Add python-2.6 support #74

jlaska · 2015-02-23T19:33:24Z

Support python-2.6

Adds a tox.ini to facilitate testing supported py_versions
Add a requirements-optional-26.txt file
Update fauxfactory/init.py to support 2.6 syntax
updates tests/test_*.py to conditionally import unittest2
Add 2.6 to .travis.yml

coveralls · 2015-02-23T19:38:56Z

Changes Unknown when pulling 9c76729 on jlaska:python26-support into * on omaciel:master*.

coveralls · 2015-02-23T19:38:56Z

Changes Unknown when pulling 9c76729 on jlaska:python26-support into * on omaciel:master*.

coveralls · 2015-02-23T19:49:49Z

Changes Unknown when pulling 98cf219 on jlaska:python26-support into * on omaciel:master*.

jlaska · 2015-02-23T20:19:06Z

Note, the travis failure is caused by a bug in pylint. I can update requirements-optional.txt to use the previous release of pylint.

Ichimonji10 · 2015-02-23T20:33:22Z

fauxfactory/__init__.py

@@ -57,7 +57,7 @@ def _make_unicode(data):
        cast is necessary, a copy of ``data`` is returned.

    """
-    if sys.version_info.major == 2:
+    if sys.version_info[0] == 2:


Why is this change being made?

I do not have a Python 2.6 system available for manual testing. But the documentation indicates that major, minor, micro, releaselevel, and serial are available. See here.

@Ichimonji10 actually it looks like sys.version_info returns a tuple "containing the five components of the version number: major, minor, micro, releaselevel, and serial" for both 2.6.x and 2.7.x but being able to access these values by name (i.e. sys.version_info.major) can only be done with 2.7.x. If this is indeed the case the change above has merit.

Answering my own question:

Python 2.6.6 (r266:84292, Sep 12 2011, 14:03:14) [GCC 4.4.5 20110214 (Red Hat 4.4.5-6)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> print sys.version_info (2, 6, 6, 'final', 0) >>> print sys.version_info.major Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: 'tuple' object has no attribute 'major' >

OK. Thanks for the clarification.

Ichimonji10 · 2015-02-23T20:42:10Z

I found it easier to migrate test collection/execution to pytest rather than conditionally using unitest2 for older python releases.

Conditionally importing unittest should be as easy as this:

from sys import version_info
if version_info.major == 2:
    import unittest2 as unittest
else:
    import unittest

That would eliminate the dependency on tox, pytest and pytest-cov. Also, should tox be listed as a requirement somewhere?

jlaska · 2015-02-23T20:49:53Z

@Ichimonji10 That's the route I went originally, just disliked having the conditional blog at the top of each module. I can go through route if desired.

Also, I think we only want unittest2 when using <= 2.6 right?

Ichimonji10 · 2015-02-23T22:03:16Z

Also, I think we only want unittest2 when using <= 2.6 right?

Yes. My example code was bad. Something like this is better:

from sys import version_info
if version_info[0:2] == (2, 6):
    import unittest2 as unittest
else:
    import unittest

jlaska · 2015-02-23T22:14:45Z

I can make the conditional import work, but I'm having difficulty getting 'Python -m unittest discover...' to work within tox on py26.

Ichimonji10 · 2015-02-23T22:21:13Z

@Ichimonji10 That's the route I went originally, just disliked having the conditional blog at the top of each module. I can go through route if desired.

Fair enough.

My personal preference is that we stick with tools from the standard library unless there's a reason to branch out. In this case, using pytest means that we don't have to perform conditional imports. Great! But the downside of using pytest is that we're forcing our readers to learn the ins and outs of yet another library, instead of allowing our readers to use their knowledge of tools from the standard lib.

I will also admit to a strong personal distaste for the "no API" design goal of pytest. One of the stand-out features of Python over other languages is that you can explicitly list what is being imported right at the top of the module. This makes Python much more readable than many other languages. pytest throws away this advantage.

I've never used tox, and if it provides some compelling advantage that cannot be acheived with standard lib tools, then that's worth taking in to consideration. Can you provide a stack trace or similar?

omaciel · 2015-02-23T22:23:55Z

@Ichimonji10 My personal preference is that we stick with tools from the standard library unless there's a reason to branch out.

