-#!/usr/bin/python
-
-#----------------------------------------------------------------------
-# Copyright (c) 2008 Board of Trustees, Princeton University
-#
-# Permission is hereby granted, free of charge, to any person obtaining
-# a copy of this software and/or hardware specification (the "Work") to
-# deal in the Work without restriction, including without limitation the
-# rights to use, copy, modify, merge, publish, distribute, sublicense,
-# and/or sell copies of the Work, and to permit persons to whom the Work
-# is furnished to do so, subject to the following conditions:
-#
-# The above copyright notice and this permission notice shall be
-# included in all copies or substantial portions of the Work.
-#
-# THE WORK IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
-# OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
-# NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
-# HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY,
-# WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-# OUT OF OR IN CONNECTION WITH THE WORK OR THE USE OR OTHER DEALINGS
-# IN THE WORK.
-#----------------------------------------------------------------------
-
-from __future__ import print_function
+#!/usr/bin/env python3
+
+"""
+A reroutable logger that can handle deep tracebacks
+
+Requirements:
+
+* for legacy, we want all our code to just do:
+
+ from sfa.util.sfalogging import logger
+ ...
+ logger.info('blabla')
+
+* depending on whether the code runs (a) inside the server,
+ (b) as part of sfa-import, or (c) as part of the sfi CLI,
+ we want these messages to be directed in different places
+
+* also because troubleshooting is very painful, we need a better way
+ to report stacks when an exception occurs.
+
+Implementation:
+
+* we use a single unique logger name 'sfa' (wrt getLogger()),
+ and provide an auxiliary function `init_logger()` that
+ accepts for its `context` parameter one of :
+ `server`, `import` `sfi` or `console`
+ It will then reconfigure the 'sfa' logger to do the right thing
+
+* also we create our own subclass of loggers, and install it
+ with logging.setLoggerClass(), so we can add our own customized
+ `log_exc()` method
+
+"""
+
+# pylint: disable=c0111, c0103, w1201
+
+
import os
+import os.path
import sys
import traceback
import logging
import logging.handlers
+import logging.config
+
+# so that users of this module don't need to import logging
+from logging import (CRITICAL, ERROR, WARNING, INFO, DEBUG)
-CRITICAL = logging.CRITICAL
-ERROR = logging.ERROR
-WARNING = logging.WARNING
-INFO = logging.INFO
-DEBUG = logging.DEBUG
-
-# a logger that can handle tracebacks
-
-
-class _SfaLogger:
-
- def __init__(self, logfile=None, loggername=None, level=logging.INFO):
- # default is to locate loggername from the logfile if avail.
- if not logfile:
- # loggername='console'
- # handler=logging.StreamHandler()
- #handler.setFormatter(logging.Formatter("%(levelname)s %(message)s"))
- logfile = "/var/log/sfa.log"
-
- if not loggername:
- loggername = os.path.basename(logfile)
- try:
- handler = logging.handlers.RotatingFileHandler(
- logfile, maxBytes=1000000, backupCount=5)
- except IOError:
- # This is usually a permissions error because the file is
- # owned by root, but httpd is trying to access it.
- tmplogfile = os.path.join(os.getenv("TMPDIR",
- os.getenv("TMP", os.path.normpath("/tmp"))),
- os.path.basename(logfile))
- tmplogfile = os.path.normpath(tmplogfile)
-
- tmpdir = os.path.dirname(tmplogfile)
- if tmpdir and tmpdir != "" and not os.path.exists(tmpdir):
- os.makedirs(tmpdir)
-
- # In strange uses, 2 users on same machine might use same code,
- # meaning they would clobber each others files
- # We could (a) rename the tmplogfile, or (b)
- # just log to the console in that case.
- # Here we default to the console.
