Re: sidnei's reindexing improvements

Rob Miller wrote:
> Sidnei da Silva wrote:
> 
>> On Fri, Dec 09, 2005 at 07:18:43PM -0200, Sidnei da Silva wrote:
>> | Yes, call it band-aid-stop-gap-measure if you want, but it's a proper
>> | fix, just the size that fits in a minor revision (1.3.6) and that took
>> | about two years with no action from anyone. Please line up your
>> | proposals for the next major revision of archetypes.
>>
>> Just to be clear, I'm not against using events. I'm against waiting
>> for the next-next tecnology when we can fix (or band-aid, or stopgap)
>> existing bugs with existing resources and *just then*, when the bug is
>> at least alleviated, wait for the next-next technology.
> 
> 
> we're on the same page here.  remember, my first message was to make 
> sure that your changes DID get merged.
> 
> for some reason, though, i think i'm seeing more multiple unindexing 
> errors when i run the CMFPlone tests using AT from the plone 2.1 bundle 
> than i am using your reindexing branch... that's why i thought the 
> changes hadn't been merged in the first place.  i'm re-running those 
> tests now, though; it's possible it was operator error causing the 
> problems.

okay, i don't think i'm off base here.  i'm running the CMFPlone tests 
using the plone 2.1 bundle.  the first thing i notice is that one of the 
testActionTool tests is failing (*grumble*).  aside from that, however, 
something seems to have changed to take away some of the gains that 
sidnei's changes made.

when i use the AT 1.3 release branch that is part of the bundle, the 
tests take about 831 seconds (on my 1.25GHz PowerBook) and there are 
1137 lines of output from the './bin/zopectl test' command that i 
executed until the 'Ran 1536 tests in... ' notifier.

when i remove the 1.3 release branch and drop in the 
sidnei-reindexing-sanity branch, with no other changes, the tests take 
627 seconds to run, and there is 663 lines of output.  a quick peek at 
the output shows that the difference is mostly due to a bunch of 
additional "Zope.ZCatalog:uncatalogObject unsuccessfully attempted to 
uncatalog an object" errors.

this is hardly a rigorous evaluation, i know, but it's all i have time 
to do right now.  anyone else have a moment to see if you get the same 
results and, if so, figure out what's causing them?

-r



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click