Server IP : 66.29.132.122 / Your IP : 3.133.129.64 Web Server : LiteSpeed System : Linux business142.web-hosting.com 4.18.0-553.lve.el8.x86_64 #1 SMP Mon May 27 15:27:34 UTC 2024 x86_64 User : admazpex ( 531) PHP Version : 7.2.34 Disable Function : NONE MySQL : OFF | cURL : ON | WGET : ON | Perl : ON | Python : ON | Sudo : OFF | Pkexec : OFF Directory : /proc/self/root/proc/self/root/proc/thread-self/root/proc/thread-self/root/opt/cloudlinux/venv/lib64/python3.11/site-packages/raven/ |
Upload File : |
from __future__ import absolute_import import os import logging try: from collections.abc import Mapping except ImportError: # Python < 3.3 from collections import Mapping from time import time from types import FunctionType from raven.utils import once from raven.utils.encoding import to_unicode from raven.utils.compat import iteritems, get_code, text_type, string_types CATEGORY_MAX_LENGTH = 128 LEVEL_MAX_LENGTH = 16 special_logging_handlers = [] special_logger_handlers = {} logger = logging.getLogger('raven') def event_payload_considered_equal(a, b): return ( a['type'] == b['type'] and a['level'] == b['level'] and a['message'] == b['message'] and a['category'] == b['category'] and a['data'] == b['data'] ) class BreadcrumbBuffer(object): def __init__(self, limit=100, message_max_length=1024): self.buffer = [] self.limit = limit self.message_max_length = message_max_length def record(self, timestamp=None, level=None, message=None, category=None, data=None, type=None, processor=None): if not (message or data or processor): raise ValueError('You must pass either `message`, `data`, ' 'or `processor`') if timestamp is None: timestamp = time() # we format here to ensure we dont bloat memory due to message size result = (self.format({ 'type': type or 'default', 'timestamp': float(timestamp), 'level': level, # hardcode message length to prevent huge crumbs 'message': message, 'category': category, # TODO(dcramer): we should trim data 'data': data, }), processor) self.buffer.append(result) del self.buffer[:-self.limit] def clear(self): del self.buffer[:] def format(self, result): result['message'] = to_unicode(result['message'])[:self.message_max_length] if result['message'] else None result['category'] = to_unicode(result['category'])[:CATEGORY_MAX_LENGTH] if result['category'] else None result['level'] = to_unicode(result['level'])[:LEVEL_MAX_LENGTH].lower() if result['level'] else None return result def get_buffer(self): rv = [] for idx, (payload, processor) in enumerate(self.buffer): if processor is not None: try: processor(payload) except Exception: raise logger.exception('Failed to process breadcrumbs. Ignored') payload = None else: # we format here to ensure we dont bloat memory due to message size payload = self.format(payload) if payload else None self.buffer[idx] = (payload, None) elif payload is not None: payload = self.format(payload) if payload is not None and \ (not rv or not event_payload_considered_equal(rv[-1], payload)): rv.append(payload) return rv class BlackholeBreadcrumbBuffer(BreadcrumbBuffer): def record(self, *args, **kwargs): pass def make_buffer(enabled=True): if enabled: return BreadcrumbBuffer() return BlackholeBreadcrumbBuffer() def record_breadcrumb(type, *args, **kwargs): # Legacy alias kwargs['type'] = type return record(*args, **kwargs) def record(message=None, timestamp=None, level=None, category=None, data=None, type=None, processor=None): """Records a breadcrumb for all active clients. This is what integration code should use rather than invoking the `captureBreadcrumb` method on a specific client. """ if timestamp is None: timestamp = time() for ctx in raven.context.get_active_contexts(): ctx.breadcrumbs.record(timestamp, level, message, category, data, type, processor) def _record_log_breadcrumb(logger, level, msg, *args, **kwargs): for handler in special_logging_handlers: if handler(logger, level, msg, args, kwargs): return handler = special_logger_handlers.get(logger.name) if handler is not None and handler(logger, level, msg, args, kwargs): return def processor(data): formatted_msg = msg format_args = args # Extract 'extra' key from kwargs and merge into data extra = kwargs.pop('extra', {}) data_value = kwargs data_value.update(extra) if args and len(args) == 1 and isinstance(args[0], Mapping) and args[0]: format_args = args[0] data_value.update(format_args) # If people log bad things, this can happen. Then just don't do # anything. try: formatted_msg = text_type(msg) if format_args: formatted_msg = msg % format_args except Exception: pass # We do not want to include exc_info as argument because it often # lies (set to a constant value like 1 or True) or even if it's a # tuple it will not be particularly useful for us as we cannot # process it anyways. kwargs.pop('exc_info', None) data.update({ 'message': formatted_msg, 'category': logger.name, 'level': logging.getLevelName(level).lower(), 'data': data_value, }) record(processor=processor) def _wrap_logging_method(meth, level=None): if not isinstance(meth, FunctionType): func = meth.im_func else: func = meth # We were patched for raven before if getattr(func, '__patched_for_raven__', False): return if level is None: args = ('level', 'msg') fwd = 'level, msg' else: args = ('msg',) fwd = '%d, msg' % level code = get_code(func) logging_srcfile = logging._srcfile if logging_srcfile is None: logging_srcfile = os.path.normpath( logging.currentframe.