Reply-To: erik.wachtmeester@bighole.iaf.nl
Raymond Pesek wrote in a message to Erik Wachtmeester:
Sorry for the delayed response, but we've been trying to get answers from 
other
channels also. This took a while.
EW ->... but afterwards it sometimes happens that even though the
   ->program thinks that a new record is appended, that actually
   ->isn't the case, and consequently existing records can be
   ->overwritten.
 RP> Some people have reported an identical problem with Visual Objects 
 RP> 2.0 in that the append returned .T., and an existing record was
 RP> overwritten because a new record was not actually appended. The fellow
 RP> who did the most testing is absolutely positive that it only happens if
 RP> the offending workstation is not the "bought-in-the-store" Windows 95a
 RP> version, but rather the "OEM-only OSR2 Windows 95b" version.
Sorry, but all machines we receive with OSR2 on them are promptly reformatted
with W95a (company policy, keep all versions the same throughout the 
etwork),
so OSR2 cannot be the offender in our case.
 RP> Please report back on what you find. If true, this could be _real_ bad.
We haven't found a real clue yet. It seems to be a combination of NetWare 
.11
(as up to date as possible), Client32 2.11 and Successware's SIx RDD 2.03 and
Clipper 5.2e. Novell points us to NW3.12 or 4.11, Successware advises us to 
use
v3 of their RDD, and CA tells us that Clipper 5.3 is much better under W95.
All this may be possible, but for the future releases of our software we've
chosen Delphi and InterBase. And in our Clipper code we're making a lot of 
se
of Clipper Tools 3, so any upgrade on the Clipper-part would be quite a 
problem.
Regards,
Erik 
---
---------------