Support build range scan plan when SQL have comparisons in combined index row

## Enhancement

### SQL
```mysql
create table t(a int, b int, c int, primary key(a, b));
explain select * from t where (a, b) > (1, 2) and (a, b) < (3, 0);
```
### Result

```
+-------------------------+----------+-----------+---------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| id | estRows | task | access object | operator info |
+-------------------------+----------+-----------+---------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| TableReader_7 | 8000.00 | root | | data:Selection_6 |
| └─Selection_6 | 8000.00 | cop[tikv] | | if(ne(test.t.a, 1), gt(test.t.a, 1), if(isnull(ne(test.t.a, 1)), NULL, gt(test.t.b, 2))), if(ne(test.t.a, 3), lt(test.t.a, 3), if(isnull(ne(test.t.a, 3)), NULL, lt(test.t.b, 0))) |
| └─TableFullScan_5 | 10000.00 | cop[tikv] | table:t | keep order:false, stats:pseudo |
+-------------------------+----------+-----------+---------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
```
Condition `(a, b) > (1, 2) and (a, b) < (3, 0)` should build to range scan

Start by reproducing the provided MySQL query and EXPLAIN output in TiDB, then trace the planner path that handles row comparisons against the combined primary index. The work is done when `(a, b) > (1, 2) AND (a, b) < (3, 0)` produces a range scan rather than a full scan with a selection filter.

Written by the indexing model from the issue text.