'filter', lambda x: x)(value) + if key in self and self[key] is value: + return + self._on_change(key, value) + dict.__setitem__(self, key, value) + + def __delitem__(self, key): + dict.__delitem__(self, key) + + def clear(self): + for key in self: + del self[key] + + def meta_get(self, key, metafield, default=None): + ''' Return the value of a meta field for a key. ''' + return self._meta.get(key, {}).get(metafield, default) + + def meta_set(self, key, metafield, value): + ''' Set the meta field for a key to a new value. This triggers the + on-change handler for existing keys. ''' + self._meta.setdefault(key, {})[metafield] = value + if key in self: + self[key] = self[key] + + def meta_list(self, key): + ''' Return an iterable of meta field names defined for a key. ''' + return self._meta.get(key, {}).keys() + + # Deprecated ConfigDict features def __getattr__(self, key): + depr('Attribute access is deprecated.') #0.12 if key not in self and key[0].isupper(): - self[key] = ConfigDict() + self[key] = self.Namespace(self, key) + if key not in self and key.startswith('__'): + raise AttributeError(key) return self.get(key) def __setattr__(self, key, value): + if key in self.__slots__: + return dict.__setattr__(self, key, value) + depr('Attribute assignment is deprecated.') #0.12 if hasattr(dict, key): raise AttributeError('Read-only attribute.') - if key in self and self[key] and isinstance(self[key], ConfigDict): + if key in self and self[key] and isinstance(self[key], self.Namespace): raise AttributeError('Non-empty namespace attribute.') self[key] = value def __delattr__(self, key): - if key in self: del self[key] + if key in self: + val = self.pop(key) + if isinstance(val, self.Namespace): + prefix = key + '.' + for key in self: + if key.startswith(prefix): + del self[prefix+key] def __call__(self, *a, **ka): - for key, value in dict(*a, **ka).items(): setattr(self, key, value) + depr('Calling ConfDict is deprecated. Use the update() method.') #0.12 + self.update(*a, **ka) return self + class AppStack(list): """ A stack-like list. Calling it returns the head of the stack. """ @@ -1921,11 +2231,27 @@ yield part +class _closeiter(object): + ''' This only exists to be able to attach a .close method to iterators that + do not support attribute assignment (most of itertools). ''' + + def __init__(self, iterator, close=None): + self.iterator = iterator + self.close_callbacks = makelist(close) + + def __iter__(self): + return iter(self.iterator) + + def close(self): + for func in self.close_callbacks: + func() + + class ResourceManager(object): ''' This class manages a list of search paths and helps to find and open - aplication-bound resources (files). - - :param base: default value for same-named :meth:`add_path` parameter. + application-bound resources (files). + + :param base: default value for :meth:`add_path` calls. :param opener: callable used to open resources. :param cachemode: controls which lookups are cached. One of 'all', 'found' or 'none'. @@ -1938,25 +2264,24 @@ #: A list of search paths. See :meth:`add_path` for details. self.path = [] - #: A cache for resolved paths. `res.cache.clear()`` clears the cache. + #: A cache for resolved paths. ``res.cache.clear()`` clears the cache. self.cache = {} def add_path(self, path, base=None, index=None, create=False): - ''' Add a new path to the list of search paths. Return False if it does - not exist. - - :param path: The new search path. Relative paths are turned into an - absolute and normalized form. If the path looks like a file (not - ending in `/`), the filename is stripped off. + ''' Add a new path to the list of search paths. Return False if the + path does not exist. + + :param path: The new search path. Relative paths are turned into + an absolute and normalized form. If the path looks like a file + (not ending in `/`), the filename is stripped off. :param base: Path used to absolutize relative search paths. - Defaults to `:attr:base` which defaults to ``./``. - :param index: Position within the list of search paths. Defaults to - last index (appends to the list). - :param create: Create non-existent search paths. Off by default. + Defaults to :attr:`base` which defaults to ``os.getcwd()``. + :param index: Position within the list of search paths. Defaults + to last index (appends to the list). The `base` parameter makes it easy to reference files installed along with a python module or package:: - + res.add_path('./resources/', __file__) ''' base = os.path.abspath(os.path.dirname(base or self.base)) @@ -1965,12 +2290,13 @@ if path in self.path: self.path.remove(path) if create and not os.path.isdir(path): - os.mkdirs(path) + os.makedirs(path) if index is None: self.path.append(path) else: self.path.insert(index, path) self.cache.clear() + return os.path.exists(path) def __iter__(self): ''' Iterate over all existing files in all registered paths. ''' @@ -2004,7 +2330,70 @@ ''' Find a resource and return a file object, or raise IOError. ''' fname = self.lookup(name) if not fname: raise IOError("Resource %r not found." % name) - return self.opener(name, mode=mode, *args, **kwargs) + return self.opener(fname, mode=mode, *args, **kwargs) + + +class FileUpload(object): + + def __init__(self, fileobj, name, filename, headers=None): + ''' Wrapper for file uploads. ''' + #: Open file(-like) object (BytesIO buffer or temporary file) + self.file = fileobj + #: Name of the upload form field + self.name = name + #: Raw filename as sent by the client (may contain unsafe characters) + self.raw_filename = filename + #: A :class:`HeaderDict` with additional headers (e.g. content-type) + self.headers = HeaderDict(headers) if headers else HeaderDict() + + content_type = HeaderProperty('Content-Type') + content_length = HeaderProperty('Content-Length', reader=int, default=-1) + + @cached_property + def filename(self): + ''' Name of the file on the client file system, but normalized to ensure + file system compatibility. An empty filename is returned as 'empty'. + + Only ASCII letters, digits, dashes, underscores and dots are + allowed in the final filename. Accents are removed, if possible. + Whitespace is replaced by a single dash. Leading or tailing dots + or dashes are removed. The filename is limited to 255 characters. + ''' + fname = self.raw_filename + if not isinstance(fname, unicode): + fname = fname.decode('utf8', 'ignore') + fname = normalize('NFKD', fname).encode('ASCII', 'ignore').decode('ASCII') + fname = os.path.basename(fname.replace('\\', os.path.sep)) + fname = re.sub(r'[^a-zA-Z0-9-_.\s]', '', fname).strip() + fname = re.sub(r'[-\s]+', '-', fname).strip('.-') + return fname[:255] or 'empty' + + def _copy_file(self, fp, chunk_size=2**16): + read, write, offset = self.file.read, fp.write, self.file.tell() + while 1: + buf = read(chunk_size) + if not buf: break + write(buf) + self.file.seek(offset) + + def save(self, destination, overwrite=False, chunk_size=2**16): + ''' Save file to disk or copy its content to an open file(-like) object. + If *destination* is a directory, :attr:`filename` is added to the + path. Existing files are not overwritten by default (IOError). + + :param destination: File path, directory or file(-like) object. + :param overwrite: If True, replace existing files. (default: False) + :param chunk_size: Bytes to read at a time. (default: 64kb) + ''' + if isinstance(destination, basestring): # Except file-likes here + if os.path.isdir(destination): + destination = os.path.join(destination, self.filename) + if not overwrite and os.path.exists(destination): + raise IOError('File exists.') + with open(destination, 'wb') as fp: + self._copy_file(fp, chunk_size) + else: + self._copy_file(destination, chunk_size) @@ -2016,7 +2405,7 @@ ############################################################################### -def abort(code=500, text='Unknown Error: Application stopped.'): +def abort(code=500, text='Unknown Error.'): """ Aborts execution and causes a HTTP error. """ raise HTTPError(code, text) @@ -2024,10 +2413,13 @@ def redirect(url, code=None): """ Aborts execution and causes a 303 or 302 redirect, depending on the HTTP protocol version. """ - if code is None: + if not code: code = 303 if request.get('SERVER_PROTOCOL') == "HTTP/1.1" else 302 - location = urljoin(request.url, url) - raise HTTPResponse("", status=code, header=dict(Location=location)) + res = response.copy(cls=HTTPResponse) + res.status = code + res.body = "" + res.set_header('Location', urljoin(request.url, url)) + raise res def _file_iter_range(fp, offset, bytes, maxread=1024*1024): @@ -2040,15 +2432,29 @@ yield part -def static_file(filename, root, mimetype='auto', download=False): +def static_file(filename, root, mimetype='auto', download=False, charset='UTF-8'): """ Open a file in a safe way and return :exc:`HTTPResponse` with status - code 200, 305, 401 or 404. Set Content-Type, Content-Encoding, - Content-Length and Last-Modified header. Obey If-Modified-Since header - and HEAD requests. + code 200, 305, 403 or 404. The ``Content-Type``, ``Content-Encoding``, + ``Content-Length`` and ``Last-Modified`` headers are set if possible. + Special support for ``If-Modified-Since``, ``Range`` and ``HEAD`` + requests. + + :param filename: Name or path of the file to send. + :param root: Root path for file lookups. Should be an absolute directory + path. + :param mimetype: Defines the content-type header (default: guess from + file extension) + :param download: If True, ask the browser to open a `Save as...` dialog + instead of opening the file with the associated program. You can + specify a custom filename as a string. If not specified, the + original filename is used (default: False). + :param charset: The charset to use for files with a ``text/*`` + mime-type. (default: UTF-8) """ + root = os.path.abspath(root) + os.sep filename = os.path.abspath(os.path.join(root, filename.strip('/\\'))) - header = dict() + headers = dict() if not filename.startswith(root): return HTTPError(403, "Access denied.") @@ -2059,41 +2465,43 @@ if mimetype == 'auto': mimetype, encoding = mimetypes.guess_type(filename) - if mimetype: header['Content-Type'] = mimetype - if encoding: header['Content-Encoding'] = encoding - elif mimetype: - header['Content-Type'] = mimetype + if encoding: headers['Content-Encoding'] = encoding + + if mimetype: + if mimetype[:5] == 'text/' and charset and 'charset' not in mimetype: + mimetype += '; charset=%s' % charset + headers['Content-Type'] = mimetype if download: download = os.path.basename(filename if download == True else download) - header['Content-Disposition'] = 'attachment; filename="%s"' % download + headers['Content-Disposition'] = 'attachment; filename="%s"' % download stats = os.stat(filename) - header['Content-Length'] = clen = stats.st_size + headers['Content-Length'] = clen = stats.st_size lm = time.strftime("%a, %d %b %Y %H:%M:%S GMT", time.gmtime(stats.st_mtime)) - header['Last-Modified'] = lm + headers['Last-Modified'] = lm ims = request.environ.get('HTTP_IF_MODIFIED_SINCE') if ims: ims = parse_date(ims.split(";")[0].strip()) if ims is not None and ims >= int(stats.st_mtime): - header['Date'] = time.strftime("%a, %d %b %Y %H:%M:%S GMT", time.gmtime()) - return HTTPResponse(status=304, header=header) + headers['Date'] = time.strftime("%a, %d %b %Y %H:%M:%S GMT", time.gmtime()) + return HTTPResponse(status=304, **headers) body = '' if request.method == 'HEAD' else open(filename, 'rb') - header["Accept-Ranges"] = "bytes" + headers["Accept-Ranges"] = "bytes" ranges = request.environ.get('HTTP_RANGE') if 'HTTP_RANGE' in request.environ: ranges = list(parse_range_header(request.environ['HTTP_RANGE'], clen)) if not ranges: return HTTPError(416, "Requested Range Not Satisfiable") offset, end = ranges[0] - header["Content-Range"] = "bytes %d-%d/%d" % (offset, end-1, clen) - header["Content-Length"] = str(end-offset) + headers["Content-Range"] = "bytes %d-%d/%d" % (offset, end-1, clen) + headers["Content-Length"] = str(end-offset) if body: body = _file_iter_range(body, offset, end-offset) - return HTTPResponse(body, header=header, status=206) - return HTTPResponse(body, header=header) + return HTTPResponse(body, status=206, **headers) + return HTTPResponse(body, **headers) @@ -2109,8 +2517,17 @@ """ Change the debug level. There is only one debug level supported at the moment.""" global DEBUG + if mode: warnings.simplefilter('default') DEBUG = bool(mode) +def http_date(value): + if isinstance(value, (datedate, datetime)): + value = value.utctimetuple() + elif isinstance(value, (int, float)): + value = time.gmtime(value) + if not isinstance(value, basestring): + value = time.strftime("%a, %d %b %Y %H:%M:%S GMT", value) + return value def parse_date(ims): """ Parse rfc1123, rfc850 and asctime timestamps and return UTC epoch. """ @@ -2120,7 +2537,6 @@ except (TypeError, ValueError, IndexError, OverflowError): return None - def parse_auth(header): """ Parse rfc2617 HTTP authentication header string (basic) and return (user,pass) tuple or None""" try: @@ -2149,6 +2565,17 @@ except ValueError: pass +def _parse_qsl(qs): + r = [] + for pair in qs.replace(';','&').split('&'): + if not pair: continue + nv = pair.split('=', 1) + if len(nv) != 2: nv.append('') + key = urlunquote(nv[0].replace('+', ' ')) + value = urlunquote(nv[1].replace('+', ' ')) + r.append((key, value)) + return r + def _lscmp(a, b): ''' Compares two strings in a cryptographically safe way: Runtime is not affected by length of common prefix. ''' @@ -2185,7 +2612,7 @@ def html_quote(string): ''' Escape and quote a string to be used as an HTTP attribute.''' - return '"%s"' % html_escape(string).replace('\n','%#10;')\ + return '"%s"' % html_escape(string).replace('\n',' ')\ .replace('\r',' ').replace('\t','	') @@ -2195,18 +2622,17 @@ takes optional keyword arguments. The output is best described by example:: a() -> '/a' - b(x, y) -> '/b/:x/:y' - c(x, y=5) -> '/c/:x' and '/c/:x/:y' - d(x=5, y=6) -> '/d' and '/d/:x' and '/d/:x/:y' + b(x, y) -> '/b/<x>/<y>' + c(x, y=5) -> '/c/<x>' and '/c/<x>/<y>' + d(x=5, y=6) -> '/d' and '/d/<x>' and '/d/<x>/<y>' """ - import inspect # Expensive module. Only import if necessary. path = '/' + func.__name__.replace('__','/').lstrip('/') - spec = inspect.getargspec(func) + spec = getargspec(func) argc = len(spec[0]) - len(spec[3] or []) - path += ('/:%s' * argc) % tuple(spec[0][:argc]) + path += ('/<%s>' * argc) % tuple(spec[0][:argc]) yield path for arg in spec[0][argc:]: - path += '/:%s' % arg + path += '/<%s>' % arg yield path @@ -2241,38 +2667,18 @@ return new_script_name, new_path_info -def validate(**vkargs): - """ - Validates and manipulates keyword arguments by user defined callables. - Handles ValueError and missing arguments by raising HTTPError(403). - """ - depr('Use route wildcard filters instead.') - def decorator(func): - @functools.wraps(func) - def wrapper(*args, **kargs): - for key, value in vkargs.items(): - if key not in kargs: - abort(403, 'Missing parameter: %s' % key) - try: - kargs[key] = value(kargs[key]) - except ValueError: - abort(403, 'Wrong parameter format for: %s' % key) - return func(*args, **kargs) - return wrapper - return decorator - - def auth_basic(check, realm="private", text="Access denied"): ''' Callback decorator to require HTTP auth (basic). TODO: Add route(check_auth=...) parameter. ''' def decorator(func): - def wrapper(*a, **ka): - user, password = request.auth or (None, None) - if user is None or not check(user, password): - response.headers['WWW-Authenticate'] = 'Basic realm="%s"' % realm - return HTTPError(401, text) - return func(*a, **ka) - return wrapper + def wrapper(*a, **ka): + user, password = request.auth or (None, None) + if user is None or not check(user, password): + err = HTTPError(401, text) + err.add_header('WWW-Authenticate', 'Basic realm="%s"' % realm) + return err + return func(*a, **ka) + return wrapper return decorator @@ -2311,8 +2717,8 @@ class ServerAdapter(object): quiet = False - def __init__(self, host='127.0.0.1', port=8080, **config): - self.options = config + def __init__(self, host='127.0.0.1', port=8080, **options): + self.options = options self.host = host self.port = int(port) @@ -2342,20 +2748,49 @@ class WSGIRefServer(ServerAdapter): - def run(self, handler): # pragma: no cover - from wsgiref.simple_server import make_server, WSGIRequestHandler - if self.quiet: - class QuietHandler(WSGIRequestHandler): - def log_request(*args, **kw): pass - self.options['handler_class'] = QuietHandler - srv = make_server(self.host, self.port, handler, **self.options) + def run(self, app): # pragma: no cover + from wsgiref.simple_server import WSGIRequestHandler, WSGIServer + from wsgiref.simple_server import make_server + import socket + + class FixedHandler(WSGIRequestHandler): + def address_string(self): # Prevent reverse DNS lookups please. + return self.client_address[0] + def log_request(*args, **kw): + if not self.quiet: + return WSGIRequestHandler.log_request(*args, **kw) + + handler_cls = self.options.get('handler_class', FixedHandler) + server_cls = self.options.get('server_class', WSGIServer) + + if ':' in self.host: # Fix wsgiref for IPv6 addresses. + if getattr(server_cls, 'address_family') == socket.AF_INET: + class server_cls(server_cls): + address_family = socket.AF_INET6 + + srv = make_server(self.host, self.port, app, server_cls, handler_cls) srv.serve_forever() class CherryPyServer(ServerAdapter): def run(self, handler): # pragma: no cover from cherrypy import wsgiserver - server = wsgiserver.CherryPyWSGIServer((self.host, self.port), handler) + self.options['bind_addr'] = (self.host, self.port) + self.options['wsgi_app'] = handler + + certfile = self.options.get('certfile') + if certfile: + del self.options['certfile'] + keyfile = self.options.get('keyfile') + if keyfile: + del self.options['keyfile'] + + server = wsgiserver.CherryPyWSGIServer(**self.options) + if certfile: + server.ssl_certificate = certfile + if keyfile: + server.ssl_private_key = keyfile + try: server.start() finally: @@ -2371,9 +2806,8 @@ class PasteServer(ServerAdapter): def run(self, handler): # pragma: no cover from paste import httpserver - if not self.quiet: - from paste.translogger import TransLogger - handler = TransLogger(handler) + from paste.translogger import TransLogger + handler = TransLogger(handler, setup_console_handler=(not self.quiet)) httpserver.serve(handler, host=self.host, port=str(self.port), **self.options) @@ -2413,7 +2847,7 @@ import tornado.wsgi, tornado.httpserver, tornado.ioloop container = tornado.wsgi.WSGIContainer(handler) server = tornado.httpserver.HTTPServer(container) - server.listen(port=self.port) + server.listen(port=self.port,address=self.host) tornado.ioloop.IOLoop.instance().start() @@ -2455,17 +2889,30 @@ class GeventServer(ServerAdapter): """ Untested. Options: - * `monkey` (default: True) fixes the stdlib to use greenthreads. * `fast` (default: False) uses libevent's http server, but has some issues: No streaming, no pipelining, no SSL. + * See gevent.wsgi.WSGIServer() documentation for more options. """ def run(self, handler): - from gevent import wsgi as wsgi_fast, pywsgi, monkey, local - if self.options.get('monkey', True): - if not threading.local is local.local: monkey.patch_all() - wsgi = wsgi_fast if self.options.get('fast') else pywsgi - log = None if self.quiet else 'default' - wsgi.WSGIServer((self.host, self.port), handler, log=log).serve_forever() + from gevent import wsgi, pywsgi, local + if not isinstance(threading.local(), local.local): + msg = "Bottle requires gevent.monkey.patch_all() (before import)" + raise RuntimeError(msg) + if not self.options.pop('fast', None): wsgi = pywsgi + self.options['log'] = None if self.quiet else 'default' + address = (self.host, self.port) + server = wsgi.WSGIServer(address, handler, **self.options) + if 'BOTTLE_CHILD' in os.environ: + import signal + signal.signal(signal.SIGINT, lambda s, f: server.stop()) + server.serve_forever() + + +class GeventSocketIOServer(ServerAdapter): + def run(self,handler): + from socketio import server + address = (self.host, self.port) + server.SocketIOServer(address, handler, **self.options).serve_forever() class GunicornServer(ServerAdapter): @@ -2539,6 +2986,7 @@ 'gunicorn': GunicornServer, 'eventlet': EventletServer, 'gevent': GeventServer, + 'geventSocketIO':GeventSocketIOServer, 'rocket': RocketServer, 'bjoern' : BjoernServer, 'auto': AutoServer, @@ -2590,7 +3038,7 @@ _debug = debug def run(app=None, server='wsgiref', host='127.0.0.1', port=8080, interval=1, reloader=False, quiet=False, plugins=None, - debug=False, **kargs): + debug=None, **kargs): """ Start a server instance. This method blocks until the server terminates. :param app: WSGI application or target string supported by @@ -2633,7 +3081,7 @@ return try: - _debug(debug) + if debug is not None: _debug(debug) app = app or default_app() if