bug: window expression does not work with preceding compound interval

[printsg]

What happened?

Rolling aggregation with compound preceding interval, e.g.,1 month - 1 day, produces wrong SQL.

import ibis
from datetime import datetime

data = {
    "id": [1, 1, 1],
    "date": [
        datetime(2024, 1, 1),
        datetime(2024, 1, 10),
        datetime(2024, 1, 20),
    ],
    "value": [10.0, 20.0, 30.0]
}
t = ibis.memtable(data)

# interv = ibis.interval(months=-1) # works! 
interv = ibis.interval(months=-1, days=1) # "Out of Range Error: Invalid RANGE FOLLOWING value"

win = ibis.window(
    group_by=["id"],
    order_by=["date"],
    between=(interv, 0),
)
expr = t.mutate(
    rolling_sum=t.value.sum().over(win)
)
# ibis.to_sql(expr)
expr.execute()

SQL:

SELECT
  "t0"."id",
  "t0"."date",
  "t0"."value",
  SUM("t0"."value") OVER (
    PARTITION BY "t0"."id"
    ORDER BY "t0"."date" ASC
    RANGE BETWEEN INTERVAL '1' DAY + INTERVAL '-1' MONTH following AND TO_DAYS(CAST(0 AS INT)) following
  ) AS "rolling_sum"
FROM "ibis_pandas_memtable_u4ikv6hl3rd6tfmfnzvlek5hb4" AS "t0"

What version of ibis are you using?

11.0.0

What backend(s) are you using, if any?

DuckDB

Relevant log output

Code of Conduct

• I agree to follow this project's Code of Conduct