Hi all,
Thought only fair to update you all who responded to my query. I spent the day on site and I came away feeling the problem has something to do with the network.
Basically there is a server (file server where the foxpro data is). The server also has a client installed (local).
The problem does not happen on the local machine.
So I switched the server round. Made of of the other clients host the data. The problem stopped on the (new) server but started happening on the old server.
We put copies of data on each client, the problem goes away.
The basic code is like this:
begin Trans
if tableupdate(1,.t.,ctablea) and tableupdate(1,.t.,ctableb) and tableupdate(1,.t.,ctablec)
end trans
else
rollback
endif
There is a grid with sereral lines roughly 30 lines. The grid includes a checkbox control. Each of these lines once checked a save/cancel command buttons are enabled. To re-create the problem, we check a line at a time, hit the save button.
It appears to work ok upto around 25 lines. I don't think the exact number has any relevance only to point out that if we do it for less than 20 times we don't get the error.
Then during any of the last 5, there is a very long pause followed by error 56. Tables ctablea and ctableb have an autoInc field and one of the 2 tables is corrupt after that. The autoinc field hasn't incremented.
Why doesn't the tableupdate() fail? It would then rollback surely - or would it?
This code has been running for over 6 years with 1000s of customers running it on networks we haven't had it. This customer is on winxp, sp2, microsoft win network / microsoft client network.
Thank you and regards
>Hi all,
>
>I have a customer who keeps crashing with error 56 (VFP) : There is not enough disk space for "name" (Error 56)
>
>It a new machine with loads of space. Any clues anyone please will be much appreciated
>
>mathias
Mathias Banda
Time is longer than a rope.