Problem committing zope "version" objects
Hi All, Several times I have attempted to use Zope "Version" objects to update sections of my web site. This works fine as long as I don't *create* any files. If I create a file, performing a "Save" appears to have no effect. All modified/created/delete objects remain locked by the version. Is this a general problem with Zope? Is there something that I am doing wrong? What is going on? -Greg (Zope 2.6.1 (binary release, python 2.1, solaris-2.8-sparc), python 2.1.3, sunos5) LEGAL NOTICE Unless expressly stated otherwise, this message is confidential and may be privileged. It is intended for the addressee(s) only. Access to this E-mail by anyone else is unauthorized. If you are not an addressee, any disclosure or copying of the contents of this E-mail or any action taken (or not taken) in reliance on it is unauthorized and may be unlawful. If you are not an addressee, please inform the sender immediately.
On Tue, Jun 03, 2003 at 10:37:44AM -0400, Warnes, Gregory R wrote:
Is this a general problem with Zope? Is there something that I am doing wrong? What is going on?
You're not alone, lots of people have problems with Versions. Check the mailing list archives. I don't touch Versions anymore. Chris, I'm CC'ing you to ask: Is there any chance the Zope Book section on versions could get a big warning stamped on it? I think it's really a disservice to new users that there's all these docs and feature lists for Zope that tout the usefulness of Versions, and inevitably the poor user discovers that versions have helpfully screwed their work. -- Paul Winkler http://www.slinkp.com Look! Up in the sky! It's BER GUY! (random hero from isometric.spaceninja.com)
Paul Winkler wrote:
On Tue, Jun 03, 2003 at 10:37:44AM -0400, Warnes, Gregory R wrote:
Is this a general problem with Zope? Is there something that I am doing wrong? What is going on?
You're not alone, lots of people have problems with Versions. Check the mailing list archives. I don't touch Versions anymore.
Chris, I'm CC'ing you to ask: Is there any chance the Zope Book section on versions could get a big warning stamped on it? I think it's really a disservice to new users that there's all these docs and feature lists for Zope that tout the usefulness of Versions, and inevitably the poor user discovers that versions have helpfully screwed their work.
Wouldn't it be better to stamp that warning sign on zope 2.6.x (and all other versions which exhibit this bug) unless this is fixed? Ahem, it will be fixed in 2.6.2, will it? cheers, oliver
On Tue, Jun 03, 2003 at 07:28:27PM +0200, Oliver Bleutgen wrote:
Wouldn't it be better to stamp that warning sign on zope 2.6.x (and all other versions which exhibit this bug) unless this is fixed?
might be a good idea. 2.5.x and 2.6.x, at least, have bad problems with versions.
Ahem, it will be fixed in 2.6.2, will it?
seems doubtful... here's another bug: http://collector.zope.org/Zope/922 no responses there. I get the impression that there are multiple version-related bugs which have not been put in the collector. And nobody seems to have the time or motivation to fix versions, so every so often they come up on the mailing list and the response is "don't use versions". Ergo, I think they should be stamped with a big warning label or even disabled until they actually work. -- Paul Winkler http://www.slinkp.com Look! Up in the sky! It's RAPIST GATOR! (random hero from isometric.spaceninja.com)
Chris, I'm CC'ing you to ask: Is there any chance the Zope Book section on versions could get a big warning stamped on it? I think it's really a disservice to new users that there's all these docs and feature lists for Zope that tout the usefulness of Versions, and inevitably the poor user discovers that versions have helpfully screwed their work.
Most people go to the documentation when they find they have a problem, in this case when its too late. I dont think stamping a big warning in the documentation is enough. With no real effort being made to solve them I would strongly suggest that versions are actually removed. Or somehow stamped in big letters in the ZMI "Buggy, liable to break things, alpha quality" :) -- Andy McKay
Andy McKay wrote:
Chris, I'm CC'ing you to ask: Is there any chance the Zope Book section on versions could get a big warning stamped on it? I think it's really a disservice to new users that there's all these docs and feature lists for Zope that tout the usefulness of Versions, and inevitably the poor user discovers that versions have helpfully screwed their work.
Most people go to the documentation when they find they have a problem, in this case when its too late. I dont think stamping a big warning in the documentation is enough.
With no real effort being made to solve them I would strongly suggest that versions are actually removed. Or somehow stamped in big letters in the ZMI "Buggy, liable to break things, alpha quality" :)
Versions are dangerous even if this is fixed. The main reason for that is that the "Join/Leave Versions" permission is snake oil. They should vanish. cheers, oliver
Oliver Bleutgen wrote at 2003-6-3 23:19 +0200:
... Versions are dangerous even if this is fixed. The main reason for that is that the "Join/Leave Versions" permission is snake oil. They should vanish.
