- Zope-Dev - Zope lists

RE: ZODBC problem
by Capesius, Alan 07 Mar '01

07 Mar '01

Thanks Dieter, I was kinda looking for a solution, I reviewed the mailing lists but do not see any solutions. I've done some more testing here: Summary: The user selects an involved financial report. The report takes about 55 seconds for Zope to prepare. The first second is the SQL server request and reply. The remaining time is Zope processing (dual 200mhz NT box). My queries are SELECT JOINS. No tables are modified at the server. The problem I see here is that Zope/ZODBCDA is holding the connection to SQL open via ODBC until all processing is complete and then issuing an SQL statement: IF @@TRANCOUNT>0 COMMIT TRAN Zope could free up this connection after one second. (at least in this case) Until Zope does the commit, other users are locked out. For read only queries, there is no need to have the request under transaction control. In any case, Zope should be able to early release the SQL connection and then process on the results. Certainly increasing processor speed will mask this problem and provide better response to users, but I'm hoping for a real fix. Thanks, Alan Network Monitor trace Wed 03/07/01 10:51:50 c:\cap1.TXT Frame Time Src MAC Addr Dst MAC Addr Protocol 1 4.413 006008F6062B COMPAQ851F04 TCP 2 4.413 006008F6062B COMPAQ851F04 TCP 3 4.413 006008F6062B COMPAQ851F04 TCP [....] 193 5.072 006008F6062B COMPAQ851F04 TCP 194 5.073 COMPAQ851F04 006008F6062B TCP 195 5.090 COMPAQ851F04 006008F6062B TCP 196 5.090 006008F6062B COMPAQ851F04 TCP 197 5.090 006008F6062B COMPAQ851F04 TCP 198 5.091 COMPAQ851F04 006008F6062B TCP 199 5.109 COMPAQ851F04 006008F6062B TCP 200 5.127 COMPAQ851F04 006008F6062B TCP // SQL server goes idle after this // python DTML processing occurs here 201 10.834 006008F6062B COMPAQ851F04 TCP // not sure what this is, Data is 0xA8 202 10.834 COMPAQ851F04 006008F6062B TCP // reply, data is empty // page returns to user at 60 second mark // then Zope closes the connection 203 60.385 COMPAQ851F04 006008F6062B TCP SQL: IF @@TRANCOUNT>0 COMMIT TRAN 204 60.386 006008F6062B COMPAQ851F04 TCP 205 60.601 COMPAQ851F04 006008F6062B TCP >Capesius, Alan writes: > > ... > > with further testing I > > have determined that the Zope server itself is not locking up, but rather > > requests to the database (MS SQL7 via ZODBCDA) seem to be queueing up. >My memory tells me faintly that such an issue has already been >discussed in the mailing lists (zope, zope-dev). >When I remember right, ZODBCDA synchronizes all database requests >--> searchable list archives... >Dieter

1 0

Attribute providers for Specialists
by Johan Carlsson 07 Mar '01

07 Mar '01

Hi, As I understand it Specialist can't get there attributes from a provider? For instance, the LoginManager knows about roles and domains, I would like to get this from attributes but I don't want to be able to get them from an attribute provider. In that way I can make a GUI interface that doesn't know about the storage, and it could be changes as easy as rewrite my Skinscripts. Well almost anyway. So is there a way to get Specialists, and specially LoginManager to get there attribute from a provider? TIA/Johan Carlsson

1 0

Announcement: CMF 1.0 beta released
by Tres Seaver 06 Mar '01

06 Mar '01

We are pleased to "release version 1.0 beta", http://cmf.zope.org/download/CMF-1.0beta of the Zope Content Management Framework (CMF). Please review the "release notes":../README.txt and the "known issues":../ISSUES.txt. >From the README: What is the CMF? The Zope Content Management Framework provides a set of services and content objects useful for building highly dynamic, content-oriented portal sites. As packaged, the CMF generates a site much like the Zope.org site. The CMF is intended to be easily customizable, in terms of both the types of content used and the policies and services it provides. Resources * The CMF "dogbowl" site, "http://cmf.zope.org":http://cmf.zope.org. * The mailing list, zope-ptk(a)zope.org. List information and online signup are available at: http://lists.zope.org/mailman/listinfo/zope-ptk. Archives of the list are at: http://lists.zope.org/pipermail/zope-ptk. Tres. -- =============================================================== Tres Seaver tseaver(a)digicool.com Digital Creations "Zope Dealers" http://www.zope.org

2 1

Implementing ZPattern: Containment
by Johan Carlsson 06 Mar '01

06 Mar '01

Hi, How could containment be implemented with ZPatterns. Container -n-----1- ContainedItem -n-----1- ContainedSubItem I suppose there would be a Containter Specialist containing a ContainerItem Specialist. But Specialists retrieve objects by IDs so there would also be a list of ContainerItem ID in the Container object. Are there any other way to do it, for instance keeping a list of object references that would be retrieve from the ContainerItem Specialist by a attribute provider in the Containter Specialist? But I still need the list of IDs don't I or how else would the ContainerItem Specialist know what objects to return? TIA, Johan Carlsson

2 2

Re: [Zope-dev] Possible 'bad' with Client.py and urllib
by Dieter Maurer 06 Mar '01

06 Mar '01

Chris Withers writes: > .... urllib & Client block Zope .... That would suggest a bug in Python's socket implementation: its "makefile().read" method would not release the global interpreter lock. Dieter

