Description

Once Xapian search has been enabled for the wiki updating a page results into IIS throwing an unhandled exception.

Steps to reproduce

run a MoinMoin wiki instance within a win32 environment using the IIS 7.5 webserver and the isapi_wsgi handler (more infos regarding the setup can be found here)
install the Xapian python bindings for the win32 platform
enable the Xapian index based search for the wiki

Component selection

unhandled excpetion is thrown within the _indexingRequest() method of the MoinMoin\search\builtin.py module.
```
   1 def _indexingRequest(self, request):
   2         """ Return a new request that can be used for index building.
   3 
   4         This request uses a security policy that lets the current user
   5         read any page. Without this policy some pages will not render,
   6         which will create broken pagelinks index.
   7 
   8         @param request: current request
   9         """
  10         import copy
  11         from MoinMoin.security import Permissions
  12         from MoinMoin.logfile import editlog
  13 
  14         class SecurityPolicy(Permissions):
  15 
  16             def read(self, *args, **kw):
  17                 return True
  18         #winpdb breakpoint for debugging purposes
  19         import rpdb2
  20         rpdb2.start_embedded_debugger("""...Removed-password-from-output...""")
  21         r = copy.copy(request)  #<-- IIS exception is caused
```

Details

It is possible to update the index entries of a wiki page by using a standalone commandline (script).

Yet despite the script succeeds the STDOUT of the script contains the following messages reporting that a runtime error has occured:

Exception RuntimeError: 'maximum recursion depth exceeded while calling a Python object' in <type 'exceptions.AttributeError'> ignored 
Exception RuntimeError: 'maximum recursion depth exceeded while calling a Python object' in <type 'exceptions.AttributeError'> ignored

The mentioned error messages do not seem to be a real exception as the script successfully finishes and updates Xapian index for the given page without having to catch the exceptions. Furthermore adding exception handling to the script does not catch the messages they are still printed to STDOUT of the script.

As the wiki is hosted as a ISAPI application within IIS the following isapi_wsgi handler is used.

I also tried to wrap the MoinMoin wsgi application with an additional Exception handler ( as described here http://lucumr.pocoo.org/2007/5/21/getting-started-with-wsgi ) because I wanted to prevent any exception being propagated to ISS. But the IIS exception still occurs. Maybe preventing the exception message being propagated to IIS was not possible for the same reasons it is not possible to suppress the messages to STDOUT within the commandline script?

MoinMoin Version	1.9.3
OS and Version	Windows 7 32bit
Python Version	2.6.6 (r266:84297, Aug 24 2010, 18:46:32) [MSC v.1500 32 bit (Intel)]
Server Setup	IIS 7.5
Server Details	MoinMoin is running as wsgi application using the isapi_wsgi_0.4.2 handler package; Xapian bindings version 1.2.8
Language you are using the wiki in (set in the browser/UserPreferences)	language_default = 'en' taken from the wikiconfig.py

Workaround

apart from disabling the Xapian based search no workaround is possible.

Discussion

I had a similary error under apache2 with mod_wsgi

[error] Exception RuntimeError: 'maximum recursion depth exceeded while calling a Python object' in <type 'exceptions.AttributeError'> ignored

Problem: Now the problem is that when checking for an exception, we do exactly that: check whether the raised exception is a subclass of a given type. If the raised exception occurred just after a recursion overflow, it can happen that subclass checking overflows the recursion count again. However, since we don't want the newly raised exception (during exception subclass checking) to overwrite the original one, so we just write out it (using PyErr_WriteUnraisable, precisely) and then discard it. See here: http://bugs.python.org/issue5508
Solution: Ignore it

Linux

I added this changeset (http://hg.moinmo.in/moin/1.9/rev/5f5faedac588) and if I try to save a page now, I receive this error.

