L
Laura Stern
Guest
Sounds that way to me. On the other hand it could be a bug in the product. I am simply relating how the product behaves. But from looking at the code it doesn’t look as if we are doing anything explicitly different for ROWID/RECID reads based on the -rereadnolock parameter. It just never goes through the code that checks that parameter. It reads the record via a different routine that will always do the re-read. Also I was the one who implemented the -rereadnolock parameter and I don’t remember any such rule as stated by the doc. But that was about 20+ years ago, so that’s not very reliable!
Continue reading...
Continue reading...