No SettingWithCopyWarning for chained indexing when .loc or .iloc is first indexer

[tdpetrou]

Code Sample, a copy-pastable example if possible

>>> df = pd.DataFrame(np.random.rand(3,4), index=['a', 'b', 'c'], columns=['w','x', 'y', 'z'])

>>> df.loc[['a', 'b']]['w'] = 1  # No warning, no change
>>> df.loc['a':]['w'] = 1 # No warning, no change
>>> df.loc['a':'b']['w'] = 1 # Warning, no change
>>> df.loc['a':'b']['w'] = .5 # Warning, modified
>>> df.loc['a']['w'] = 1 # No warning, modified
>>> df.loc['a', ['w', 'x']]['w'] = 99 # No warning, no change
>>> df.loc[df.x > .2]['w'] = 111 # Warns only when boolean index contains a mix of True/False, no change

>>> df.iloc[[0, 2]]['w'] = 99 # Warning, no change
>>> df.iloc[:2]['w'] = 99 # Warning, no change
>>> df.iloc[2]['w'] = 99 # No warning, modified
>>> df.iloc[:2, :2]['w'] = 100 # No warning, no change

Problem description

I am completely confused by this behavior. I would expect that all of these should trigger warnings as they are all using chained indexing to make assignments. There is one example from each .loc\iloc that does modify df.

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.3.final.0
python-bits: 64
OS: Darwin
OS-release: 15.6.0
machine: x86_64
processor: i386
byteorder: little
LC_ALL: None
LANG: en_US.UTF-8
LOCALE: en_US.UTF-8

pandas: 0.21.0
pytest: 3.2.1
pip: 9.0.1
setuptools: 36.5.0.post20170921
Cython: 0.26.1
numpy: 1.13.3
scipy: 0.19.1
pyarrow: None
xarray: None
IPython: 6.1.0
sphinx: 1.6.3
patsy: 0.4.1
dateutil: 2.6.1
pytz: 2017.3
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.2
feather: None
matplotlib: 2.1.0
openpyxl: 2.4.8
xlrd: 1.1.0
xlwt: 1.2.0
xlsxwriter: 1.0.2
lxml: 4.1.0
bs4: 4.6.0
html5lib: 0.999999999
sqlalchemy: 1.1.13
pymysql: None
psycopg2: None
jinja2: 2.9.6
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: 0.5.0