- Zope-Dev - Zope lists

KeyError in UnIndex
by Morten W. Petersen 26 Oct '01

26 Oct '01

Hi guys, I've got an issue here, the following traceback can be seen after several users modify and view the same object in a short (2-3) minute period. I've tried recataloging the objects but that doesn't help either; this error is seen every time.. Site Error An error was encountered while publishing this resource. KeyError Sorry, a site error occurred. Traceback (innermost last): File /usr/local/zope/lib/python/ZPublisher/Publish.py, line 223, in publish_module File /usr/local/zope/lib/python/ZPublisher/Publish.py, line 187, in publish File /usr/local/zope/lib/python/Zope/__init__.py, line 226, in zpublisher_exception_hook (Object: api) File /usr/local/zope/lib/python/ZPublisher/Publish.py, line 171, in publish File /usr/local/zope/lib/python/ZPublisher/mapply.py, line 160, in mapply (Object: index_html) File /usr/local/zope/lib/python/ZPublisher/Publish.py, line 112, in call_object (Object: index_html) File /usr/local/zope/lib/python/Products/warp_framework/object_manager_inter face.py, line 88, in index_html (Object: api) File /usr/local/zope/lib/python/Shared/DC/Scripts/Bindings.py, line 324, in __call__ (Object: object_manager_index) File /usr/local/zope/lib/python/Shared/DC/Scripts/Bindings.py, line 354, in _bindAndExec (Object: object_manager_index) File /usr/local/zope/lib/python/App/special_dtml.py, line 244, in _exec (Object: object_manager_index) File /usr/local/zope/lib/python/DocumentTemplate/DT_In.py, line 770, in renderwob (Object: get_meta_types) File /usr/local/zope/lib/python/DocumentTemplate/DT_Util.py, line 231, in eval (Object: get_number_of_objects(meta_type=_['sequence-item'])) (Info: get_number_of_objects) File <string>, line 0, in ? File /usr/local/zope/lib/python/Products/warp_framework/object_manager.py, line 115, in get_number_of_objects (Object: api) File /usr/local/zope/lib/python/Products/WarpFramework/catalog.py, line 210, in __call__ (Object: LockableItem) File /usr/local/zope/lib/python/Products/ZCatalog/ZCatalog.py, line 597, in searchResults (Object: LockableItem) File /usr/local/zope/lib/python/Products/ZCatalog/Catalog.py, line 649, in searchResults File /usr/local/zope/lib/python/Products/ZCatalog/Catalog.py, line 601, in _indexedSearch File /usr/local/zope/lib/python/Products/PluginIndexes/common/UnIndex.py, line 474, in keyForDocument (Object: PluggableIndex) KeyError: 1074768801 Any help much appreciated, Morten -- About Thingamy: http://www.thingamy.com - For the CEO http://www.thingamy.org - The good stuff http://www.thingamy.net - Rants

1 0

Extending Zope with Eiffel
by Eric Roby 26 Oct '01

26 Oct '01

Ok ... don't laugh (too hard) yet. I know this sounds of blasphemy but let me explain. I spent many years as a meteorologist in the Navy. One tool that would have been invaluable to me (to provide better predictions and warning capabilities) is an expert system. I am sure NOAA probably has one but the Navy doesn't. And they are the ones that would benefit from such a system the most because their forecasting responsibilities are global in nature where NOAA has a much smaller yard to watch over. I have pondered how to build such a system for over 10 years. Things began to fall into place this year when I stumbled upon Zope and when the Navy decided to get serious about Web Services. I like Python ... but I don't think (as a scripting language) it is up to the challenge of an expert system. A language that I feel is up to the challenge is Eiffel (especially in conjunction with ISE EiffelStudio .. an awesome modeling and development environment). So a question for the Zope guru's out there. Since Python is such an adept "glue language" and Eiffel systems are distributed as C-packages, from a technical perspective, would it be absurd to develop the system in Eiffel and integrate it into Zope as a product (or better yet a Web Service in the future). Am I way off base here ? Or am I wrong about my assessment of Python's ability to build an expert system ? Thanks (in advance) for any insight...

