BUG: read_fwf() does not obey skip_blank_lines

[dnicolodi]

The read_fwf() function does not obey the skip_blank_lines argument:

data = io.StringIO("""\
 1 2 3
 4 5 6
""")

d = pandas.read_fwf(data, header=None, colspec=[(0, 2), (2, 4), (4, 6)])

results in:

   0  1  2
0  1  2  3
1  4  5  6

while

data = io.StringIO("""\
 1 2 3

 4 5 6
""")

d = pandas.read_fwf(data, header=None, colspec=[(0, 2), (2, 4), (4, 6)])

results in

     0    1    2
0  1.0  2.0  3.0
1  NaN  NaN  NaN
2  4.0  5.0  6.0

The skip_blank_lines argument defaults to True, but also explicitly setting it results in the same behavior. The read_fwf() function documentation says that it takes all arguments supported by TextFileReader thus this is unexpected.

I have tested this with Pandas 1.1.3 as distributed by conda, however, the relevant code does not seem to have changed on master.