__code__.co_filename ) # This requires a bit of explanation why we're doing this. Due to how # logging itself works we need to pretend that the method actually was # created within the logging module. There are a few ways to detect # this and we fake all of them: we use the same function globals (the # one from the logging module), we create it entirely there which # means that also the filename is set correctly. This fools the # detection code in logging and it makes logging itself skip past our # code when determining the code location. # # Because we point the globals to the logging module we now need to # refer to our own functions (original and the crumb recording # function) through a closure instead of the global scope. # # We also add a lot of newlines in front of the code so that the # code location lines up again in case someone runs inspect.getsource # on the function. ns = {} eval(compile('''%(offset)sif 1: def factory(original, record_crumb): def %(name)s(self, %(args)s, *args, **kwargs): record_crumb(self, %(fwd)s, *args, **kwargs) return original(self, %(args)s, *args, **kwargs) return %(name)s \n''' % { 'offset': '\n' * (code.co_firstlineno - 3), 'name': func.__name__, 'args': ', '.join(args), 'fwd': fwd, 'level': level, }, logging_srcfile, 'exec'), logging.__dict__, ns) new_func = ns['factory'](meth, _record_log_breadcrumb) new_func.__doc__ = func.__doc__ assert code.co_firstlineno == get_code(func).co_firstlineno # In theory this should already be set correctly, but in some cases # it is not. So override it. new_func.__module__ == func.__module__ new_func.__name__ == func.__name__ new_func.__patched_for_raven__ = True return new_func def _patch_logger(): cls = logging.Logger methods = { 'debug': logging.DEBUG, 'info': logging.INFO, 'warning': logging.WARNING, 'warn': logging.WARN, 'error': logging.ERROR, 'exception': logging.ERROR, 'critical': logging.CRITICAL, 'fatal': logging.FATAL } for method_name, level in iteritems(methods): new_func = _wrap_logging_method( getattr(cls, method_name), level) setattr(logging.Logger, method_name, new_func) logging.Logger.log = _wrap_logging_method( logging.Logger.log) @once def install_logging_hook(): """Installs the logging hook if it was not installed yet. Otherwise does nothing. """ _patch_logger() def ignore_logger(name_or_logger, allow_level=None): """Ignores a logger during breadcrumb recording. """ def handler(logger, level, msg, args, kwargs): if allow_level is not None and \ level >= allow_level: return False return True register_special_log_handler(name_or_logger, handler) def register_special_log_handler(name_or_logger, callback): """Registers a callback for log handling. The callback is invoked with given arguments: `logger`, `level`, `msg`, `args` and `kwargs` which are the values passed to the logging system. If the callback returns true value the default handling is disabled. Only one callback can be registered per one logger name. Logger tree is not traversed so calling this method with `spammy_module` argument will not silence messages from `spammy_module.child`. """ if isinstance(name_or_logger, string_types): name = name_or_logger else: name = name_or_logger.name special_logger_handlers[name] = callback def register_logging_handler(callback): """Registers a callback for log handling. The callback is invoked with given arguments: `logger`, `level`, `msg`, `args` and `kwargs` which are the values passed to the logging system. If the callback returns true value the default handling is disabled. Registering multiple handlers is allowed. """ special_logging_handlers.append(callback) hooked_libraries = {} def libraryhook(name): def decorator(f): f = once(f) hooked_libraries[name] = f return f return decorator @libraryhook('requests') def _hook_requests(): try: from requests.sessions import Session except ImportError: return real_send = Session.send def send(self, request, *args, **kwargs): def _record_request(response): record(type='http', category='requests', data={ 'url': request.url, 'method': request.method, 'status_code': response and response.status_code or None, 'reason': response and response.reason or None, }) try: resp = real_send(self, request, *args, **kwargs) except Exception: _record_request(None) raise else: _record_request(resp) return resp Session.send = send ignore_logger('requests.packages.urllib3.connectionpool', allow_level=logging.WARNING) @libraryhook('httplib') def _install_httplib(): try: from httplib import HTTPConnection except ImportError: from http.client import HTTPConnection real_putrequest = HTTPConnection.putrequest real_getresponse = HTTPConnection.getresponse def putrequest(self, method, url, *args, **kwargs): self._raven_status_dict = status = {} host = self.host port = self.port default_port = self.default_port def processor(data): real_url = url if not real_url.startswith(('http://', 'https://')): real_url = '%s://%s%s%s' % ( default_port == 443 and 'https' or 'http', host, port != default_port and ':%s' % port or '', url, ) data['data'] = { 'url': real_url, 'method': method, } data['data'].update(status) return data record(type='http', category='requests', processor=processor) return real_putrequest(self, method, url, *args, **kwargs) def getresponse(self, *args, **kwargs): rv = real_getresponse(self, *args, **kwargs) status = getattr(self, '_raven_status_dict', None) if status is not None and 'status_code' not in status: status['status_code'] = rv.status status['reason'] = rv.reason return rv HTTPConnection.putrequest = putrequest HTTPConnection.getresponse = getresponse def hook_libraries(libraries): if libraries is None: libraries = hooked_libraries.keys() for lib in libraries: func = hooked_libraries.get(lib) if func is None: raise RuntimeError('Unknown library %r for hooking' % lib) func() import raven.context