[Mailman-Developers] Anyone else seeing this?

[Barry A. Warsaw]

--M7zH6CPYdt
Content-Type: text/plain; charset=us-ascii
Content-Description: message body text
Content-Transfer-Encoding: 7bit


Corbett,

Everything looks okay to me, and I just did a fresh install, as root,
from the 1.0b5 tarball.  Then I created a dummy list and had no
problem viewing the list-of-lists and the specific list info page.
Note that I am running this in a Solaris 2.6 system, what OS are you
using?   We've had one other report of something similar happening on
Linux (so I wonder if some OS problem could happening).

Since I can't reproduce the problem on my end, I'm going to need your
help.  We have trace through the driver script to see where it's
crapping out.  Here's a version that does some very simple print-style 
tracing to /tmp/mailman.log.  Can you substitute this for
scripts/driver and see if it gives you a clue as to what is failing?

-Barry


--M7zH6CPYdt
Content-Type: text/plain
Content-Description: test driver script
Content-Disposition: inline;
	filename="driver"
Content-Transfer-Encoding: 7bit

#! /usr/bin/env python
#
# Copyright (C) 1998 by the Free Software Foundation, Inc.
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
# 
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
# 
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software 
# Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.


# Useful for debugging.  When an error occurs, this attaches the file name to
# the exception string and re-raises (using the bogus Python 1.5 semantics)
# this may be unnecessary in Python 1.5.2

## realopen = open
## def open(filename, mode='r', bufsize=-1, realopen=realopen):
##     try:
##         return realopen(filename, mode, bufsize)
##     except IOError, e:
##         strerror = e.strerror + ': ' + filename
##         e.strerror = strerror
##         e.args = (e.args[0], strerror)
##         # Python 1.5
##         import sys
##         raise e, None, sys.exc_info()[2]
##         # Python 1.5.1
##         #raise
## import __builtin__
## __builtin__.__dict__['open'] = open



# This standard driver script is used to run CGI programs, wrapped in code
# that catches errors, and displays them as HTML.  This guarantees that
# (almost) any problems in the Mailman software doesn't result in a Web server
# error.  It is much more helpful to generate and show a traceback, which the
# user could send to the administrator, than to display a server error and
# have to trudge through server logs.

# Note: this isn't 100% perfect!  Here are some things that can go wrong that
# are not caught and reported as traceback-containing HTML:
#
# - This file could contain a syntax error.  In that case, you would indeed
#   get a Web server error since this file wouldn't even compile, and there's
#   no way to catch that.
#
# - The sys module could be royally screwed.  Either we couldn't import it, or 
#   it didn't have a sys.stderr attribute.  Both those would indicate serious
#   problems in the Python installation.  These won't generate Web server
#   errors, but neither will they give meaningful tracebacks.
#
# - We couldn't import the traceback module, or traceback.print_exc()
#   failed.  Same diagnosis and effect as with sys being broken.
#
# I consider all these pretty unlikely.  Note that it is also possible that
# the environment variables wouldn't be printed, perhaps because os couldn't
# be imported or there was some problem with os.environ.  Again, not likely.



def run_main():
    try:
        # insert the relative path to the parent of the Mailman package
        # directory, so we can pick up the Utils module
        fp.write('import os\n')
        import os
        # sys gets imported at module level below
        fp.write('sys.path.insert()\n')
        sys.path.insert(0, os.pardir)
        # map stderr to a logger, if possible
        fp.write('import StampedLogger\n')
        from Mailman.Logging.StampedLogger import StampedLogger
        fp.write('import MultiLogger\n')
        from Mailman.Logging.MultiLogger import MultiLogger
        fp.write('instantiating logger\n')
        logger = StampedLogger('error',
                               label='admin',
                               manual_reprime=1,
                               nofail=0)
        fp.write('instantiating multi\n')
        multi = MultiLogger(sys.__stdout__, logger)
        # The name of the module to run is passed in argv[1].  What we
        # actually do is import the module named by argv[1] that lives in the
        # Mailman.Cgi package.  That module must have a main() function, which
        # we dig out and call.
        #
        fp.write('should now get normal error reporting\n')
        scriptname = sys.argv[1]
        # See the reference manual for why we have to do things this way.
        # Note that importing should have no side-effects!
        pkg = __import__('Mailman.Cgi', globals(), locals(), [scriptname])
        module = getattr(pkg, scriptname)
        main = getattr(module, 'main')
        try:
            main()
        except SystemExit:
            # this is a valid way for the function to exit
            pass
    except:
        print_traceback(logger, multi)
        print_environment(logger)



def print_traceback(logger, multi):
    print """\
Content-type: text/html

<p><h3>We're sorry, we hit a bug!</h3>

<p>If you would like to help us identify the problem, please
email a copy of this page to the webmaster for this site with
a description of what happened.  Thanks!

<h4>Traceback:</h4>
<p><pre>
"""
    logger.write('[----- Traceback ------]\n')
    try:
        import traceback
        # in normal situation, this will get logged to the MultiLogger created
        # above, which will write the data to both the real live stdout, and
        # the StampedLogger
        traceback.print_exc(file=multi)
    except:
        multi.write('[failed to get a traceback]\n')
    print '\n\n</pre>'



def print_environment(logger):
    try:
        import os
        print '''\
<p><hr><h4>Environment variables:</h4>

<p><table>
<tr><td><strong>Variable</strong></td>
<td><strong>Value</strong></td></tr>
'''
        logger.write('[----- Environment Variables -----]\n')
        for varname, value in os.environ.items():
            print '<tr><td>', varname, '</td><td>', value, '</td></tr>'
            logger.write('\t%s: %s\n' % (varname, value))
        print '</table>'
    except:
        print '<p><hr>[environment variables are not available]'


try:
    fp = open('/tmp/mailman.log', 'w+')
    fp.write('=====\nimport sys\n')
    import sys
    try:
        fp.write('run_main()\n')
        run_main()
        fp.write('done.\n')
    finally:
        # this is probably not strictly necessary since the script is exiting
        # soon anyway
        sys.stderr = sys.__stderr__
except:
    # Jeez, we couldn't even import sys, or sys didn't have a stderr
    # attribute!
    print """\
Content-type: text/html

<p><h3>We're sorry, we hit a bug!</h3>

Mailman experienced a very low level failure and could not even generate
a useful traceback.  Please report this to the Mailman administrator at
this site.
"""

--M7zH6CPYdt--