Yesterday I upgraded a client's application from VFP7 to VFP8. Before upgrading I found that there were several tables that triggered error 2091 in VFP8. I added code to repair the tables to my error handler and it worked well. After upgrading the user, the table repair worked and the user was able to operate. However, after about 30 minutes, one of the tables became "corrupted" again. They had to get everyone out and let the error handler repair the table again. About 15 minutes after that, they had another table trigger error 2091. My error handler checks every table in the application directory and fixes every table it finds that shows a corruption. The application does not use a .dbc or transactions. It has run very reliably for a long time at several locations.

They are running a mixture of Windows 98 and Windows XP computers.

For now, I have set tablevalidate to 0 and redistributed the app, but I would like to know how to track down what is causing the problem. Throughout the course of their operation, 3 different files became corrupted, all from different parts of the software, however they have never complained of data loss. I am at a loss!