- Zope-Dev - Zope lists

OpenFlow workflow engine
by Vincenzo Di Somma 04 Sep '01

04 Sep '01

Icube is pleased to announce the availability of OpenFlow 0.4 http://www.zope.org/Members/vds/OpenFlow, the open source workflow engine. Openflow includes the ability to specify multiple maps of activities, support for conditional splits and parallel executions. Each activity can be implemented using every Zope object. A web based management interface completes the engine itself. -- Vincenzo Di Somma - Responsabile Ricerca e Sviluppo - Icube S.r.l. Sede: Via Ridolfi 15 - 56124 Pisa (PI), Italia E-mail: e.disomma(a)icube.it WWW: www.icube.it Tel: (+39) 050 97 02 07 Fax: (+39) 050 31 36 588

2 2

CoreSession, ZEO and BerkeleyStorage = CorruptedDataError
by Tony McDonald 04 Sep '01

04 Sep '01

I'm pooped. I'll apologise in advance that there doesn't seem enough info. In this email, but the setup is so complex that to put it all in would exceed the mailing list message size :) (and to tell the truth, I'm hoping that someone can point a finger at a likely spot). I really like CST, but if this doesn't work, I'm thinking about using SQLSession as I'm really getting out of my depth here. Any thoughts would be really appreciated. I've got a ZEO system set up using this HowTo (which is very good BTW) 1) http://dev.zope.org/Members/dshaw/AdvancedSiteSetup Then I installed BerkeleyStorage from a combination of 2) http://www.zope.org/Members/Duncan/BerkeleyStorage and 3) http://www.zope.org/Members/andyd/zopeberkeleydb Then following the instructions at 4) http://www.zope.org/Members/randy/ZEO-Sessions I set up a ZEO based CoreSessionTracking system and set up several client zope systems (from (1)) and had them accessed 'randomly' using Apache rewriterules. I managed to get the system up and running. But once I 'invalidate' a CoreSession, I get this error message; Error Type: CorruptedDataError Error Value: data_container/manage_container Traceback:  Versions Solaris 8 Python 2.1 Zope 2.4.0 src Berkeley3.2 CoreSessionTracking 0.8 -- Dr Tony McDonald, Assistant Director, FMCC, http://www.fmcc.org.uk/ The Medical School, Newcastle University Tel: +44 191 243 6140 A Zope list for UK HE/FE http://www.fmcc.org.uk/mailman/listinfo/zope

3 3

RE: [Zope-dev] declarePublic and users
by Randall F. Kern 03 Sep '01

03 Sep '01

I for one have hacked around this, so all methods authenticate, even if they are declarePublic'd. I don't recall having to make any changes to ZMI. -Randy > -----Original Message----- > From: Chris McDonough [mailto:chrism@digicool.com] > Sent: Monday, September 03, 2001 2:04 PM > To: seb bacon > Cc: richard(a)bizarsoftware.com.au; zope-dev(a)zope.org > Subject: Re: [Zope-dev] declarePublic and users > > > seb bacon wrote: > > * Chris McDonough <chrism(a)digicool.com> [010903 18:42]: > > > >>Yeah, this is due to a shortcut in authentication. I don't > particularly > >>like this behavior. It bites lots of people. I wonder if we could > >>change it. > > > > I think so, FWIW. As you say, this behaviour is a result of an > > implementation shortcut, rather than by design; I doubt the > > performance benefits outweigh the silliness of the resulting > > behaviour. > > > > I thought I remembered someone from ZC saying it was > changed for 2.4, > > though? > > Geez, I don't think so... maybe you mean the "Authenticated" > role being > added? > > I'd say ditch the current behavior if we can be reasonably > certain that > nothing important (especially in the ZMI) depends on it. > > Anybody want to write a fishbowl proposal? ;-) > > -- > Chris McDonough Zope Corporation > http://www.zope.org http://www.zope.com > "Killing hundreds of birds with thousands of stones" > > > _______________________________________________ > Zope-Dev maillist - Zope-Dev(a)zope.org > http://lists.zope.org/mailman/listinfo/zope-dev > ** No cross posts or HTML encoding! ** > (Related lists - > http://lists.zope.org/mailman/listinfo/zope-announce > http://lists.zope.org/mailman/listinfo/zope ) >