Agreed. I want to make sure that we keep external (to the standard library) dependencies to a minimal.

Now, I wish I had a 2.6 system laying around... :/

jlaska · 2015-02-24T01:30:48Z

Fair enough. PR updated for unittest2 support.

coveralls · 2015-02-24T01:33:29Z

Changes Unknown when pulling 074981e on jlaska:python26-support into * on omaciel:master*.

omaciel · 2015-02-24T03:10:22Z

fauxfactory/__init__.py

@@ -532,7 +532,7 @@ def gen_ipaddr(ip3=False, ipv6=False):

    if ipv6:
        # StackOverflow.com questions: generate-random-ipv6-address
-        ipaddr = u':'.join('{:x}'.format(
+        ipaddr = u':'.join('{0:x}'.format(


omaciel · 2015-02-24T03:22:11Z

@jlaska other than my still pending doubt about being able to call sys.version_info.major from a 2.6 environment, and a question about the additional requirements file, everything looks good but we still need to handle the this failure. Do we need to update .travis.yaml to take into consideration how lintmay behave under 2.6???

omaciel · 2015-02-24T03:28:39Z

Ahhh I missed the comment about the bug with pylint and python 2.6. Guess one way to work around it would be to follow the same .travis.yaml and tox configurations shown here?

jlaska · 2015-02-24T12:04:45Z

Ahhh I missed the comment about the bug with pylint and python 2.6. Guess one way to work around it would be to follow the same .travis.yaml and tox configurations shown here?

Yeah. My preferred approach is to have Travis use tox, so it mimics how one would test locally. I'll have that up for review shortly.

Ichimonji10 · 2015-02-24T14:25:00Z

tests/test_html.py


 class TestHTML(unittest.TestCase):
    """Test HTML generator."""
+    # support for py2.6
+    # matcher = re.compile('^<.*?>(.*?)</.*>$')


I don't understand this comment. Can it be clarified or dropped?

jlaska · 2015-02-24T14:30:55Z

@Ichimonji10 ah, good catch. That should not have landed in the PR. Removed and updated PR.

Ichimonji10 · 2015-02-24T15:13:44Z

.travis.yml

-  - 3.3
-  - 3.4
+env:
+  - TOXENV=py26


Can you add a comment above the list of environments noting that this is a duplicate of what is in tox.ini? Something like this:

env: # Should be identical to envlist in tox.ini. See: # http://tox.readthedocs.org/en/latest/config.html#confval-envlist=CSV - TOXENV=py26 - TOXENV=py27 - TOXENV=py33 - TOXENV=py34

@Ichimonji10 Sure! It's easier to just have travis just run tox. However, if you want travis to run a job for each supported python version, we'll need to duplicate the TOXENV in .travis.yml. I presume that's how you'll want the results, since it's consistent with how tests were previously run.

I'll update the .travis.yml per your feedback.

I like being able to view a matrix of results in Travis. There are times when a bug will only manifest itself in certain Python versions, and making that information easily discoverable is immensely valuable. That's why I'm OK with duplicating this information. If there is some method by which we can de-duplicate this information and still get a matrix of easily discoverable results, then that would be even cooler. But I don't know how to do that. I just don't know enough about the interactions between Travis and tox.

Ichimonji10 · 2015-02-24T15:14:12Z

ACK pending comment and commits being squashed.

* Adds a tox.ini to facilitate testing supported py_versions * Add a requirements-optional-26.txt file * Update fauxfactory/__init__.py to support 2.6 syntax * Updates tests/test_*.py to conditionally import unittest2 * Update travis to run tests using tox

jlaska · 2015-02-24T16:38:48Z

Commits squashed and incorporated feedback from @Ichimonji10

omaciel · 2015-02-24T17:01:15Z

ACK and thank you!!!