4 3

[Bug+Patch] Microseconds in DateTime
by Nicola Larosa 25 Oct '01

25 Oct '01

[Sent to main list, ignored. Retrying here...] The __str__ function in DateTime.DateTime cuts off two precision digits from the time returned by time.time(), falling off to tenths of milliseconds instead of keeping the microseconds. If it feels unrighteous to put the hard work of gettimeofday() to waste, apply the patch below to lib/python/DateTime/DateTime.py, and get back that microsecond precision. Yes, on some hardware/software platforms those microseconds are significant, and yes, they can be useful in many situations. Here is the patch. --- DateTime.py Thu Jul 05 17:55:29 2001 +++ DateTime.py.NEW Mon Oct 22 16:00:36 2001 @@ -1558,11 +1558,11 @@ y,m,d =self._year,self._month,self._day h,mn,s,t=self._hour,self._minute,self._second,self._tz if(h+mn+s): - if (s-int(s))> 0.0001: + if (s-int(s)) >= 0.000001: try: # For the seconds, print two digits # before the decimal point. - subsec = split('%g' % s, '.')[1] + subsec = split('%f' % s, '.')[1] return '%4.4d/%2.2d/%2.2d %2.2d:%2.2d:%2.2d.%s %s' % ( y,m,d,h,mn,s,subsec,t) except: -- Nicola Larosa - nico(a)tekNico.net

3 4

login prompt after letting user change his password.
by Clark OBrien 25 Oct '01

25 Oct '01

Hi all I have written some code to alow a user to change his password (below) The problem is that after executing this code the login dialog pops up. The login requires the user to enter his NEW password. container.REQUEST.set('roles',request.AUTHENTICATED_USER.getRoles()) container.REQUEST.set('domains',request.AUTHENTICATED_USER.getDomains()) container.REQUEST.set('name',request.AUTHENTICATED_USER.getUserName()) container.REQUEST.set('password',request['new_password']) container.REQUEST.set('confirm',request['new_password_confirm']) context.acl_users.manage_users('Change',container.REQUEST) context.REQUEST.RESPONSE.redirect("changepwd?status=ok")

2 1

product distribution compliance
by Behrens Matt - Grand Rapids 25 Oct '01

25 Oct '01

The TTW Product distribution tab is not compliant with <http://www.zope.org/Wikis/DevSite/Proposals/FinishedProductGuidelines>. Products/..., not lib/python/Products/... is what it should be doing. The attached patch makes it so. Incidentally, where are we supposed to be sending patches? Trawling recent mailing list archives comes up with several conflicting answers. klm@, ColDev, the ML...? -- Matt Behrens <matt.behrens(a)kohler.com> System Analyst, Baker Furniture

1 0

ZCatalog bug, probably BTrees, what Zope version is this fixed in?
by Chris Withers 25 Oct '01

25 Oct '01

Hi, I have a ZCatalog search that returns 213 results. As soon as I add a sort_on and sort_order to the request doing this search, the number of results drops to about 60. This reminds me of some bugs in the BTrees package that I remember being solved in the not too distant past. Can someone please confirm to me what version of Zope I need to use to get these fixes? If it's not in a release yet, _please_ can it make it for 2.4.2. cheers, Chris

1 1

howto: background processes, queued events and ZEO?
by Jay, Dylan 25 Oct '01

25 Oct '01

I have a very computationally expensive task that I need to perform for users of my system. The idea is that when they access the page that needs the results of this process they will be returned a note saying to wait while the calculations are done. The calculation request is queued and eventually picked up by a background process and run. Only one calculation is run at once. My current attempt at implementation is using Xron DTML Methods. However from looking at the code I gather that this will not scale well if I decide to use ZEO (I think every client will have a background dispatcher so the task would be run multiple times). Using ZEO I'd like to nominate one or more servers to be the workhorses and do all the calculations. Anyone tried doing something like this or can suggest the best approach?