isinstance(app, basestring): app = load_app(app) @@ -2770,11 +3218,19 @@ def search(cls, name, lookup=[]): """ Search name in all directories specified in lookup. First without, then with common extensions. Return first hit. """ - if os.path.isfile(name): return name + if not lookup: + depr('The template lookup path list should not be empty.') #0.12 + lookup = ['.'] + + if os.path.isabs(name) and os.path.isfile(name): + depr('Absolute template path names are deprecated.') #0.12 + return os.path.abspath(name) + for spath in lookup: - fname = os.path.join(spath, name) - if os.path.isfile(fname): - return fname + spath = os.path.abspath(spath) + os.sep + fname = os.path.abspath(os.path.join(spath, name)) + if not fname.startswith(spath): continue + if os.path.isfile(fname): return fname for ext in cls.extensions: if os.path.isfile('%s.%s' % (fname, ext)): return '%s.%s' % (fname, ext) @@ -2799,8 +3255,8 @@ """ Render the template with the specified local variables and return a single byte or unicode string. If it is a byte string, the encoding must match self.encoding. This method must be thread-safe! - Local variables may be provided in dictionaries (*args) - or directly, as keywords (**kwargs). + Local variables may be provided in dictionaries (args) + or directly, as keywords (kwargs). """ raise NotImplementedError @@ -2845,7 +3301,7 @@ class Jinja2Template(BaseTemplate): - def prepare(self, filters=None, tests=None, **kwargs): + def prepare(self, filters=None, tests=None, globals={}, **kwargs): from jinja2 import Environment, FunctionLoader if 'prefix' in kwargs: # TODO: to be removed after a while raise RuntimeError('The keyword argument `prefix` has been removed. ' @@ -2853,6 +3309,7 @@ self.env = Environment(loader=FunctionLoader(self.loader), **kwargs) if filters: self.env.filters.update(filters) if tests: self.env.tests.update(tests) + if globals: self.env.globals.update(globals) if self.source: self.tpl = self.env.from_string(self.source) else: @@ -2871,184 +3328,247 @@ return f.read().decode(self.encoding) -class SimpleTALTemplate(BaseTemplate): - ''' Deprecated, do not use. ''' - def prepare(self, **options): - depr('The SimpleTAL template handler is deprecated'\ - ' and will be removed in 0.12') - from simpletal import simpleTAL - if self.source: - self.tpl = simpleTAL.compileHTMLTemplate(self.source) - else: - with open(self.filename, 'rb') as fp: - self.tpl = simpleTAL.compileHTMLTemplate(tonat(fp.read())) - - def render(self, *args, **kwargs): - from simpletal import simpleTALES - for dictarg in args: kwargs.update(dictarg) - context = simpleTALES.Context() - for k,v in self.defaults.items(): - context.addGlobal(k, v) - for k,v in kwargs.items(): - context.addGlobal(k, v) - output = StringIO() - self.tpl.expand(context, output) - return output.getvalue() - - class SimpleTemplate(BaseTemplate): - blocks = ('if', 'elif', 'else', 'try', 'except', 'finally', 'for', 'while', - 'with', 'def', 'class') - dedent_blocks = ('elif', 'else', 'except', 'finally') - - @lazy_attribute - def re_pytokens(cls): - ''' This matches comments and all kinds of quoted strings but does - NOT match comments (#...) within quoted strings. (trust me) ''' - return re.compile(r''' - (''(?!')|""(?!")|'{6}|"{6} # Empty strings (all 4 types) - |'(?:[^\\']|\\.)+?' # Single quotes (') - |"(?:[^\\"]|\\.)+?" # Double quotes (") - |'{3}(?:[^\\]|\\.|\n)+?'{3} # Triple-quoted strings (') - |"{3}(?:[^\\]|\\.|\n)+?"{3} # Triple-quoted strings (") - |\#.* # Comments - )''', re.VERBOSE) - - def prepare(self, escape_func=html_escape, noescape=False, **kwargs): + + def prepare(self, escape_func=html_escape, noescape=False, syntax=None, **ka): self.cache = {} enc = self.encoding self._str = lambda x: touni(x, enc) self._escape = lambda x: escape_func(touni(x, enc)) + self.syntax = syntax if noescape: self._str, self._escape = self._escape, self._str - @classmethod - def split_comment(cls, code): - """ Removes comments (#...) from python code. """ - if '#' not in code: return code - #: Remove comments only (leave quoted strings as they are) - subf = lambda m: '' if m.group(0)[0]=='#' else m.group(0) - return re.sub(cls.re_pytokens, subf, code) - @cached_property def co(self): return compile(self.code, self.filename or '<string>', 'exec') @cached_property def code(self): - stack = [] # Current Code indentation - lineno = 0 # Current line of code - ptrbuffer = [] # Buffer for printable strings and token tuple instances - codebuffer = [] # Buffer for generated python code - multiline = dedent = oneline = False - template = self.source or open(self.filename, 'rb').read() - - def yield_tokens(line): - for i, part in enumerate(re.split(r'\{\{(.*?)\}\}', line)): - if i % 2: - if part.startswith('!'): yield 'RAW', part[1:] - else: yield 'CMD', part - else: yield 'TXT', part - - def flush(): # Flush the ptrbuffer - if not ptrbuffer: return - cline = '' - for line in ptrbuffer: - for token, value in line: - if token == 'TXT': cline += repr(value) - elif token == 'RAW': cline += '_str(%s)' % value - elif token == 'CMD': cline += '_escape(%s)' % value - cline += ', ' - cline = cline[:-2] + '\\\n' - cline = cline[:-2] - if cline[:-1].endswith('\\\\\\\\\\n'): - cline = cline[:-7] + cline[-1] # 'nobr\\\\\n' --> 'nobr' - cline = '_printlist([' + cline + '])' - del ptrbuffer[:] # Do this before calling code() again - code(cline) - - def code(stmt): - for line in stmt.splitlines(): - codebuffer.append(' ' * len(stack) + line.strip()) - - for line in template.splitlines(True): - lineno += 1 - line = touni(line, self.encoding) - sline = line.lstrip() - if lineno <= 2: - m = re.match(r"%\s*#.