On Sat, Apr 12, 2014 at 02:10:13PM -0400, Jan Wieck wrote:
> Since it doesn't seem to produce any side effects, I'd think that
> making the snapshot unique within txid_current_snapshot() and
> filtering duplicates on input should be sufficient and eligible for
> backpatching.
Agreed.
> The attached patch adds a unique loop to the internal
> sort_snapshot() function and skips duplicates on input. The git
> commit is here:
>
> https://github.com/wieck/postgres/commit/a88a2b2c25b856478d7e2b012fc718106338fe00
> static void
> sort_snapshot(TxidSnapshot *snap)
> {
> + txid last = 0;
> + int nxip, idx1, idx2;
> +
> if (snap->nxip > 1)
> + {
> qsort(snap->xip, snap->nxip, sizeof(txid), cmp_txid);
> + nxip = snap->nxip;
> + idx1 = idx2 = 0;
> + while (idx1 < nxip)
> + {
> + if (snap->xip[idx1] != last)
> + last = snap->xip[idx2++] = snap->xip[idx1];
> + else
> + snap->nxip--;
> + idx1++;
> + }
> + }
> }
I think you need to do SET_VARSIZE also here. Alternative is to
move SET_VARSIZE after sort_snapshot().
And it seems the drop-double-txid logic should be added also to
txid_snapshot_recv(). It seems weird to have it behave differently
from txid_snapshot_in().
--
marko
