[CLN] [BLD] Fix many compiler warnings

[jbrockmendel]

Checked in OSX in 2.7 and 3.7, on Ubuntu on 2.7 and 3.5, this fixes just about all the warnings that are fixable at my pay-grade.

The ones that are left:

• Ubiquitous NPY_NO_DEPRECATED_API NPY_1_7_API_VERSION --> I expect there is a compiler flag that can be set in setup.py to silence this (and only this) warning, haven't figured it out.
• In Py3.7:

pandas/_libs/src/ujson/python/objToJSON.c:430:15: warning: initializing 'char *' with an expression of type 'const char *' discards qualifiers [-Wincompatible-pointer-types-discards-qualifiers] char *data = PyUnicode_AsUTF8AndSize(obj, &len); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ pandas/_libs/src/ujson/python/objToJSON.c:436:14: warning: 'PyUnicode_EncodeUTF8' is deprecated [-Wdeprecated-declarations] newObj = PyUnicode_EncodeUTF8(PyUnicode_AS_UNICODE(obj), ^ /usr/local/Cellar/python/3.7.0/Frameworks/Python.framework/Versions/3.7/include/python3.7m/unicodeobject.h:1324:7: note: 'PyUnicode_EncodeUTF8' has been explicitly marked deprecated here ) Py_DEPRECATED(3.3); ^ /usr/local/Cellar/python/3.7.0/Frameworks/Python.framework/Versions/3.7/include/python3.7m/pyport.h:493:54: note: expanded from macro 'Py_DEPRECATED' #define Py_DEPRECATED(VERSION_UNUSED) __attribute__((__deprecated__)) ^ 3 warnings generated. 

• On Linux:

In file included from pandas/_libs/src/datetime/np_datetime.h:21:0, from pandas/_libs/src/datetime/np_datetime_strings.c:33: /usr/include/python3.5m/datetime.h:191:25: warning: ‘PyDateTimeAPI’ defined but not used [-Wunused-variable] static PyDateTime_CAPI *PyDateTimeAPI = NULL; 

• Also on Linux a bunch of "this may be unitialized" warnings.

[codecov]

Codecov Report

Merging #22013 into master will not change coverage.
The diff coverage is n/a.

@@ Coverage Diff @@ ## master #22013 +/- ## ======================================= Coverage 91.99% 91.99% ======================================= Files 167 167 Lines 50578 50578 ======================================= Hits 46530 46530 Misses 4048 4048

Flag	Coverage Δ
#multiple	90.4% <ø> (ø)	⬆️
#single	42.17% <ø> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 67b6277...c19e7e5. Read the comment docs.

[jreback]

https://travis-ci.org/pandas-dev/pandas/jobs/406697012

was this just worse before? there are still many unsigned compare with signed warnings.

[jreback]

see comments

[jbrockmendel]

was this just worse before? there are still many unsigned compare with signed warnings.

Yes. The first commit here caused test failures (that I didn't get locally), so I reverted fixes in interval.pyx.

[jreback]

ok let's hope this and your other PR fixes more warnings :>