rql_test failes statistically

When running rql_test sometimes the `split` function gets stuck and gets to a point of out of memory.
After a lot of debugging i found out that the way the `equalExp` function implemented is wrong.

The check
```go
if s1[i][0] == '(' && s2[j][0] == '(' {
found = equalExp(s1[i][1:len(s1[i])-1], s2[j][1:len(s2[j])-1])
}
```

is wrong because if we get `(` at the start we are removing the end of the string, not the next `)` found which is the problem
there is a corruption of data `(a = b AND c = d) OR (a = c AND (c = d OR d = f))` will be `a = b AND c = d) OR (a = c AND (c = d OR d = f)`

The problem starts here and get to the `split` function which is not validating the given data and not breaking out of the `while` loop after it failes at the start with finding `)`

Look at the rql_test file and the equalExp function. The bug is in how parentheses are stripped: it removes the last character instead of the matching ')'. The split function then gets stuck. Start by reproducing the failure in rql_test, then examine the parsing logic around parentheses. The fix involves correcting the substring logic and ensuring split validates input to avoid infinite loops.

Scritto dal modello di indicizzazione a partire dal testo della issue.