1 0

/lib/python in INSTANCE_HOME
by Chris Withers 03 Sep '01

03 Sep '01

Hi, Was I just on something funny or do I remember from somewhere that you can now have the equivalent of a lib/python in your INSTANCE_HOME? If so, how would I got about doing it? cheers, Chris

2 3

Something to brighten your day!
by FotoAsia 03 Sep '01

03 Sep '01

1 0

Publish site to end user
by Sunil Kumar Gupta 03 Sep '01

03 Sep '01

Hi, I have used CMFDefault product to create a template. I want it to be viewed by end users without Zope standard headers and footers. One way to remove it is by removing the link from the template. But then the problem is owner of the template does not get any link to edit the template. Is there any way so that both the things can be achieved. 1. No standard headers shoulds be displayed for Anonymous end users. 2. Owner, reviewer, manager should be able to see these header and footer. Regards, sunil

1 0

RE: [Zope-dev] monkeypatching Products
by Adrian Hungate 02 Sep '01

02 Sep '01

Yes. Look at my PatchKit - http://www.zope.org/Members/haqa/PatchKit If you need help with how to create a patch, drop me a line. Adrian... -- Adrian Hungate Try Zope - http://www.zope.org - You might like it. If you do, try my Zope products - http://www.zope.org/Members/haqa - You might like them too! -----Original Message----- From: sean.bowman(a)acm.org [mailto:sean.bowman@acm.org] Sent: 02 September 2001 22:02 To: zope-dev(a)zope.org Subject: [Zope-dev] monkeypatching Products hello, Is it possible for a Product to "monkey patch" another Product? If so, is there an example of this anywhere? I'm trying to add a method to a class in a different product, and I'd like to aviod subclassing. Thanks, Sean _______________________________________________ Zope-Dev maillist - Zope-Dev(a)zope.org http://lists.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope )

2 1

monkeypatching Products
by sean.bowman＠acm.org 02 Sep '01

02 Sep '01

hello, Is it possible for a Product to "monkey patch" another Product? If so, is there an example of this anywhere? I'm trying to add a method to a class in a different product, and I'd like to aviod subclassing. Thanks, Sean

