At 3:07 PM -0800 4/6/99, Kenneth Albanowski wrote:
>And DmResetRecordStates() doesn't crash in this context? (It's fairly
>obvious it doesn't as devices don't usually crash right about then. But
>I'm curious how it actually manages to works, for once.)
Nope, DmResetRecordStates doesn't crash there.
The DmResetRecordStates bug is that it doesn't do the write enable stuff
for the storage heap, so the first time it tries to fix a record when you
call it, it generates a permission error (or bus error, or address error,
or something like that.) When the data manager calls it from
DmOpenDatabase, the storage heap is already write enabled, so no crash.
That one will get fixed too one of these days, but there's really no point
in calling DmResetRecordStates yourself, since DmOpenDatabase does it for
you. :)
--Bob