*coding[:=]\s*([-\w.]+)", sline) - if m: self.encoding = m.group(1) - if m: line = line.replace('coding','coding (removed)') - if sline and sline[0] == '%' and sline[:2] != '%%': - line = line.split('%',1)[1].lstrip() # Full line following the % - cline = self.split_comment(line).strip() - cmd = re.split(r'[^a-zA-Z0-9_]', cline)[0] - flush() # You are actually reading this? Good luck, it's a mess :) - if cmd in self.blocks or multiline: - cmd = multiline or cmd - dedent = cmd in self.dedent_blocks # "else:" - if dedent and not oneline and not multiline: - cmd = stack.pop() - code(line) - oneline = not cline.endswith(':') # "if 1: pass" - multiline = cmd if cline.endswith('\\') else False - if not oneline and not multiline: - stack.append(cmd) - elif cmd == 'end' and stack: - code('#end(%s) %s' % (stack.pop(), line.strip()[3:])) - elif cmd == 'include': - p = cline.split(None, 2)[1:] - if len(p) == 2: - code("_=_include(%s, _stdout, %s)" % (repr(p[0]), p[1])) - elif p: - code("_=_include(%s, _stdout)" % repr(p[0])) - else: # Empty %include -> reverse of %rebase - code("_printlist(_base)") - elif cmd == 'rebase': - p = cline.split(None, 2)[1:] - if len(p) == 2: - code("globals()['_rebase']=(%s, dict(%s))" % (repr(p[0]), p[1])) - elif p: - code("globals()['_rebase']=(%s, {})" % repr(p[0])) - else: - code(line) - else: # Line starting with text (not '%') or '%%' (escaped) - if line.strip().startswith('%%'): - line = line.replace('%%', '%', 1) - ptrbuffer.append(yield_tokens(line)) - flush() - return '\n'.join(codebuffer) + '\n' - - def subtemplate(self, _name, _stdout, *args, **kwargs): - for dictarg in args: kwargs.update(dictarg) + source = self.source + if not source: + with open(self.filename, 'rb') as f: + source = f.read() + try: + source, encoding = touni(source), 'utf8' + except UnicodeError: + depr('Template encodings other than utf8 are no longer supported.') #0.11 + source, encoding = touni(source, 'latin1'), 'latin1' + parser = StplParser(source, encoding=encoding, syntax=self.syntax) + code = parser.translate() + self.encoding = parser.encoding + return code + + def _rebase(self, _env, _name=None, **kwargs): + if _name is None: + depr('Rebase function called without arguments.' + ' You were probably looking for {{base}}?', True) #0.12 + _env['_rebase'] = (_name, kwargs) + + def _include(self, _env, _name=None, **kwargs): + if _name is None: + depr('Rebase function called without arguments.' + ' You were probably looking for {{base}}?', True) #0.12 + env = _env.copy() + env.update(kwargs) if _name not in self.cache: self.cache[_name] = self.__class__(name=_name, lookup=self.lookup) - return self.cache[_name].execute(_stdout, kwargs) - - def execute(self, _stdout, *args, **kwargs): - for dictarg in args: kwargs.update(dictarg) + return self.cache[_name].execute(env['_stdout'], env) + + def execute(self, _stdout, kwargs): env = self.defaults.copy() + env.update(kwargs) env.update({'_stdout': _stdout, '_printlist': _stdout.extend, - '_include': self.subtemplate, '_str': self._str, - '_escape': self._escape, 'get': env.get, - 'setdefault': env.setdefault, 'defined': env.__contains__}) - env.update(kwargs) + 'include': functools.partial(self._include, env), + 'rebase': functools.partial(self._rebase, env), '_rebase': None, + '_str': self._str, '_escape': self._escape, 'get': env.get, + 'setdefault': env.setdefault, 'defined': env.__contains__ }) eval(self.co, env) - if '_rebase' in env: - subtpl, rargs = env['_rebase'] - rargs['_base'] = _stdout[:] #copy stdout + if env.get('_rebase'): + subtpl, rargs = env.pop('_rebase') + rargs['base'] = ''.join(_stdout) #copy stdout del _stdout[:] # clear stdout - return self.subtemplate(subtpl,_stdout,rargs) + return self._include(env, subtpl, **rargs) return env def render(self, *args, **kwargs): """ Render the template using keyword arguments as local variables. """ - for dictarg in args: kwargs.update(dictarg) - stdout = [] - self.execute(stdout, kwargs) + env = {}; stdout = [] + for dictarg in args: env.update(dictarg) + env.update(kwargs) + self.execute(stdout, env) return ''.join(stdout) +class StplSyntaxError(TemplateError): pass + + +class StplParser(object): + ''' Parser for stpl templates. ''' + _re_cache = {} #: Cache for compiled re patterns + # This huge pile of voodoo magic splits python code into 8 different tokens. + # 1: All kinds of python strings (trust me, it works) + _re_tok = '((?m)[urbURB]?(?:\'\'(?!\')|""(?!")|\'{6}|"{6}' \ + '|\'(?:[^\\\\\']|\\\\.)+?\'|"(?:[^\\\\"]|\\\\.)+?"' \ + '|\'{3}(?:[^\\\\]|\\\\.|\\n)+?\'{3}' \ + '|"{3}(?:[^\\\\]|\\\\.|\\n)+?"{3}))' + _re_inl = _re_tok.replace('|\\n','') # We re-use this string pattern later + # 2: Comments (until end of line, but not the newline itself) + _re_tok += '|(#.