ERROR 2012-06-09 01:07:32,141 MoinMoin.wsgiapp:293 An exception has occurred [https://www.heavy.ch/rock/Lost%20and%20Found].
Traceback (most recent call last):
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 282, in __call__
    response = run(context)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 88, in run
    response = dispatch(request, context, action_name)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 136, in dispatch
    response = handle_action(context, pagename, action_name)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 195, in handle_action
    handler(context.page.page_name, context)
  File "/home/lotek/lib/python2.6/MoinMoin/action/edit.py", line 164, in execute
    savemsg = pg.saveText(savetext, rev, trivial=trivial, comment=comment)
  File "/home/lotek/lib/python2.6/MoinMoin/PageEditor.py", line 1148, in saveText
    results = events.send_event(e)
  File "/home/lotek/lib/python2.6/MoinMoin/events/__init__.py", line 257, in send_event
    retval = handle(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/xapian_index.py", line 77, in handle
    handle_changed(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/xapian_index.py", line 51, in handle_changed
    index.update_item(event.page.page_name)
  File "/home/lotek/lib/python2.6/MoinMoin/search/builtin.py", line 180, in update_item
    self.do_queued_updates()
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 216, in do_queued_updates
    self._index_page(request, connection, pagename, mode='update')
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 372, in _index_page
    updated = self._index_page_rev(request, connection, pagename, revno, mode=mode)
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 409, in _index_page_rev
    mimetype = 'text/%s' % page.pi['format']  # XXX improve this
  File "/home/lotek/lib/python2.6/MoinMoin/Page.py", line 252, in get_pi
    self.__pi = self.parse_processing_instructions()
  File "/home/lotek/lib/python2.6/MoinMoin/Page.py", line 961, in parse_processing_instructions
    request.setPragma(key, val)
AttributeError: 'FakeRequest' object has no attribute 'setPragma'

the error is generated by my pragma stuff (in the wiki page itself):

#pragma keywords Hard, Rock, Magazin, Heavy, Metal, Schweiz, Glam, Hairmetal, HeavyMetal, Hardrock, Glamrock, Pop, PopMetal, Zeitschrift, Magazin, eZine, Newspaper, Information, Musik, Music, RockHeavy, Rock.Heavy.ch
#pragma description Lost'n'Found - Das Heavy Metal Wiki und Online Rock Magazin unter Rock.Heavy.ch

after i removed those pragma, I could save the page.

SystemInformation

Linux version 2.6.32-220.4.2.el6.x86_64 (mockbuild@c6b18n3.bsys.dev.centos.org) (gcc version 4.4.6 20110731 (Red Hat 4.4.6-3) (GCC) ) #1 SMP Tue Feb 14 04:00:16 GMT 2012
Python 2.6
Apache/2.2.17
Mod_wsgi 3.2

bye -- MarcelHäfner 2012-06-09 08:04:59

Thanks for the feedback. Such stuff was somehow expected as (due to copy.copy not really working), we started with a rather empty FakeRequest object there. I'll have a look ASAP. -- ThomasWaldmann 2012-06-09 14:52:12

Thanks, it works now without any error message. -- MarcelHäfner 2012-06-09 16:03:10

RenamePage Fails

If I try to rename a page it fails with AttributeError: 'FakeRequest' object has no attribute 'parsePageLinks_running'. I use the latest rev and same configuration as obove (linux, mod-wsgi, python 2.6, etc.)

ERROR 2012-06-27 13:14:24,189 MoinMoin.wsgiapp:293 An exception has occurred [https://www.heavy.ch/lotek/MoinMoinWiki/Development1.7].
Traceback (most recent call last):
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 282, in __call__
    response = run(context)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 88, in run
    response = dispatch(request, context, action_name)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 136, in dispatch
    response = handle_action(context, pagename, action_name)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 195, in handle_action
    handler(context.page.page_name, context)
  File "/home/lotek/lib/python2.6/MoinMoin/action/RenamePage.py", line 183, in execute
    RenamePage(pagename, request).render()
  File "/home/lotek/lib/python2.6/MoinMoin/action/__init__.py", line 214, in render
    success, self.error = self.do_action()
  File "/home/lotek/lib/python2.6/MoinMoin/action/RenamePage.py", line 66, in do_action
    success, msgs = self.page.renamePage(newpagename, comment)
  File "/home/lotek/lib/python2.6/MoinMoin/PageEditor.py", line 652, in renamePage
    send_event(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/__init__.py", line 257, in send_event
    retval = handle(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/xapian_index.py", line 73, in handle
    handle_renamed(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/xapian_index.py", line 29, in handle_renamed
    index.update_item(event.page.page_name)
  File "/home/lotek/lib/python2.6/MoinMoin/search/builtin.py", line 180, in update_item
    self.do_queued_updates()
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 216, in do_queued_updates
    self._index_page(request, connection, pagename, mode='update')
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 372, in _index_page
    updated = self._index_page_rev(request, connection, pagename, revno, mode=mode)
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 425, in _index_page_rev
    multivalued_fields['linkto'] = page.getPageLinks(request)
  File "/home/lotek/lib/python2.6/MoinMoin/Page.py", line 1490, in getPageLinks
    links = self.parsePageLinks(request)
  File "/home/lotek/lib/python2.6/MoinMoin/Page.py", line 1513, in parsePageLinks
    if request.parsePageLinks_running.get(pagename, False):
AttributeError: 'FakeRequest' object has no attribute 'parsePageLinks_running'

