Jos -

Thanks for taking the time to look at this and for your good wish.

Yes, the code was just an sample.

Update: It's looking like the problem never was the variables and that we only thought we "fixed" it because they never recurred at that point in the code.

The cause is narrowing to #3, #4, or #5 on your list.

In one client environment all problems disappeared when the data was moved to another server.

At another client site a bumper crop of c0000005 errors started this week, firing from anywhere and everywhere. They are moving equipment, changing offices, changing servers, replacing bad power supplies on new equipment, etc. Indexes have been rebuilt 3 times in 4 days.


>Eleanor, I can see no easy answer and there is clearly nothing wrong with the code per se. Just one point which I do not think is material - your sample code opens table1 but you do the select from table2. I assume this is just a sample code snippet.
>
>Since the code itself seems ok I would now:
>
>1) Create a mini app which just does your code snippet. Does it still happen? i.e. isolate whether its your code snippet in your app of whether its just the code snippet.
>
>2) Is your app loading any other third party libraries, activex controls etc? Can you not load them and still make it happen? i.e. isolate whether its your code snippet with these third party products or whether its just the code snippet.
>
>3) Suspect the hardware. If this occurs on more than one box then:
>
>4) Suspect the copy of the runtime files of VFP being used. Re-install clean runtimes. If it still ocurrs:
>
>5) Suspect the environment. Do all machines run a particular background app or something?
>
>In other words I dont think its VFP myself and so you now need to do the "usual suspects" thing :)
>
>Good luck!
>
>
>>Jos-
>>
>>Yes, it is easy enough to track. There is nothing complex about the code this occurs in. We are using VFP7 SP1 for development and clients either use VFP7 SP1 or VFP7 runtime files and SP1. In all cases this occurred with VFP 6 code recompiled in VFP7. The VFP 6 code ran without incident.
>>I wondered if anyone else was experiencing this.
>>
>>Example 1:
>>Local l_csubject
>>if !used("table1")
>>use table1 in 0
>>endif
>>select * from table2 into cursor csamples readwrite
>>l_csubject= .t.
>>** C0000005
>>** Adding l_csubject=.f. under the local l_csubject line fixed the problem.
>>** that makes no sense since it should already be .f.
>>
>>
>>Example 2:
>>** if lazily using the following approach:
>>xyz="my name"
>>** C0000005 (this does not occur consistently)
>>** adding xyz="" before xyz="my name" consistently fixes this problem.
>>