use py.builtin._basestring for py2/3 compatibility

[montefra]

when visiting a path, check that the input values are a string in a python2/3 compatible way. See issue #81

[RonnyPfannschmidt]

good work, for avoiding regressions, can you perhaps put the type variation into the tests for the feature as well?

[montefra]

@RonnyPfannschmidt : this should do.

I'll think if there is an easy way to test it. I don't think that importing the __future__ is ok. Probably I need something using pytest.mark.parametrize and skipif.

Do I have to add some changelog or similary anywhere else?

Edit:

I was writing when you replied. I'll do it soon.

[montefra]

I did try to run tox but I got E.g.

py34 create: /data01/montefra/Code/Gitrepos/py/.tox/py34
py34 installdeps: pytest~=2.9.0
py34 inst: /data01/montefra/Code/Gitrepos/py/.tox/dist/py-1.4.32.dev1.zip
py34 installed: py==1.4.32.dev1,pytest==2.9.2
py34 runtests: PYTHONHASHSEED='2583340150'
py34 runtests: commands[0] | py.test --confcutdir=.. -rfsxX --junitxml=/data01/montefra/Code/Gitrepos/py/.tox/py34/log/junit-py34.xml
Traceback (most recent call last):
  File "../.tox/py34/bin/py.test", line 7, in 
    from pytest import main
  File "/data01/montefra/Code/Gitrepos/py/.tox/py34/lib/python3.4/site-packages/pytest.py", line 21, in 
    from _pytest.config import (
  File "/data01/montefra/Code/Gitrepos/py/.tox/py34/lib/python3.4/site-packages/_pytest/config.py", line 8, in 
    import py
ImportError: No module named 'py'
ERROR: InvocationError: '/data01/montefra/Code/Gitrepos/py/.tox/py34/bin/py.test --confcutdir=.. -rfsxX --junitxml=/data01/montefra/Code/Gitrepos/py/.tox/py34/log/junit-py34.xml'   

ps: I can't access the documentation: http://pylib.readthedocs.io/en/latest/ returns me:

Welcome to Read the Docs
This is an autogenerated index file.

Please create a /home/docs/checkouts/readthedocs.org/user_builds/pylib/checkouts/latest/doc/index.rst or /home/docs/checkouts/readthedocs.org/user_builds/pylib/checkouts/latest/doc/README.rst file with your own content.

If you want to use another markup, choose a different builder in your settings.

[RonnyPfannschmidt]

the docs issue is my fault, we will reconfigure readthedocs sometime today

[montefra]

on a side note: can I add python 3.5 in tox.ini?

[RonnyPfannschmidt]

of course, please do so, i forgot doing it after the move

[montefra]

I have added py35 environment. But I've noticed one thing that puzzles me testing/path/common.py is not run by tox, nor by Travis CI. Why is that? My changes are not run in the test suite. And I have no idea how to do it: I did go a bit through the fixtures and setups and I haven't understood much of it.

This is what I get:

-> tox -e py35          
GLOB sdist-make: /data/Codes/Gitrepos/py/setup.py
py35 inst-nodeps: /data/Codes/Gitrepos/py/.tox/dist/py-1.4.32.dev1.zip
py35 installed: py==1.4.32.dev1,pytest==2.9.2
py35 runtests: PYTHONHASHSEED='3938473908'
py35 runtests: commands[0] | py.test --confcutdir=.. -rfsxX --junitxml=/data/Codes/Gitrepos/py/.tox/py35/log/junit-py35.xml
========================================================================= test session starts ==========================================================================
platform linux -- Python 3.5.1, pytest-2.9.2, py-1.4.32.dev1, pluggy-0.3.1
rootdir: /data/Codes/Gitrepos/py, inifile: tox.ini
collected 478 items / 4 skipped 

test_iniconfig.py ...........................................
code/test_assertion.py ...............ss.......
code/test_code.py ..............
code/test_excinfo.py ...........s........s....................................................
code/test_source.py .....................................................................
io_/test_capture.py ..................s.............s.............s................s.....................
io_/test_saferepr.py ........
io_/test_terminalwriter.py ............................................
log/test_log.py .................
log/test_warning.py x.....
path/test_cacheutil.py .............
process/test_cmdexec.py .....
process/test_forkedfunc.py ..............
process/test_killproc.py .
root/test_builtin.py ...................
root/test_error.py .....
root/test_py_imports.py ................
root/test_std.py ...
root/test_xmlgen.py ...................