- if os.path.exists(tmplogfile) and not os.access(tmplogfile, os.W_OK):
- loggername = loggername + "-console"
- handler = logging.StreamHandler()
- else:
- handler = logging.handlers.RotatingFileHandler(
- tmplogfile, maxBytes=1000000, backupCount=5)
- handler.setFormatter(logging.Formatter(
- "%(asctime)s - %(levelname)s - %(message)s"))
- self.logger = logging.getLogger(loggername)
- self.logger.setLevel(level)
- # check if logger already has the handler we're about to add
- handler_exists = False
- for l_handler in self.logger.handlers:
- if l_handler.baseFilename == handler.baseFilename and \
- l_handler.level == handler.level:
- handler_exists = True
-
- if not handler_exists:
- self.logger.addHandler(handler)
-
- self.loggername = loggername
-
- def setLevel(self, level):
- self.logger.setLevel(level)
+
+class SfaLogger(logging.getLoggerClass()):
+ """
+ a rewrite of old _SfaLogger class that was way too cumbersome
+ keep this as much as possible though
+ """
# shorthand to avoid having to import logging all over the place
def setLevelDebug(self):
- self.logger.setLevel(logging.DEBUG)
+ self.setLevel(DEBUG)
def debugEnabled(self):
- return self.logger.getEffectiveLevel() == logging.DEBUG
+ return self.getEffectiveLevel() == logging.DEBUG
# define a verbose option with s/t like
- # parser.add_option("-v", "--verbose", action="count", dest="verbose", default=0)
+ # parser.add_option("-v", "--verbose", action="count",
+ # dest="verbose", default=0)
# and pass the coresponding options.verbose to this method to adjust level
def setLevelFromOptVerbose(self, verbose):
if verbose == 0:
- self.logger.setLevel(logging.WARNING)
+ self.setLevel(logging.WARNING)
elif verbose == 1:
- self.logger.setLevel(logging.INFO)
+ self.setLevel(logging.INFO)
elif verbose >= 2:
- self.logger.setLevel(logging.DEBUG)
- # in case some other code needs a boolean
+ self.setLevel(logging.DEBUG)
- def getBoolVerboseFromOpt(self, verbose):
+ # in case some other code needs a boolean
+ @staticmethod
+ def getBoolVerboseFromOpt(verbose):
return verbose >= 1
- def getBoolDebugFromOpt(self, verbose):
+ @staticmethod
+ def getBoolDebugFromOpt(verbose):
return verbose >= 2
- ####################
- def info(self, msg):
- self.logger.info(msg)
-
- def debug(self, msg):
- self.logger.debug(msg)
+ def log_exc(self, message, limit=100):
+ """
+ standard logger has an exception() method but this will
+ dump the stack only between the frames
+ (1) that does `raise` and (2) the one that does `except`
- def warn(self, msg):
- self.logger.warn(msg)
+ log_exc() has a limit argument that allows to see deeper than that
- # some code is using logger.warn(), some is using logger.warning()
- def warning(self, msg):
- self.logger.warning(msg)
-
- def error(self, msg):
- self.logger.error(msg)
-
- def critical(self, msg):
- self.logger.critical(msg)
-
- # logs an exception - use in an except statement
- def log_exc(self, message):
+ use limit=None to get the same behaviour as exception()
+ """
self.error("%s BEG TRACEBACK" % message + "\n" +
- traceback.format_exc().strip("\n"))
+ traceback.format_exc(limit=limit).strip("\n"))
self.error("%s END TRACEBACK" % message)
- def log_exc_critical(self, message):
- self.critical("%s BEG TRACEBACK" % message + "\n" +
- traceback.format_exc().strip("\n"))
- self.critical("%s END TRACEBACK" % message)
-
# for investigation purposes, can be placed anywhere
- def log_stack(self, message):
- to_log = "".join(traceback.format_stack())
+ def log_stack(self, message, limit=100):
+ to_log = "".join(traceback.format_stack(limit=limit))
self.info("%s BEG STACK" % message + "\n" + to_log)
self.info("%s END STACK" % message)
- def enable_console(self, stream=sys.stdout):
+ def enable_console(self):
formatter = logging.Formatter("%(message)s")
- handler = logging.StreamHandler(stream)
+ handler = logging.StreamHandler(sys.stdout)
handler.setFormatter(formatter)
- self.logger.addHandler(handler)
-
-
-info_logger = _SfaLogger(loggername='info', level=logging.INFO)
-debug_logger = _SfaLogger(loggername='debug', level=logging.DEBUG)
-warn_logger = _SfaLogger(loggername='warning', level=logging.WARNING)
-error_logger = _SfaLogger(loggername='error', level=logging.ERROR)
-critical_logger = _SfaLogger(loggername='critical', level=logging.CRITICAL)
-logger = info_logger
-sfi_logger = _SfaLogger(logfile=os.path.expanduser("~/.sfi/") + 'sfi.log',
- loggername='sfilog', level=logging.DEBUG)
-########################################
-import time
-
-
-def profile(logger):
- """
- Prints the runtime of the specified callable. Use as a decorator, e.g.,
-
- @profile(logger)
- def foo(...):
- ...
