BUG: resample on TimedeltaIndex with closed=right takes no effect

[tfiasco]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

In [1]: import pandas as pd

In [2]: index = pd.Series([pd.Timedelta(seconds=120 + i * 30) for i in range(10)])

In [3]: series = pd.Series(range(10), index=index)

In [4]: series
Out[4]: 
0 days 00:02:00    0
0 days 00:02:30    1
0 days 00:03:00    2
0 days 00:03:30    3
0 days 00:04:00    4
0 days 00:04:30    5
0 days 00:05:00    6
0 days 00:05:30    7
0 days 00:06:00    8
0 days 00:06:30    9
dtype: int64

In [5]: series.resample('T', closed='left', label='left').sum()
Out[5]: 
0 days 00:02:00     1
0 days 00:03:00     5
0 days 00:04:00     9
0 days 00:05:00    13
0 days 00:06:00    17
Freq: T, dtype: int64

In [6]: series.resample('T', closed='right', label='right').sum()  # WRONG!
Out[6]: 
0 days 00:02:00     1
0 days 00:03:00     5
0 days 00:04:00     9
0 days 00:05:00    13
0 days 00:06:00    17
Freq: T, dtype: int64

Issue Description

closed=right argument takes no effect, the output is same as closed=left.

Expected Behavior

The same computation on a DatetimeIndex, which is the expected behavior:

In [7]: index = pd.Series([pd.Timedelta(seconds=120 + i * 30) for i in range(10)])

In [8]: index = index + pd.to_datetime('2000-01-01')

In [9]: series = pd.Series(range(10), index=index)

In [10]: series
Out[10]: 
2000-01-01 00:02:00    0
2000-01-01 00:02:30    1
2000-01-01 00:03:00    2
2000-01-01 00:03:30    3
2000-01-01 00:04:00    4
2000-01-01 00:04:30    5
2000-01-01 00:05:00    6
2000-01-01 00:05:30    7
2000-01-01 00:06:00    8
2000-01-01 00:06:30    9
dtype: int64

In [11]: series.resample('T', closed='left', label='left').sum()
Out[11]: 
2000-01-01 00:02:00     1
2000-01-01 00:03:00     5
2000-01-01 00:04:00     9
2000-01-01 00:05:00    13
2000-01-01 00:06:00    17
Freq: T, dtype: int64

In [12]: series.resample('T', closed='right', label='right').sum()  # RIGHT
Out[12]: 
2000-01-01 00:02:00     0
2000-01-01 00:03:00     3
2000-01-01 00:04:00     7
2000-01-01 00:05:00    11
2000-01-01 00:06:00    15
2000-01-01 00:07:00     9
Freq: T, dtype: int64

Installed Versions

INSTALLED VERSIONS

commit : 66e3805
python : 3.8.8.final.0
python-bits : 64
OS : Linux
OS-release : 4.9.130-0409130-generic
Version : #201809291131 SMP Sat Sep 29 11:33:40 UTC 2018
machine : x86_64
processor :
byteorder : little
LC_ALL : None
LANG : None
LOCALE : en_US.UTF-8

pandas : 1.3.5
numpy : 1.19.5
pytz : 2021.1
dateutil : 2.8.1
pip : 21.0.1
setuptools : 52.0.0.post20210125
Cython : None
pytest : 6.2.3
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : 1.0.2
psycopg2 : None
jinja2 : 2.11.3
IPython : 7.22.0
pandas_datareader: None
bs4 : None
bottleneck : None
fsspec : 0.9.0
fastparquet : 0.7.1
gcsfs : None
matplotlib : 3.3.4
numexpr : None
odfpy : None
openpyxl : 3.0.7
pandas_gbq : None
pyarrow : 6.0.1
pyxlsb : None
s3fs : None
scipy : 1.6.2
sqlalchemy : 1.4.23
tables : None
tabulate : 0.8.9
xarray : 0.19.0
xlrd : 2.0.1
xlwt : None
numba : 0.53.1

[attack68]

#40245 should deprecate this arg in favour of inclusive