>>>>>I FoxPro 2.5/6 whenever there was a problem in the table header with a table an error 15 - Not a database file - was generated. This would occur when you attempted to open and consequently in most cases before anything bad happened. VFP apparently does not generate this error and consequently bad things most assuredly do happen. I WANT ERROR 15 BACK!
>>>>
>>>>Calvin,
>>>>
>>>>The instances where error 15 appeared in FPW and doesn't in VFP are usually connected with the record count in the header not matching the actual number of records. In FPW this would generate an error, in VFP it doesn't. I've used this feature a number of time to "fix" tables used by FPW. I simply open the table in exclusive mode in VFP, pack it and it's good to go in FPW again.
>>>
>>>Yes, that is nice and handy - when you have 2.6 around to show the problem in the first place.
>>
>>You don't need 2.x. Simply read the value from the header and compare it to RECCOUNT(). If they don't match, you can fix it. If you couldn't open the table because the record count was wrong, you'd have to rely on a third party tool to fix it. Now this error only occurs when the header is really hosed.
>>
>>Thanks very much, but I don't want it back the way it was before.
>
>George/Calvin
>
>The real problem is that once the header becomes corrupted, no further records are added to the table, and data is lost, until a user becomes aware of this. Fixing the table is no sweat, but the data loss is a serious flaw in data integrity. It was reported as a 'problem' in a KB article, which now acknowledges it as a bug. I would be very surprised and disappointed if this is not fixed in VFP8 (since there won't be an SP2).
>
Geoff,

Since you can't open the table under FPW, of course, this is true. However, under VFP it is possible to add/delete records. The most serious problem is when the number of records in the header is less than the actual total.

In 10+ years in dealing with data stored on a network server, I've seen this problem arise less than a handful of times. If its a persistant problem, however, it suggests that there is either a serious problem with latency or that the hardware/software combination be used is either being mis-used or is sub-standard.