Yet Another 2GB Problem

[ieo at gmx.ph]

group, hope its all good on your end
going to get right into it...

dev environment details:
W7x64,fP5.7.00,NTFS
key/data filepair and data is in excess of 2GB 

no issue in production IUA'ing to this file now ~2.6GB

however

-adding fields corrupts the set resulting in the segment mismatch error
-expanding the file any number of records reports invalid argument error

i did some poking around..

1st) tested fpcopy workflows. (with data, without data)
used copy to try chunks of records thinking perhaps corruption in the set somewhere
only failed results after 2GB.

i was thinking perhaps a pre5.0 dataset in the 5.7 environment had some compatability issues?

2nd.0) created an empty file expanded data to a size under 2gb. results: success with field adds/reconstruction
2nd.1) add a field and reconstruct so the file is now over 2gb. results: IUA works,expand fails
2nd.2) now the file is over 2gb a second attempt to add field. results: segment mismatch corruption

i thought the 2gb bug was squashed as it fixed our issue for production.


any insight into the behavior?


i also noticed when expanding the empty file the first attemt would limit the expand much less than available (100,000)
after expanding to 100,000 it would allow a more saturating 700,000 
i set pfNODF in hopes that the freespace bug was related to the main problem but not such luck.


i realize the switch/extend is a trick of the trade of yesteryear. ill be testing that workflow.
hoping for some input here before moving onto that. understanding is paramount

cheers