-3 I like versions and use them from time to time to install changes automatically on life servers. I do not care at all about the "Join/Leave Versions" permission. The only thing one can get is to see work in progres. I have no problem with this. Dieter
It's farly difficult to maintain documentation that is riddled with very specific references to open-ended time-sensitive bugs (sort of like maintaining C code with lots of #ifdefs in them). But apparently this bug has been fixed for 2.6.2, so it's probably a good idea to mention in the book that "versions are somewhat broken up to Zope 2.6.2". Does this make sense? - C On Tue, 2003-06-03 at 16:16, Andy McKay wrote:
Chris, I'm CC'ing you to ask: Is there any chance the Zope Book section on versions could get a big warning stamped on it? I think it's really a disservice to new users that there's all these docs and feature lists for Zope that tout the usefulness of Versions, and inevitably the poor user discovers that versions have helpfully screwed their work.
Most people go to the documentation when they find they have a problem, in this case when its too late. I dont think stamping a big warning in the documentation is enough.
With no real effort being made to solve them I would strongly suggest that versions are actually removed. Or somehow stamped in big letters in the ZMI "Buggy, liable to break things, alpha quality" :) -- Andy McKay
_______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
On Tue, 2003-06-03 at 16:16, Andy McKay wrote:
Chris, I'm CC'ing you to ask: Is there any chance the Zope Book section on versions could get a big warning stamped on it? I think it's really a disservice to new users that there's all these docs and feature lists for Zope that tout the usefulness of Versions, and inevitably the poor user discovers that versions have helpfully screwed their work.
Most people go to the documentation when they find they have a problem, in this case when its too late. I dont think stamping a big warning in the documentation is enough.
As a newbie who was bitten by Versions and has read the 2.6 docs, I'd like to promote the idea of a warning label. Although a ton of direct help is provided by posters, many times I've seen the (appropriate) suggestion to 'read the book'. I don't know who maintains the book or what their workload is, but when a problem keeps snagging people wouldn't it be worth the time to stop that by flagging the docs or adding an errata page, especially for something like Versions that sounds like such a good idea? Just my two cents. No response required. I understand this is fixed in newer versions. Thanks, Tom P. Allred, MCSE, RHCE, CCNA Allred Data Consulting, Inc. 704-633-0311 x826 www.AllredData.com Tom@AllredData.com
On Tue, 2003-06-03 at 23:03, Chris McDonough wrote:
It's farly difficult to maintain documentation that is riddled with very specific references to open-ended time-sensitive bugs (sort of like maintaining C code with lots of #ifdefs in them).
Absolutely.
But apparently this bug has been fixed for 2.6.2, so it's probably a good idea to mention in the book that "versions are somewhat broken up to Zope 2.6.2".
IHMO there are more bugs with Versions than just this and I think for most users this is too late since most dont read the book until its too late. -- Andy McKay
Warnes, Gregory R wrote at 2003-6-3 10:37 -0400:
Several times I have attempted to use Zope "Version" objects to update sections of my web site. This works fine as long as I don't *create* any files. If I create a file, performing a "Save" appears to have no effect. All modified/created/delete objects remain locked by the version.
That's a very nice problem description. When it allows me to reproduce the problem, then I will investigate and we will (probably) get rid of this nasty bug. Dieter
Attention: cross post Warnes, Gregory R wrote at 2003-6-3 10:37 -0400:
... If I create a file, performing a "Save" appears to have no effect. All modified/created/delete objects remain locked by the version.
Is this a general problem with Zope?
It is a bug in "FileStorage._commitVersion", hopefully fixed by the attached patch. The object records belonging to a version form a linked list via the "pv" link field. The head of the list is pointed to by "fs._vindex[version]". When the version is committed or aborted, the linked list must be traversed to determine the set of objects modified in this version. Apparently, someone misinterpreted the traversal loop (which looks for objects belonging to the version) as following some backpointers and left the loop as soon as the first object was found. As a consequence, just this one object was moved from the version, all other objects (if any) remained with the old version (and therefore remained locked). As the code expected to commit (or abort) the complete version, "fs._vindex[version]" was later deleted such that the version appeared to be empty while there were still records associated with the version. This is a critical bug, that should soon be fixed. Dieter
This bug is fixed in CVS and in the Zope 2.6.2 beta. Jeremy
participants (8)
-
Andy McKay -
Chris McDonough -
Dieter Maurer -
Jeremy Hylton -
Oliver Bleutgen -
Paul Winkler -
Tom P. Allred -
Warnes, Gregory R