*)' + # 3,4: Keywords that start or continue a python block (only start of line) + _re_tok += '|^([ \\t]*(?:if|for|while|with|try|def|class)\\b)' \ + '|^([ \\t]*(?:elif|else|except|finally)\\b)' + # 5: Our special 'end' keyword (but only if it stands alone) + _re_tok += '|((?:^|;)[ \\t]*end[ \\t]*(?=(?:%(block_close)s[ \\t]*)?\\r?$|;|#))' + # 6: A customizable end-of-code-block template token (only end of line) + _re_tok += '|(%(block_close)s[ \\t]*(?=$))' + # 7: And finally, a single newline. The 8th token is 'everything else' + _re_tok += '|(\\r?\\n)' + # Match the start tokens of code areas in a template + _re_split = '(?m)^[ \t]*(\\\\?)((%(line_start)s)|(%(block_start)s))(%%?)' + # Match inline statements (may contain python strings) + _re_inl = '%%(inline_start)s((?:%s|[^\'"\n]*?)+)%%(inline_end)s' % _re_inl + + default_syntax = '<% %> % {{ }}' + + def __init__(self, source, syntax=None, encoding='utf8'): + self.source, self.encoding = touni(source, encoding), encoding + self.set_syntax(syntax or self.default_syntax) + self.code_buffer, self.text_buffer = [], [] + self.lineno, self.offset = 1, 0 + self.indent, self.indent_mod = 0, 0 + + def get_syntax(self): + ''' Tokens as a space separated string (default: <% %> % {{ }}) ''' + return self._syntax + + def set_syntax(self, syntax): + self._syntax = syntax + self._tokens = syntax.split() + if not syntax in self._re_cache: + names = 'block_start block_close line_start inline_start inline_end' + etokens = map(re.escape, self._tokens) + pattern_vars = dict(zip(names.split(), etokens)) + patterns = (self._re_split, self._re_tok, self._re_inl) + patterns = [re.compile(p%pattern_vars) for p in patterns] + self._re_cache[syntax] = patterns + self.re_split, self.re_tok, self.re_inl = self._re_cache[syntax] + + syntax = property(get_syntax, set_syntax) + + def translate(self): + if self.offset: raise RuntimeError('Parser is a one time instance.') + while True: + m = self.re_split.search(self.source[self.offset:]) + if m: + text = self.source[self.offset:self.offset+m.start()] + self.text_buffer.append(text) + self.offset += m.end() + if m.group(1): # New escape syntax + line, sep, _ = self.source[self.offset:].partition('\n') + self.text_buffer.append(m.group(2)+m.group(5)+line+sep) + self.offset += len(line+sep)+1 + continue + elif m.group(5): # Old escape syntax + depr('Escape code lines with a backslash.') #0.12 + line, sep, _ = self.source[self.offset:].partition('\n') + self.text_buffer.append(m.group(2)+line+sep) + self.offset += len(line+sep)+1 + continue + self.flush_text() + self.read_code(multiline=bool(m.group(4))) + else: break + self.text_buffer.append(self.source[self.offset:]) + self.flush_text() + return ''.join(self.code_buffer) + + def read_code(self, multiline): + code_line, comment = '', '' + while True: + m = self.re_tok.search(self.source[self.offset:]) + if not m: + code_line += self.source[self.offset:] + self.offset = len(self.source) + self.write_code(code_line.strip(), comment) + return + code_line += self.source[self.offset:self.offset+m.start()] + self.offset += m.end() + _str, _com, _blk1, _blk2, _end, _cend, _nl = m.groups() + if code_line and (_blk1 or _blk2): # a if b else c + code_line += _blk1 or _blk2 + continue + if _str: # Python string + code_line += _str + elif _com: # Python comment (up to EOL) + comment = _com + if multiline and _com.strip().endswith(self._tokens[1]): + multiline = False # Allow end-of-block in comments + elif _blk1: # Start-block keyword (if/for/while/def/try/...) + code_line, self.indent_mod = _blk1, -1 + self.indent += 1 + elif _blk2: # Continue-block keyword (else/elif/except/...) + code_line, self.indent_mod = _blk2, -1 + elif _end: # The non-standard 'end'-keyword (ends a block) + self.indent -= 1 + elif _cend: # The end-code-block template token (usually '%>') + if multiline: multiline = False + else: code_line += _cend + else: # \n + self.write_code(code_line.strip(), comment) + self.lineno += 1 + code_line, comment, self.indent_mod = '', '', 0 + if not multiline: + break + + def flush_text(self): + text = ''.join(self.text_buffer) + del self.text_buffer[:] + if not text: return + parts, pos, nl = [], 0, '\\\n'+' '*self.indent + for m in self.re_inl.finditer(text): + prefix, pos = text[pos:m.start()], m.end() + if prefix: + parts.append(nl.join(map(repr, prefix.splitlines(True)))) + if prefix.endswith('\n'): parts[-1] += nl + parts.append(self.process_inline(m.group(1).strip())) + if pos < len(text): + prefix = text[pos:] + lines = prefix.splitlines(True) + if lines[-1].endswith('\\\\\n'): lines[-1] = lines[-1][:-3] + elif lines[-1].endswith('\\\\\r\n'): lines[-1] = lines[-1][:-4] + parts.append(nl.join(map(repr, lines))) + code = '_printlist((%s,))' % ', '.join(parts) + self.lineno += code.count('\n')+1 + self.write_code(code) + + def process_inline(self, chunk): + if chunk[0] == '!': return '_str(%s)' % chunk[1:] + return '_escape(%s)' % chunk + + def write_code(self, line, comment=''): + line, comment = self.fix_backward_compatibility(line, comment) + code = ' ' * (self.indent+self.indent_mod) + code += line.lstrip() + comment + '\n' + self.code_buffer.append(code) + + def fix_backward_compatibility(self, line, comment): + parts = line.strip().split(None, 2) + if parts and parts[0] in ('include', 'rebase'): + depr('The include and rebase keywords are functions now.') #0.12 + if len(parts) == 1: return "_printlist([base])", comment + elif len(parts) == 2: return "_=%s(%r)" % tuple(parts), comment + else: return "_=%s(%r, %s)" % tuple(parts), comment + if self.lineno <= 2 and not line.strip() and 'coding' in comment: + m = re.match(r"#.