I wonder if it is the database itself. It it is very simple then try running recover on it and put the recovered copy back up to see it it still happens.
You could also do something like save as a clone and then reimport all the data back in which will rebuild the indexes.
and iff it is really really simple just rebuild a fresh database and export data from the old and import into the new.
That would mean any index corruption and or layout corruption or other corruption is fixed in the database and it would either fix the problem or at least eliminate the database as the issue.
You could also do something like save as a clone and then reimport all the data back in which will rebuild the indexes.
and iff it is really really simple just rebuild a fresh database and export data from the old and import into the new.
That would mean any index corruption and or layout corruption or other corruption is fixed in the database and it would either fix the problem or at least eliminate the database as the issue.
<[email protected]> wrote:
Please do keep in mind that this was not unique to Andrew's situation. I reported the very same problem on or about the same day. I had not see this problem in any of my prior projects, so it appears perhaps it may actually be a bug (a.k.a. flaw in the plumbing).
With so many OS updates that come out nowadays in both Windows & OSX, and a couple of you have mentioned the dependence on AppleEvents, could there be some connection to an automatic OS update resulting in what was assumed to be a protected single-threaded database connection to actually mix up results between database threads? Just a thought... anyone have any ideas?
Thanks.
Roland Dumas wrote:
Ok, I have a copy of his FMP file, the taf, and dropped it on my server. I made some tweaks to clean up some things, but found nothing wacky enough to cause two users to get the same results. Specifically, two users hitting the same _function with different search arguments getting the same results - correct for one visitor and incorrect for the other. If they run the same queries with some time in between them, they get the right answers for each.
It APPEARS that with FMP as the datasource and using appleevents connection, the single threaded database is returning visitor #1's resultset to visitor #2.
The FMP table is very simple and small. Queries sometimes run very slowly, slower than I experience with larger tables in FMP.
This really does look like a flaw in the plumbing, not the coding.
On Feb 28, 2006, at 3:18 PM, Andrew Derry wrote:
Hello,
We set up a few simple Tango apps a few years ago that I believe were Tango3.6 - Which were not all that long ago ported up to Witango Server for OS X,currently v5.0.1.065.
One app is a simple FileMaker DB search (If you can call FileMaker a DB?) ;)
However since the upgrade, people have complained about getting resultsother than for what they searched... It turns out that if you hit the pagefrom two different browsers at very close to the same time, they'll both getthe results of one of the searches.
It's basically just a simple form where the user just selects some valuesout of the table to bring up the resulting matching records.
We just go:
<FORM METHOD=POSTACTION="">
And then the TAF just branches on function and does a search on the@POSTARG's
Actually, it didn't have that @USERREFERENCEARGUMENT at first, I added thatout of desperation... But no luck...
I also thought to check the DEFAULTSCOPE in the witango.ini - which for somereason was set to request, which I couldn't even find in the help (though myhelp seems to not be easily searchable) - So I tried setting it to user andre-starting the witangod, but it didn't make any difference either.
Any ideas, anyone..? Do you need more info? It's been at least a year ortwo since I've had to do any Witango programming, and I never did doanything particularly complicated, so it might just be something obvious toothers..
Any thoughts or suggestions would be very much appreciated.
Thanks
- Andrew Derry
--Andrew Derry - [EMAIL PROTECTED]Simon Fraser UniversityApplication Integration Support - MBC1415Burnaby, BCCanada V5A 1S6
________________________________________________________________________TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Roland A. Dumas310 W. Bellevue Ave.San Mateo, CA 94402650-347-1373415-412-9300 (cell)AIM: radumas
________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
-- Dan Stein MSN Certified FileMaker 7 Developer Digital Software Solution, LLC 799 Evergreen Circle 215-799-0192 610-256-2843
________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
