Instead of using pgproto3.FieldDescription through pgconn and pgx. This
lets the lowest level pgproto3 still be as memory efficient as possible.
https://github.com/jackc/pgx/pull/1281
Previously, a batch with 10 unique parameterized statements executed
100 times would entail 11 network round trips. 1 for each prepare /
describe and 1 for executing them all. Now pipeline mode is used to
prepare / describe all statements in a single network round trip. So it
would only take 2 round trips.
- Remove rarely used ScanPlan.Scan arguments
- Plus other refactorings and fixes that fell out of this change.
- Plus rows Scan now handles checking for changed type.
This patch fixes jackc/pgx#841. The meat of the fix lives
in [a PR to the pgconn repo][1]. This change just checks
for errors after executing a prepared statement and informs
the underlying stmtcache about them so that it can properly
clean up. We don't try to get fancy with retries or anything
like that, just return the error and allow the application to handle it.
I had to make [some][1] [changes][2] to to the jackc/pgconn package as well
as this package.
Fixes#841
[1]: https://github.com/jackc/pgconn/pull/56
[2]: https://github.com/jackc/pgconn/pull/55
Previously, the Scan documentation stated that scanning into a []byte
will skip the decoding process and directly copy the raw bytes received
from PostgreSQL.
This has not been true for at least 2 months. It is also undesirable
behavior in some cases such as a binary formatted jsonb. In that case
the '1' prefix needs to be stripped to have valid JSON. If the raw
bytes are desired this can easily be accomplished by scanning into
pgtype.GenericBinary or using Rows.RawValues.
In light of the fact that the new behavior is superior, and that it has
been in place for a significant amount of time, I have decided to
document the new behavior rather than change back to the old behavior.
