TST: Improve compatibility with pypy error messages

[rlamy]

This fixes a bunch of spurious errors when running the test suite on pypy3.

• closes #xxxx
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[WillAyd]

Hmm I'm surprised pypy would impact these. Can you post pd.show_versions()?

[rlamy]

INSTALLED VERSIONS
------------------
commit           : a9085e80b0a7dd07f398f2c0b87f68b5c680402e
python           : 3.6.9.final.0
python-bits      : 64
OS               : Linux
OS-release       : 4.4.0-160-generic
machine          : x86_64
processor        : x86_64
byteorder        : little
LC_ALL           : None
LANG             : en_GB.UTF-8
LOCALE           : fr_FR.UTF-8

pandas           : 0.26.0.dev0+502.ga9085e8
numpy            : 1.18.0.dev0+68bd6e3
pytz             : 2019.2
dateutil         : 2.8.0
pip              : 19.2.3
setuptools       : 41.2.0
Cython           : 0.29.13
pytest           : 5.1.3
hypothesis       : 4.36.2
sphinx           : None
blosc            : None
feather          : None
xlsxwriter       : None
lxml.etree       : None
html5lib         : None
pymysql          : None
psycopg2         : None
jinja2           : None
IPython          : None
pandas_datareader: None
bs4              : None
bottleneck       : None
fastparquet      : None
gcsfs            : None
lxml.etree       : None
matplotlib       : None
numexpr          : None
odfpy            : None
openpyxl         : None
pandas_gbq       : None
pyarrow          : None
pytables         : None
s3fs             : None
scipy            : 1.3.1
sqlalchemy       : None
tables           : None
xarray           : None
xlrd             : None
xlwt             : None
xlsxwriter       : None

pd.show_versions() is missing an important bit of information though, i.e.

Python 3.6.9 (16472df37081, Sep 27 2019, 01:00:13)
[PyPy 7.3.0-alpha0 with GCC 6.2.0 20160901] on linux

[WillAyd]

I suspect this is probably due to numpy versioning instead of pypy but in any case I think OK. @simonjayhawkins care to look?

[simonjayhawkins]

Thanks @rlamy lgtm.

[jbrockmendel]

LGTM.

Should we consider adding a CI build for pypy?

[rlamy]

There are lots of failures ATM, I want to fix more of them before suggesting adding it to CI.

[jbrockmendel]

There are lots of failures ATM

@rlamy yep, I'm seeing 750 failures and 35 errors. Some of those are mismatching exception messages like this PR, but I'm also seeing a lot of OverflowErrors. Are you involved with pypy? i.e. are you able/willing to help troubleshoot these?

[WillAyd]

Not sure where we strayed from #9532 but @rlamy if something you are interested in supporting I would suggest opening a dedicated issue and ideally coming up with bullet points on things that need to be tackled