*coding[:=]\s*([-\w.]+)", comment) + if m: + depr('PEP263 encoding strings in templates are deprecated.') #0.12 + enc = m.group(1) + self.source = self.source.encode(self.encoding).decode(enc) + self.encoding = enc + return line, comment.replace('coding','coding*') + return line, comment + + def template(*args, **kwargs): ''' Get a rendered template as a string iterator. @@ -3057,26 +3577,26 @@ or directly (as keyword arguments). ''' tpl = args[0] if args else None - template_adapter = kwargs.pop('template_adapter', SimpleTemplate) - if tpl not in TEMPLATES or DEBUG: + adapter = kwargs.pop('template_adapter', SimpleTemplate) + lookup = kwargs.pop('template_lookup', TEMPLATE_PATH) + tplid = (id(lookup), tpl) + if tplid not in TEMPLATES or DEBUG: settings = kwargs.pop('template_settings', {}) - lookup = kwargs.pop('template_lookup', TEMPLATE_PATH) - if isinstance(tpl, template_adapter): - TEMPLATES[tpl] = tpl - if settings: TEMPLATES[tpl].prepare(**settings) + if isinstance(tpl, adapter): + TEMPLATES[tplid] = tpl + if settings: TEMPLATES[tplid].prepare(**settings) elif "\n" in tpl or "{" in tpl or "%" in tpl or '$' in tpl: - TEMPLATES[tpl] = template_adapter(source=tpl, lookup=lookup, **settings) + TEMPLATES[tplid] = adapter(source=tpl, lookup=lookup, **settings) else: - TEMPLATES[tpl] = template_adapter(name=tpl, lookup=lookup, **settings) - if not TEMPLATES[tpl]: + TEMPLATES[tplid] = adapter(name=tpl, lookup=lookup, **settings) + if not TEMPLATES[tplid]: abort(500, 'Template (%s) not found' % tpl) for dictarg in args[1:]: kwargs.update(dictarg) - return TEMPLATES[tpl].render(kwargs) + return TEMPLATES[tplid].render(kwargs) mako_template = functools.partial(template, template_adapter=MakoTemplate) cheetah_template = functools.partial(template, template_adapter=CheetahTemplate) jinja2_template = functools.partial(template, template_adapter=Jinja2Template) -simpletal_template = functools.partial(template, template_adapter=SimpleTALTemplate) def view(tpl_name, **defaults): @@ -3097,6 +3617,8 @@ tplvars = defaults.copy() tplvars.update(result) return template(tpl_name, **tplvars) + elif result is None: + return template(tpl_name, defaults) return result return wrapper return decorator @@ -3104,7 +3626,6 @@ mako_view = functools.partial(view, template_adapter=MakoTemplate) cheetah_view = functools.partial(view, template_adapter=CheetahTemplate) jinja2_view = functools.partial(view, template_adapter=Jinja2Template) -simpletal_view = functools.partial(view, template_adapter=SimpleTALTemplate) @@ -3124,6 +3645,7 @@ #: A dict to map HTTP status codes (e.g. 404) to phrases (e.g. 'Not Found') HTTP_CODES = httplib.responses HTTP_CODES[418] = "I'm a teapot" # RFC 2324 +HTTP_CODES[422] = "Unprocessable Entity" # RFC 4918 HTTP_CODES[428] = "Precondition Required" HTTP_CODES[429] = "Too Many Requests" HTTP_CODES[431] = "Request Header Fields Too Large" @@ -3134,11 +3656,10 @@ ERROR_PAGE_TEMPLATE = """ %%try: %%from %s import DEBUG, HTTP_CODES, request, touni - %%status_name = HTTP_CODES.get(e.status, 'Unknown').title() <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"> <html> <head> - <title>Error {{e.status}}: {{status_name}}</title> + <title>Error: {{e.status}}</title> <style type="text/css"> html {background-color: #eee; font-family: sans;} body {background-color: #fff; border: 1px solid #ddd; @@ -3147,10 +3668,10 @@ </style> </head> <body> - <h1>Error {{e.status}}: {{status_name}}</h1> + <h1>Error: {{e.status}}</h1> <p>Sorry, the requested URL <tt>{{repr(request.url)}}</tt> caused an error:</p> - <pre>{{e.output}}</pre> + <pre>{{e.body}}</pre> %%if DEBUG and e.exception: <h2>Exception:</h2> <pre>{{repr(e.exception)}}</pre> @@ -3167,7 +3688,7 @@ %%end """ % __name__ -#: A thread-safe instance of :class:`LocalRequest`. If accessed from within a +#: A thread-safe instance of :class:`LocalRequest`. If accessed from within a #: request callback, this instance always refers to the *current* request #: (even on a multithreaded server). request = LocalRequest() @@ -3186,7 +3707,7 @@ #: A virtual package that redirects import statements. #: Example: ``import bottle.ext.sqlite`` actually imports `bottle_sqlite`. -ext = _ImportRedirect(__name__+'.ext', 'bottle_%s').module +ext = _ImportRedirect('bottle.ext' if __name__ == '__main__' else __name__+".ext", 'bottle_%s').module if __name__ == '__main__': opt, args, parser = _cmd_options, _cmd_args, _cmd_parser @@ -3202,10 +3723,11 @@ sys.modules.setdefault('bottle', sys.modules['__main__']) host, port = (opt.bind or 'localhost'), 8080 - if ':' in host: + if ':' in host and host.rfind(']') < host.rfind(':'): host, port = host.rsplit(':', 1) - - run(args[0], host=host, port=port, server=opt.server, + host = host.strip('[]') + + run(args[0], host=host, port=int(port), server=opt.server, reloader=opt.reload, plugins=opt.plugin, debug=opt.debug)