sql: FETCH from DECLARE CURSOR does not reset statement_timeout

Today, after declaring a cursor, `FETCH` is subject to the `statement_timeout`. However, the timeout is from the beginning of `DECLARE CURSOR`, not from the `FETCH`.

See this reproduction:

```
root@127.0.0.1:26257/movr> set statement_timeout = '5s';
SET

Time: 1ms total (execution 1ms / network 0ms)

root@127.0.0.1:26257/movr> CREATE TABLE t (a int primary key);
CREATE TABLE

Time: 6ms total (execution 6ms / network 0ms)

root@127.0.0.1:26257/movr> insert into t values (1), (2);
INSERT 0 2

Time: 7ms total (execution 6ms / network 0ms)

root@127.0.0.1:26257/movr> begin;
BEGIN

Time: 0ms total (execution 0ms / network 0ms)

root@127.0.0.1:26257/movr OPEN> declare c cursor for SELECT * FROM t;
DECLARE CURSOR

Time: 3ms total (execution 3ms / network 0ms)

root@127.0.0.1:26257/movr OPEN> fetch 1 from c;
a
-----
1
(1 row)

Time: 1ms total (execution 1ms / network 0ms)

root@127.0.0.1:26257/movr OPEN> -- wait 5 seconds ......
-> ;
OK

Time: 0ms total (execution 0ms / network 0ms)

root@127.0.0.1:26257/movr OPEN> fetch 1 from c;

ERROR: query execution canceled due to statement timeout
SQLSTATE: 57014
```

quote @rafiss
> ah actually maybe the fix should be to make sure fetch always resets the timer, so that the calculation here doesn't hit the timeout https://github.com/cockroachdb/cockroach/blob/e402d828641007572b0e44ae70a628ce5bf0982c/pkg/sql/conn_executor_exec.go#L530

otan note:
> hmm i'm not sure it's so straightforward to change conn_executor_exec - the abstraction for calling FETCH is all hidden behind the InternalExecutor/connExecutor and pausing and resuming the timeouts on each FETCH is difficult. i think we have to move the statement timeout logic to fetchNode to make this work.
>
> if we still want statement_timeout for fetch, we should set the statement timeout to 0s on the internalexecutor setting up the cursor. if we really want it to work soon & backported, i'd recommend a session variable that overrides statement timeouts for 22.2 for DECLARE CURSOR - true by default for 23.1+, false by default 22.2

Jira issue: CRDB-24063

Epic CRDB-60818

Start with the DECLARE CURSOR/FETCH reproduction in the issue, then read pkg/sql/conn_executor_exec.go around the referenced timeout calculation and inspect fetchNode, which the discussion identifies as a possible location. Determine how statement_timeout is applied across DECLARE and FETCH; done means a later FETCH is timed from its own execution while still respecting the intended timeout behavior, with coverage for the reproduction.

Written by the indexing model from the issue text.