Add python-2.6 support

omaciel · 2015-02-24T17:03:55Z

Will release new version later today

jlaska · 2015-02-24T17:05:03Z

Yay, thanks @omaciel and @Ichimonji10

Ichimonji10 · 2015-02-24T20:12:14Z

Thank you, @jlaska. This is a definite improvement.

jlaska mentioned this pull request Feb 23, 2015

python-2.6 support #73

Closed

Ichimonji10 reviewed Feb 23, 2015
View reviewed changes

jlaska force-pushed the python26-support branch from 98cf219 to 074981e Compare February 24, 2015 01:30

omaciel reviewed Feb 24, 2015
View reviewed changes

jlaska force-pushed the python26-support branch 2 times, most recently from c76f44d to 92edf38 Compare February 24, 2015 13:46

Ichimonji10 reviewed Feb 24, 2015
View reviewed changes

jlaska force-pushed the python26-support branch from 92edf38 to 0b0dcb3 Compare February 24, 2015 14:29

Ichimonji10 reviewed Feb 24, 2015
View reviewed changes

Support python-2.6

36787d7

* Adds a tox.ini to facilitate testing supported py_versions * Add a requirements-optional-26.txt file * Update fauxfactory/__init__.py to support 2.6 syntax * Updates tests/test_*.py to conditionally import unittest2 * Update travis to run tests using tox

jlaska force-pushed the python26-support branch from 0b0dcb3 to 36787d7 Compare February 24, 2015 16:38

omaciel added a commit that referenced this pull request Feb 24, 2015

Merge pull request #74 from jlaska/python26-support

baa8c30

Add python-2.6 support

omaciel merged commit baa8c30 into omaciel:master Feb 24, 2015

jlaska deleted the python26-support branch February 24, 2015 17:04

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add python-2.6 support #74

Add python-2.6 support #74

jlaska commented Feb 23, 2015

coveralls commented Feb 23, 2015

coveralls commented Feb 23, 2015

coveralls commented Feb 23, 2015

jlaska commented Feb 23, 2015

Ichimonji10 Feb 23, 2015

Ichimonji10 Feb 23, 2015

omaciel Feb 24, 2015

omaciel Feb 24, 2015

Ichimonji10 Feb 24, 2015

Ichimonji10 commented Feb 23, 2015

jlaska commented Feb 23, 2015

Ichimonji10 commented Feb 23, 2015

jlaska commented Feb 23, 2015 via email

Ichimonji10 commented Feb 23, 2015

omaciel commented Feb 23, 2015

jlaska commented Feb 24, 2015

coveralls commented Feb 24, 2015

omaciel Feb 24, 2015

omaciel commented Feb 24, 2015

omaciel commented Feb 24, 2015

jlaska commented Feb 24, 2015

Ichimonji10 Feb 24, 2015

jlaska commented Feb 24, 2015

Ichimonji10 Feb 24, 2015

jlaska Feb 24, 2015

Ichimonji10 Feb 24, 2015

Ichimonji10 commented Feb 24, 2015

jlaska commented Feb 24, 2015

omaciel commented Feb 24, 2015

omaciel commented Feb 24, 2015

jlaska commented Feb 24, 2015

Ichimonji10 commented Feb 24, 2015

Add python-2.6 support #74

Add python-2.6 support #74

Conversation

jlaska commented Feb 23, 2015

coveralls commented Feb 23, 2015

coveralls commented Feb 23, 2015

coveralls commented Feb 23, 2015

jlaska commented Feb 23, 2015

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Ichimonji10 commented Feb 23, 2015

jlaska commented Feb 23, 2015

Ichimonji10 commented Feb 23, 2015

jlaska commented Feb 23, 2015 via email

Ichimonji10 commented Feb 23, 2015

omaciel commented Feb 23, 2015

jlaska commented Feb 24, 2015

coveralls commented Feb 24, 2015

Choose a reason for hiding this comment

omaciel commented Feb 24, 2015

omaciel commented Feb 24, 2015

jlaska commented Feb 24, 2015

Choose a reason for hiding this comment

jlaska commented Feb 24, 2015

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Ichimonji10 commented Feb 24, 2015

jlaska commented Feb 24, 2015

omaciel commented Feb 24, 2015

omaciel commented Feb 24, 2015

jlaska commented Feb 24, 2015

Ichimonji10 commented Feb 24, 2015