1 0

adding uniqueValuesFor to CatalogQuery
by Steve Alexander 01 Sep '01

01 Sep '01

Hi folks, I've been playing around with adding a uniqueValuesFor capability to Casey Duncan's CatalogQuery product. This is more a proof-of-concept, so it isn't straightforward to use just yet. There are probably bugs that I've overlooked, especially due to assumptions about the types of things that BTrees and Buckets return. Patches and explanations gratefully accepted! Interestingly, I came up with two algorithms for uniqueValuesFor. One of them uses the fact that the difference method of IMerge returns items, not just keys. The other algorithm is less tricky and more laborious. I decided to test the algorithms using a Catalog I have lying around. It isn't all that full, maybe a few hundred records, and a few tens of unique values for the index I'm interested in. I found that the first algorithm runs about ten times as fast as the second. I also tried it on a larger Catalog (one with 2000 records), and the first algorithm ran 30 times as fast as the second. Attached are the diffs and the test script I used. The test script is specific to a particular application I have, but it should be easy to run it against another catalog. Oh, by the way, Casey's CatalogQuery is a MonkeyPatch (those run-time patches formerly known as hotfixes) on the ZCatalog product. I did, for just a second, consider releasing this as a MonkeyPatch product that patches CatalogQuery. So, it would be a MonkeyPatch of a MonkeyPatch, or perhaps a MonkeyMonkeyPatch. -- Steve Alexander Software Engineer Cat-Box limited import Zope, ZPublisher #c=Zope.app().school.Instructors.Catalog c=Zope.app().school.Pupils.Catalog from Products.CatalogQuery import CatalogQuery cq=CatalogQuery(c._catalog, 'current==1') field="area_upper" a1=cq(uniqueValuesFor=field) a2=cq(uv2=field) a3=cq(uv3=field) from time import time m1=time() a1=cq(uniqueValuesFor=field) m2=time() a2=cq(uv2=field) m3=time() a3=cq(uv3=field) m4=time() print "a1 took", m2-m1 print "a2 took", m3-m2 print "a3 took", m4-m3 # ---- Catalog with a few hundred records # a1 took 0.00448000431061 # a2 took 0.0652350187302 # a3 took 0.0657860040665 # ---- Catalog with 2000 records # a1 took 0.0895169973373 # a2 took 2.71825802326 # a3 took 2.75240898132 *** original CatalogQuery.py --- new CatalogQuery.py *************** *** 43,48 **** --- 43,51 ---- from Acquisition import Implicit from Products.ZCatalog.Lazy import LazyMap from BTrees.IIBTree import intersection, weightedIntersection, union, weightedUnion, difference, IISet + from BTrees import IOBTree + from BTrees.OIBTree import OISet + from types import IntType CQL_re = re.compile( ('\s*([_\-a-zA-Z0-9]+)\s*' *************** *** 167,173 **** if hasattr(self, '_v_all_keys'): del self._v_all_keys ! def __call__(self, md={}, used=None, **kw): """Execute the query on the catalog""" # Combine any keywords into the namespace mapping --- 170,176 ---- if hasattr(self, '_v_all_keys'): del self._v_all_keys ! def __call__(self, md={}, used=None, uniqueValuesFor=None, uv2=None, uv3=None, **kw): """Execute the query on the catalog""" # Combine any keywords into the namespace mapping *************** *** 186,191 **** --- 189,237 ---- w, rs = weightedIntersection(rs, r) elif block.logic == 'or': w, rs = weightedUnion(rs, r) + + if uniqueValuesFor: + inverse_rs=IOBTree.IOSet(difference(self._getAllKeys(), rs)) + index=self._catalog.indexes[uniqueValuesFor] + results=IOBTree.difference(index._unindex, inverse_rs) + resultset=OISet() + for l in results.values(): + resultset.update(l) + self._cleanUp() + return resultset.keys() + if uv2: + index=self._catalog.indexes[uv2] + doc_ids=intersection(rs, IISet(index._unindex.keys())) + d={} # could use an OISet here for automatic sorted order + for k,v in index._index.items(): + if isinstance(v, IntType): + if v in doc_ids: + d[k]=None + else: + for i in v.keys(): + if i in doc_ids: + d[k]=None + continue + resultset=d.keys() + resultset.sort() + self._cleanUp() + return resultset + if uv3: + index=self._catalog.indexes[uv3] + doc_ids=intersection(rs, IISet(index._unindex.keys())) + resultset=OISet() + for k,v in index._index.items(): + if isinstance(v, IntType): + if v in doc_ids: + resultset.insert(k) + else: + for i in v.keys(): + if i in doc_ids: + resultset.insert(k) + continue + self._cleanUp() + return resultset.keys() + # Create a lazy result map of the results if hasattr(rs, 'keys'): rs=rs.keys()

1 0

BTrees bug
by Steve Alexander 01 Sep '01

01 Sep '01

>>> from BTrees.IOBTree import IOBucket >>> q=IOBucket({1:'a'}) >>> q.keys() [1] >>> q.values() [1] >>> q.items() [(1, 'a')] I'd expect q.values() to give me ['a']. This patch seems to fix it. *** original BucketTemplate.c --- new BucketTemplate.c *************** *** 686,692 **** for (i=low; i <= high; i++) { ! COPY_KEY_TO_OBJECT(v, self->keys[i]); UNLESS (v) goto err; if (PyList_SetItem(r, i, v) < 0) goto err; } --- 686,692 ---- for (i=low; i <= high; i++) { ! COPY_VALUE_TO_OBJECT(v, self->values[i]); UNLESS (v) goto err; if (PyList_SetItem(r, i, v) < 0) goto err; } cc: Collector item 2561 -- Steve Alexander Software Engineer Cat-Box limited

1 0