On Tue, Nov 9, 2010 at 12:21 AM, Jonas Sicking <[email protected]> wrote:
> This discussion seemed to die off with no clear resolution. > > Since I had forgotten about this thread I specified that the first > item is always the one returned for _NO_DUPLICATE cursors. Where > "first" means "with lowest object-store key". > It seems as though "first" should mean with the highest key in the case of reverse cursors. This is how it's implemented in Chromium. J > I don't feel strongly either way if they should be removed or not. SQL > has 'unique', but we of course we're not aiming to match SQL's feature > set. > > / Jonas > >
