Oy-

I don't know that we (RTI in last three years, or any time before that) ever documented the possiility that this could happen. We (as WinWin and prior) knew that this could happen. If I run a select that creates 10 T* records in the syslists file as this user on this workstation, and later ran another select that creates 3 T* records, there was a chance that it might bump into the old list.

At some point in the past we caught on to this and began using listnames in the select statements, and deleting them before we ran the statement:

Delete_Save_Select('LISTNAME')

Rlist(statement, 4, 'LISTNAME', , )

We also had a routine called 'CLEAN_LISTS' which would select all SYSLISTS records starting with 'T*', 'W*', or 'Q*' and delete them. We'd only run it when no other users were on- scheduled for the middle of the night.

I was just talking to Andrew about this issue- it basically comes down to when we know that a list has been used up, and we can delete it. A current idea is as we're writing out these temporary records we delete the next one in the sequence- eg, as I write out *1, delete *2.

So, that's where we are right now. We're still looking at and trying this for the 7.0 release. You're on the beta list, correct? Have you tried this in the new build?

Thanks

Whatever became of this? Has this issue been addressed?

Dan-

I believe it was fixed in either release 7.0 or 7.0.1

Mike

The 7.0 Readme in the Miscellaneous section says it was addressed in that release.

Bill

View this thread on the Works forum...