Almost forgot, there was another serious bug that came to light during the same conversion in my previous post. A report update with 2 file accesses, 2nd file access is on a control break. Even when the read of the primary file failed (keys in default logic were not valid) the LS continued through the logic on the file with the control break (the control break was not &$EOF, which might make sense). I have seen this on different systems in the past 2 weeks so it is not isolated. THe other system was also 4.6 although it was on AIX.

I am curious. What was the sub version of 1.5 you are upgrading from? We have found differences depending on the release, 1.5103r44 onwards upgrades smoothly with no apparant problems, upgrading from r13 and roundabouts require a bit of reworking where windows are concerned. All our new development is in 4.6, the only two difficulties we have had is the installation kit for r100 (unix) missing some files and the non-functioning of @TIMEOUT under NT. We are not experiencing ANY stability problems on either NT or Unix (Indeed 4.6r107 NT allows the killing of hung processes without resorting to stopping the service, a major benefit ) There are some problems with the new facilities in 4.6, but when has anything new ever worked straight out of the box!!!!!! All files are Pro-Isam.

Trev.