2 3

LoginUser and Dictionaries
by Johan Carlsson 06 Mar '01

06 Mar '01

Hi, I'm working on a User object that will support Groups and Groups-to-Roles mappings. I modeled it a bit and come up with this model: (I use Coad's constraints notion, putting constraints with the item the constrain) User -n------n- Role | n | | 1 GroupRoleItem -1----- n- Role (*same as the other role above) | 1 | | n Group To make it as simple as possible I whan to keep Group and GroupRoleItem in a dictionary inside the User instance. E.g. {'groupname': ('rolename','rolename',), ... } In my LM implementation I want to store GroupRoleItems in a SQL table and Groups in another SQL table (which records awailable groups). The LoginUser object should have a attribute grouproles_dict that gets its data from the GroupRoleItems table. Here's some code: class GroupLoginUser(LoginUser): ... def getGroupRoles(self): # group_list contains the following brains from the LoginManager (Skinscript/SQL method): # (brains.username) # brains.rolename # brains.groupname # groups = () grouproles = {} if self.group_list: for g in self.group_list: if g.groupname in groups: grouproles[g.groupname].append(g.rolename) else: grouproles[g.groupname] = [g.rolename] groups = groups + (g.groupname,) return grouproles The GroupLoginUser is not stored persitently in my implementation, but I like to cache the grouproles for as long as possible. For example I'm planning a method "getGroups" in the GroupLoginUser which could just return grouproles.keys(). So because GroupLoginUser is a virtual dataskin I'm not sure where I should save the grouproles. In self._v_grouproles or as a skinscripted attribute? E.g. WITH self.getGroupRoles() COMPUTE grouproles = RESULT I might be thinking wrong in the model as well, so any hints and tips for corrections are welcome. Regards, Johan Carlsson

1 0

Class attr hotfix
by Casey Duncan 05 Mar '01

05 Mar '01

I noticed that the most recent hotfix completely denies access to the getClassAttr, setClassAttr and delClassAttr from DTML. Is this going to be the behavior for future Zope versions? I know this will break code, so it might be prudent to alert people when the fix is rolled into the next Zope release. I would like to argue for deprecating these functions in favor of making the ZClass act like a dictionary so that: ZClass.set('name',value) ZClass.get('name'[,inherit]) ZClass['name'] could be called securely from scripts. I would be willing to write the code for this if you think it would be a good idea. ===== | Casey Duncan | Kaivo, Inc. | cduncan(a)kaivo.com `-----------------> __________________________________________________ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/

4 3

is there a hook for before the transaction is committed
by Tim McLaughlin 05 Mar '01

05 Mar '01

Is there a hook for before the transaction is committed for objects which subclass Persistent? I found __inform_commit__ for a "registered" object, but I can't seem to get that to work as I thought it did. I also tried subclassing TM like a DA, but to no avail. TIA, Tim ___________________________________________________________ Tim McLaughlin BCSwebservices.net Director, Technical Group 1950 Old Gallows Road tel: (703) 790.8081 x111 Suite 201 tim(a)bcswebservices.net Vienna, VA 22182 www .bcswebservices. net

4 4

namespace binding in pythonscript
by R. David Murray 05 Mar '01

05 Mar '01

Has the problem with the binding of the namespace in pythonscript been fixed in cvs (if so will it be in b2)? It's a pain to having to call my pythonscript methods using xyz(_=_). --RDM

1 1

RE: [Zope-dev] is there a hook for before the transaction is comm itted
by Tim McLaughlin 05 Mar '01

05 Mar '01

In what respect are the _v_* attribs gonna cause problems. My guestimate was that they disappeared upon transaction commit/abort. I'm also not sure as to why I would need subtrans since I'm only messing with properties of the object. To my knowledge, subtrans are only necessary to conserve resources. Anyway, what you gave me works! Thanks. It seems I need to override the _vote method (since it is only called once and allows exceptions). As to the other stuff, I'm sure you can enlighten me further as to why. I appreciate the help. Cheers. Tim -----Original Message----- From: John D. Heintz [mailto:jheintz@isogen.com] Sent: Monday, March 05, 2001 12:45 PM To: Tim McLaughlin Cc: 'zope-dev(a)zope.org' Subject: Re: [Zope-dev] is there a hook for before the transaction is committed Hi Tim, I have two suggestions, I hope one of them helps. 1) Attached is a TM.py file that I wrote based on the one you mention below. I've tried to make it more obvious and better documented. 2) Don't use this kind of functionality, but rather use sub-transaction commits. The first suggestion has more overhead than what I assume you would need, but the second one won't work for all situations. A Fishbowl proposal of mine, HashingSupport, was going to use the same kind of hook you are asking about. In this case though, using sub-transaction commits made a lot more sense. In general though, I think that _v_* attributes pose a non-trivial problem that probably requires a hook on abort() if not commit() as well. John Tim McLaughlin wrote: > Is there a hook for before the transaction is committed for objects which > subclass Persistent? I found __inform_commit__ for a "registered" object, > but I can't seem to get that to work as I thought it did. I also tried > subclassing TM like a DA, but to no avail. > > TIA, > Tim > > ___________________________________________________________ > Tim McLaughlin BCSwebservices.net > Director, Technical Group 1950 Old Gallows Road > tel: (703) 790.8081 x111 Suite 201 > tim(a)bcswebservices.net Vienna, VA 22182 > www .bcswebservices. net > > > _______________________________________________ > 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 ) -- . . . . . . . . . . . . . . . . . . . . . . . . John D. Heintz | Senior Engineer 1016 La Posada Dr. | Suite 240 | Austin TX 78752 T 512.633.1198 | jheintz(a)isogen.com w w w . d a t a c h a n n e l . c o m

2 1