The last time I saw one of these was around 1982 on a Reality.
So. Resurrected in D3/SCO 7.1.15. Here are the circumstances.
In accessing a file using a BASIC program which has been well behaved
for about 6 years, the user got some kind of Referencing Illegal Frame
abort which is absolutely reproducable. To get out, the user types OFF.
The next file-save reports a bad count field, so we investigate. This
turns out to be in the middle of a group in a well sized file ( no
overflow) consisting of very small records. Typically under 30 bytes
each record including the count fields.
We use the gfe handler to fix the offending data and lose about 40
records which we restore from the previous night's file save.
Coincidentally ;) the RIF problem goes away at the same time.
I have very good reason to believe there was no data changed in the
group, although new records may have been added and therefore would have
gone to the end of the group.
So we treated and solved the problem in the time honoured way, but the
question that I cannot answer, and would like at least a clue to, is
what do you think are the possible causes of a bad count field in stable
Martin Taylor (Author of "Pick for Users" and half of "Unix and Unidata")
Datamatters Ltd http://www.dmatters.co.uk
13 Market Place, Heywood, Lancashire, UK