L
Laura Stern
Guest
I haven’t read the kbase entry. So this may be redundant. But it sounds like the kbase doesn’t really explain why the error is happening again. You have an invalid value in the tt record, which stays there because the modification was not undone. Any operation that would cause the record to be flushed out (like reading another record into the same buffer or doing a release), will cause the validation and thus the error to happen again, ad infinitum. This bug was reported before, and if my memory serves me correctly, to “fix” it, the AVM was modified to treat delete as a special case that would not complain and make the error stop. I suppose you have a case for the EMPTY-TEMP-TABLE method doing the same. Though it seems to me that making the TT UNDO might be the cheaper option!
Continue reading...
Continue reading...