2 1

startup security status (say that five times fast... well, ok, it wasn't so tough after all)
by Behrens Matt - Grand Rapids 24 Oct '01

24 Oct '01

I have a patch in hand that addresses MOST of the issues I brought up, but the biggie (tricking root into killing arbitrary processes) is a hard one to solve. I have many options, and I'd like opinions... Right now, the pid file is written out by the user that ZServer drops to after starting. This is bad because if that user is compromised, the pid file can be overwritten, and root can be tricked into killing (an) arbitrary process(es) the next time 'stop' is run. The obvious solution was to move the writing of the file up before the setuid() call. Now, the unprivileged user can't *change* the pid file. However, because the var directory must be writable by the unprivileged user, the unprivileged user can *remove* then *rewrite* the pid file, and we're back where we started. Solutions: 1. Have the stop script check ownership of the pid file to make sure it's still root's baby. This solution seems easiest, but something about it doesn't seem right to me. When something doesn't feel right to me, there's probably a way to fool it... 2. Enforce the sticky bit on the var directory. From Solaris' chmod(2) manpage: If a directory is writable and has S_ISVTX (the sticky bit) set, files within that directory can be removed or renamed only if one or more of the following is true (see unlink(2) and rename(2)): o the user owns the file o the user owns the directory o the file is writable by the user o the user is a privileged user (Privileged user means 'root'.) We only need to enforce the sticky bit if we start as root and are doing the requisite setuid(). My patch already has a test for this. 3. Have the pid file written into another directory that only root can write to. The rest of this should probably be another mail, but I figured I'd cover what my patch also does: 1. No longer defaults to running as 'nobody'. As I've explained, running as 'nobody' and the requisite permission settings that need to go with running as 'nobody' can set your Zope data up for compromise on your local system. If -u is not specified z2.py will SystemExit. 2. Warns you if you decide that you REALLY want to run as 'nobody', either with -u or by being nobody when starting z2. 3. Tells you when and who it actually setuid()s to. 4. Warns you if your umask isn't sufficient to protect your data files (experiment: pack your Data.fs and check its permissions.) All messages in this patch are going through zLOG for cleanliness' sake. -- Matt Behrens <matt.behrens(a)kohler.com> System Analyst, Baker Furniture

2 3

CoreSessionTracking onStart
by Godefroid Chapelle 24 Oct '01

24 Oct '01

Hi, I am trying to use the onStart method of a session data manager.... But I do not get how to give the path so that my method get called... I could not find any information of the real way to use it. Can someone help me ? Thanks -- Godefroid Chapelle BubbleNet sprl rue Victor Horta, 30 1348 Louvain-la-Neuve Belgium Tel + 32 (10) 457490 Mob + 32 (477) 363942 TVA 467 093 008 RC Niv 49849

2 1

Re: [Zope-dev] Versions (still)
by Oliver Bleutgen 24 Oct '01

24 Oct '01

> So there I was in this discussion about Zope versioning (again) and there > were two features requested that seemed perfectly reasonable at the time, > - to have a list of all the objects changed by a version Sorry if this is obvious, but at least neither ZopeFind nor locked_in_version() seem to be mentioned anywhere in zope's (2.3.3) online docu, so I thought I might post it: <dtml-in "ZopeFind(PARENTS[-1],search_sub=1,obj_expr='locked_in_version()')"> <LI>found <dtml-var title_or_id html_quote> in Version <dtml-var locked_in_version> </dtml-in> > - to be able to individually commit or discard changes in a version on a > per > object basis > This is would be extremely useful for those times when you forget to check > out of a version and accidentally lock the root folder (This happens quite > a > bit). Simple I thought, but ugh versions go all the way down to > FileStorage > and got into ZODB stuff I didnt understand. I would have thought having a > little refactoring to give two more methods: getVersionContents and > commitObject would be possible, but Im scratching my head at FileStorage > now. Hehe, I feel with you, looking at how versions work takes one straight to Filestorage.commitVersion and some very low level code. > Before I embark down this path is there an easier way? The question is if this is worth the effort, in the light of the new versioning discussion. The weaknesses of the current implementation of versioning puts some pressure to create something new IMO. cheers, oliver

2 1