Incorrect results when using `daft.lit(1).over(window)` (at least, they differ from duckdb / pyspark)

[MarcoGorelli]

Describe the bug

The code below outputs

daft
╭───────┬───────┬────────╮
│ a     ┆ i     ┆ b      │
│ ---   ┆ ---   ┆ ---    │
│ Int64 ┆ Int64 ┆ UInt64 │
╞═══════╪═══════╪════════╡
│ 1     ┆ 0     ┆ 1      │
├╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌┤
│ 1     ┆ 1     ┆ 1      │
├╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌┤
│ 2     ┆ 2     ┆ 1      │
╰───────┴───────┴────────╯

(Showing first 3 of 3 rows)
pyspark
+---+---+---+
|  a|  i|  b|
+---+---+---+
|  1|  0|  2|
|  1|  1|  2|
|  2|  2|  1|
+---+---+---+

To Reproduce

import daft
from pyspark.sql import SparkSession
import pyspark.sql.functions as F
from pyspark.sql.window import Window as W
from datetime import datetime

data = {"a": [1, 1, 2], "i": [0, 1, 2]}

print("daft")
print(
    daft.from_pydict(data)
    .with_column(
        "b", daft.functions.count(daft.lit(1)).over(daft.Window().partition_by("a"))
    )
    .collect()
)


session = SparkSession.builder.getOrCreate()
rows = [
    {key: value[i] for key, value in data.items()}
    for i in range(len(data[next(iter(data.keys()))]))
]
df = session.createDataFrame(rows)
df = df.withColumn("b", F.count(F.lit(1)).over(W.partitionBy("a")))
print("pyspark")
df.show()

Expected behavior

Same as pyspark (note: duckdb also matches pyspark)

Component(s)

Built-in Functions

Additional context

Narwhals