- """
- def logger_profile(callable):
- def wrapper(*args, **kwds):
- start = time.time()
- result = callable(*args, **kwds)
- end = time.time()
- args = map(str, args)
- args += ["%s = %s" % (name, str(value))
- for (name, value) in kwds.iteritems()]
- # should probably use debug, but then debug is not always enabled
- logger.info("PROFILED %s (%s): %.02f s" %
- (callable.__name__, ", ".join(args), end - start))
- return result
- return wrapper
- return logger_profile
-
-
-if __name__ == '__main__':
- print('testing sfalogging into logger.log')
- logger1 = _SfaLogger('logger.log', loggername='std(info)')
- logger2 = _SfaLogger('logger.log', loggername='error', level=logging.ERROR)
- logger3 = _SfaLogger('logger.log', loggername='debug', level=logging.DEBUG)
-
- for logger, msg in ((logger1, "std(info)"), (logger2, "error"), (logger3, "debug")):
-
- print("====================", msg, logger.logger.handlers)
-
- logger.enable_console()
- logger.critical("logger.critical")
- logger.error("logger.error")
- logger.warn("logger.warning")
- logger.info("logger.info")
- logger.debug("logger.debug")
- logger.setLevel(logging.DEBUG)
- logger.debug("logger.debug again")
-
- @profile(logger)
- def sleep(seconds=1):
- time.sleep(seconds)
-
- logger.info('console.info')
- sleep(0.5)
- logger.setLevel(logging.DEBUG)
- sleep(0.25)
+ self.addHandler(handler)
+
+
+# install our class as the default
+logging.setLoggerClass(SfaLogger)
+
+
+# configure
+# this is *NOT* passed to dictConfig as-is
+# instead we filter 'handlers' and 'loggers'
+# to contain just one entry
+# so make sure that 'handlers' and 'loggers'
+# have the same set of keys
+def logging_config(context):
+ if context == 'server':
+ # use stdout and let journalctl do the heavy lifting
+ handlername = 'stdout'
+ #filename = '/var/log/sfa.log'
+ level = 'DEBUG'
+ elif context == 'import':
+ handlername = 'file'
+ filename = '/var/log/sfa-import.log'
+ level = 'INFO'
+ elif context == 'cli':
+ handlername = 'file'
+ filename = os.path.expanduser("~/.sfi.log")
+ level = 'DEBUG'
+ elif context == 'console':
+ handlername = 'stdout'
+ #filename = 'ignored'
+ level = 'INFO'
+ else:
+ print("Cannot configure logging - exiting")
+ exit(1)
+
+ config = {
+ 'version': 1,
+ # IMPORTANT: we may be imported by something else, so:
+ 'disable_existing_loggers': False,
+ 'formatters': {
+ 'standard': {
+ 'datefmt': '%m-%d %H:%M:%S',
+ 'format': ('%(asctime)s %(levelname)s '
+ '%(filename)s:%(lineno)d %(message)s'),
+ },
+ },
+ # fill in later with just the one needed
+ # otherwise a dummy 'ignored' file gets created
+ 'handlers': {
+ },
+ 'loggers': {
+ 'sfa': {
+ 'handlers': [handlername],
+ 'level': level,
+ 'propagate': False,
+ },
+ },
+ }
+ if handlername == 'stdout':
+ config['handlers']['stdout'] = {
+ 'level': level,
+ 'formatter': 'standard',
+ 'class': 'logging.StreamHandler',
+ }
+ else:
+ config['handlers']['file'] = {
+ 'filename': filename,
+ 'level': level,
+ 'formatter': 'standard',
+ 'class': 'logging.handlers.TimedRotatingFileHandler',
+ # every monday and during 3 months
+ 'when': 'w0',
+ 'interval': 1,
+ 'backupCount': 12,
+ }
+ return config
+
+
+logger = logging.getLogger('sfa')
+
+
+def init_logger(context):
+ logging.config.dictConfig(logging_config(context))
+
+
+# if the user process does not do anything
+# like for the miscell testers and other certificate
+# probing/dumping utilities
+init_logger('console')
git://git.onelab.eu / sfa.git / blobdiff