2012-07-05: AttributeError: 'FakeRequest' object has no attribute 'clock'

And still not working now "clock" attribut error.

ERROR 2012-07-05 18:53:28,508 MoinMoin.wsgiapp:293 An exception has occurred [ht                                                                                                                    tps://www.heavy.ch/rock/best].
Traceback (most recent call last):
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 282, in __call__
    response = run(context)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 88, in run
    response = dispatch(request, context, action_name)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 136, in dispatch
    response = handle_action(context, pagename, action_name)
  File "/home/lotek/lib/python2.6/MoinMoin/wsgiapp.py", line 195, in handle_acti                                                                                                                    on
    handler(context.page.page_name, context)
  File "/home/lotek/lib/python2.6/MoinMoin/action/RenamePage.py", line 183, in e                                                                                                                    xecute
    RenamePage(pagename, request).render()
  File "/home/lotek/lib/python2.6/MoinMoin/action/__init__.py", line 214, in ren                                                                                                                    der
    success, self.error = self.do_action()
  File "/home/lotek/lib/python2.6/MoinMoin/action/RenamePage.py", line 66, in do                                                                                                                    _action
    success, msgs = self.page.renamePage(newpagename, comment)
  File "/home/lotek/lib/python2.6/MoinMoin/PageEditor.py", line 652, in renamePa                                                                                                                    ge
    send_event(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/__init__.py", line 257, in sen                                                                                                                    d_event
    retval = handle(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/xapian_index.py", line 73, in                                                                                                                     handle
    handle_renamed(event)
  File "/home/lotek/lib/python2.6/MoinMoin/events/xapian_index.py", line 29, in                                                                                                                     handle_renamed
    index.update_item(event.page.page_name)
  File "/home/lotek/lib/python2.6/MoinMoin/search/builtin.py", line 180, in upda                                                                                                                    te_item
    self.do_queued_updates()
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 216,                                                                                                                     in do_queued_updates
    self._index_page(request, connection, pagename, mode='update')
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 372,                                                                                                                     in _index_page
    updated = self._index_page_rev(request, connection, pagename, revno, mode=mo                                                                                                                    de)
  File "/home/lotek/lib/python2.6/MoinMoin/search/Xapian/indexing.py", line 425,                                                                                                                     in _index_page_rev
    multivalued_fields['linkto'] = page.getPageLinks(request)
  File "/home/lotek/lib/python2.6/MoinMoin/Page.py", line 1490, in getPageLinks
    links = self.parsePageLinks(request)
  File "/home/lotek/lib/python2.6/MoinMoin/Page.py", line 1521, in parsePageLink                                                                                                                    s
    request.clock.start('parsePageLinks')
AttributeError: 'FakeRequest' object has no attribute 'clock'

See repo or please test fakerequest.diff.

Plan

Priority:
Assigned to:
Status:
- fixed by http://hg.moinmo.in/moin/1.9/rev/5f5faedac588 - needs more testing by xapian indexing users on all platforms
- fixed setPragma/getPragma AttributeError: http://hg.moinmo.in/moin/1.9/rev/aaf543f558d9
- fixed getPageLinks_running AttributeError: http://hg.moinmo.in/moin/1.9/rev/aba21f8db240
- more fixes: http://hg.moinmo.in/moin/1.9/rev/9fd0e9e1b4a5 + refactored in http://hg.moinmo.in/moin/1.9/rev/fc11712e0df0

CategoryMoinMoinBugFixed

MoinMoin: MoinMoinBugs/1.9.3XapianNoIndexUpdateAtPageUpdateIfRunningIIS (last edited 2012-07-22 19:09:08 by ThomasWaldmann)