----------------------------------------------- generated xml file: /data/Codes/Gitrepos/py/.tox/py35/log/junit-py35.xml -----------------------------------------------
======================================================================= short test summary info ========================================================================
SKIP [4] testing/io_/test_capture.py:236: text output different for bytes on python3
SKIP [4] /data/Codes/Gitrepos/py/testing/path/common.py:445: sys.version_info < (3,6)
SKIP [1] /data/Codes/Gitrepos/py/testing/code/test_excinfo.py:288: could not import 'jinja2'
SKIP [2] /data/Codes/Gitrepos/py/testing/code/test_assertion.py:170: could not import 'py._code._assertionold'
SKIP [1] /data/Codes/Gitrepos/py/testing/code/test_excinfo.py:183: could not import 'decorator'
XFAIL log/test_warning.py::test_forwarding_to_warnings_module

[RonnyPfannschmidt]

Good find, ill check

[RonnyPfannschmidt]

there was a mistake in a pr i merged previously causing all tests to be skipped

fortunately pytest 3.0 made me find it, and i'm cleaning things up now

[montefra]

@RonnyPfannschmidt : Good to know.

Let you know when you are done so I'll merge master into this branch and see if the tests run.

[RonnyPfannschmidt]

@montefra currently the needed changes are on #88 - since its large i'd like to have another pair of eyes from a core maintainer on it just in case i missed something

[montefra]

Thank for the work. I'll wait for it to land in master before proceeding.

[RonnyPfannschmidt]

@montefra sorry for the delay, the cleanups on the other branch started to break out of proportion i started #90 to stop blocking your good work and get something to the point done

[RonnyPfannschmidt]

@montefra the new run exposed a typo, also now your tests run as expected, thanks for the patience

[montefra]

Thank you. I'll work on this branch in the next couple of days.

[montefra]

@RonnyPfannschmidt : do you prefer me to merge or rebase this branch on master?

[RonnyPfannschmidt]

@montefra please rebase so we can keep the number of merges to a minimum

[RonnyPfannschmidt]

@montefra good work

interesting, any idea why did the byte string fail on python3?

[montefra]

I was about to write about this:

self = <py._path.common.Visitor object at 0x7f58774094a8>, path = svnwc('/tmp/pytest-of-susefra/pytest-17/test_make_repo0/path1wc/otherdir')

    def gen(self, path):
        try:
            entries = path.listdir()
        except self.ignore:
            return
        rec = self.rec
        dirs = self.optsort([p for p in entries
                    if p.check(dir=1) and (rec is None or rec(p))])
        if not self.breadthfirst:
            for subdir in dirs:
                for p in self.gen(subdir):
                    yield p
        for p in self.optsort(entries):
>           if self.fil is None or self.fil(p):
E           TypeError: 'bytes' object is not callable

Byte string are not base_string in python3, so isinstance(fil, py.builtin._basestring) is False. Does FNMatcher support byte strings? If yes should be easy to add byte strings.

[montefra]

You asked about non ascii character. Per se would be easy to add it somehow. I would need a new test, using tmpdir and ``tmpdir.ensure("some-string-with-non-ascii)" and then run visit('non-ascii').

Would be such a test be picked up automatically? I find the logic tests complicated (e.g. my new the tests in testing/path/test_local.py are clearly run, but the module itself is not present in the list from pytest and the three tests appears in three different positions, as the tests of 1d95912 appears to show)

[RonnyPfannschmidt]

@montefra yes, the structure used there is quite old, back then py.test had no fixtures and was part of pylib

[RonnyPfannschmidt]

@montefra such a test would be picked up

[montefra]

@RonnyPfannschmidt : I'll work on it later today, if I manage

[RonnyPfannschmidt]

@montefra as your change looks good, i'd like to merge, any objections?

[montefra]

If you don't need tests for non-ascii go ahead.