The change I made this morning in loadTree (to close the cursor when
there's an Exception) is wrong. Will correct it shortly. I am working on
additional tests.
On Thu, Mar 21, 2013 at 1:24 PM, Mike McMahon <email address hidden> wrote:
> > The code path is complex, so I would appreciate confirmation that this
> is the case. But from my reading I can only see a path that reads from a
> PersistitValuePValueSource.
>
> Yes, we are good.
>
> Look at Sorter.loadTree() and its calls to createKey() and createValue().
> The entire incoming Row is encoded in the value and that is always the
> output source. This allows it to support cases where you sort on a subset
> of the columns that pass through. No particular attempt is made to suppress
> duplication (as you say, like a covering index).
> --
>
> https://code.launchpad.net/~pbeaman/akiban-persistit/fix-1157809-aioobe-on-append-key-segment/+merge/154422
> You are the owner of
> lp:~pbeaman/akiban-persistit/fix-1157809-aioobe-on-append-key-segment.
>
The change I made this morning in loadTree (to close the cursor when
there's an Exception) is wrong. Will correct it shortly. I am working on
additional tests.
On Thu, Mar 21, 2013 at 1:24 PM, Mike McMahon <email address hidden> wrote:
> > The code path is complex, so I would appreciate confirmation that this ValueSource. /code.launchpad .net/~pbeaman/ akiban- persistit/ fix-1157809- aioobe- on-append- key-segment/ +merge/ 154422
> is the case. But from my reading I can only see a path that reads from a
> PersistitValueP
>
> Yes, we are good.
>
> Look at Sorter.loadTree() and its calls to createKey() and createValue().
> The entire incoming Row is encoded in the value and that is always the
> output source. This allows it to support cases where you sort on a subset
> of the columns that pass through. No particular attempt is made to suppress
> duplication (as you say, like a covering index).
> --
>
> https:/
> You are the owner of
> lp:~pbeaman/akiban-persistit/fix-1157809-aioobe-on-append-key-segment.
>