* v1.0d15.4 -  1/16/92 - Fixed bug in HEPasteInternal where the vertical scroll
*                        bar wasn't being invalidated. If you had a record with
*                        an amount of data such that the scroll bar was
*                        inactive, adding new bytes so that it should become
*                        active would not update it properly.
*                      - Since a zero-length resource has no real context (and
*                        who knows, it may kill the Resource Manager), when
*                        attempting to delete the last byte in the resource,
*                        the shell's delete item code is called (via the
*                        fSelectorRequest callback). This gives the user the
*	     option of deleting the resource, or canceling the
*                        delete/clear operation.
*                      - When 'Edit Item...' is chosen, first priority is now
*                        given to ScriptEdit. If it can't find a script for
*                        the type, an internal dispatch to HexEdit occurs.
*                      - Modified all calls to fLoadResource that used the
*                        special = $FFFF (detach) to call fDetachResource
*                        directly.
*                      - Inserted some paranoia HLocks during HEClearInternal
*                        and HEPasteInternal calls. I _think_ it would be ok
*                        during sub-calls to HEUpdateInternal, but I'll have to
*                        check carefully.
*                      - There is some really weird behaviour in largish
*                        records during rapid deletes/cursor mode changes. Data
*                        doesn't appear to be corrupted, but for a thus far
*                        unexplicable reason, the screen is 'jumping' during
*                        the delete update. I've no idea why...
*                      - More wacky unexplainable behavior involves the 
*                        underscore ('_') character. Find one of them in a 
*                        record somewhere and put the cursor in the same column.
*                        Scroll the record so the char moves off of the screen.
*                        As you scroll it back onto the screen and under the
*                        ASCII cursor, the cursor is left on the screen. This
*                        is the most bizarre behaviour I have ever seen, since
*                        cursor hilighting and drawing has no relationship
*                        whatsoever to the character in that position.
