When an import error happens ``testtools.run`` will now show the full
error rather than just the name of the module that failed to import.
(Robert Collins)

Change-Id: I2d9d7a4659b51a5e3ea8000301ecb799ed30f463

.gitignore

@@ -16,3 +16,6 @@ doc/_build
 testtools.egg-info
 /build/
 /.env/
+/.eggs/
+AUTHORS
+ChangeLog

.travis.yml

@@ -6,20 +6,24 @@ python:
   - "3.3"
   - "3.4"
   - "pypy"
-  - "pypy3"
 
 # We have to pin Jinja2 < 2.7  for Python 3.2 because 2.7 drops/breaks support:
 # http://jinja.pocoo.org/docs/changelog/#version-2-7
+# And Spinx to < 1.3 for pypy3 and python 3.2 similarly.
 #
 # See also:
 # http://stackoverflow.com/questions/18252804/syntax-error-in-jinja-2-library
 matrix:
   include:
     - python: "3.2"
-      env: JINJA_REQ="jinja2<2.7, Pygments<2.0"
+      env:
+        - JINJA_REQ="jinja2<2.7, Pygments<2.0"
+        - SPHINX="<1.3"
+    - python: "pypy3"
+      env: SPHINX="<1.3"
 
 install:
-  - pip install fixtures $JINJA_REQ sphinx
+  - pip install fixtures $JINJA_REQ sphinx$SPHINX
   - python setup.py install
 
 script:

MANIFEST.in

@@ -4,7 +4,4 @@ include MANIFEST.in
 include NEWS
 include README.rst
 include .gitignore
-graft doc
-graft doc/_static
-graft doc/_templates
 prune doc/_build

NEWS

@@ -7,6 +7,78 @@ Changes and improvements to testtools_, grouped by release.
 NEXT
 ~~~~
 
+1.8.0
+~~~~~
+
+Improvements
+------------
+
+* AsynchronousDeferredRunTest now correctly attaches the test log.
+  Previously it attached an empty file. (Colin Watson)
+
+1.7.1
+~~~~~
+
+Improvements
+------------
+
+* Building a wheel on Python 3 was missing ``_compat2x.py`` needed for Python2.
+  This was a side effect of the fix to bug #941958, where we fixed a cosmetic
+  error. (Robert Collins, #1430534)
+
+* During reporting in ``TextTestResult`` now always uses ``ceil`` rather than
+  depending on the undefined rounding behaviour in string formatting.
+  (Robert Collins)
+
+1.7.0
+~~~~~
+
+Improvements
+------------
+
+* Empty attachments to tests were triggering a file payload of None in the
+  ``ExtendedToStreamDecorator`` code, which caused multiple copies of
+  attachments that had been output prior to the empty one.
+  (Robert Collins, #1378609)
+
+1.6.1
+~~~~~
+
+Changes
+-------
+
+* Fix installing when ``extras`` is not already installed. Our guards
+  for the absence of unittest2 were not sufficient.
+  (Robert Collins, #1430076)
+
+1.6.0
+~~~~~
+
+Improvements
+------------
+
+* ``testtools.run`` now accepts ``--locals`` to show local variables
+  in tracebacks, which can be a significant aid in debugging. In doing
+  so we've removed the code reimplementing linecache and traceback by
+  using the new traceback2 and linecache2 packages.
+  (Robert Collins, github #111)
+
+Changes
+-------
+
+* ``testtools`` now depends on ``unittest2`` 1.0.0 which brings in a dependency
+  on ``traceback2`` and via it ``linecache2``. (Robert Collins)
+
+1.5.0
+~~~~~
+
+Improvements
+------------
+
+* When an import error happens ``testtools.run`` will now show the full
+  error rather than just the name of the module that failed to import.
+  (Robert Collins)
+
 1.4.0
 ~~~~~
 

doc/for-framework-folk.rst

@@ -92,6 +92,13 @@ cause ``testtools.RunTest`` to fail the test case after the test has finished.
 This is useful when you want to cause a test to fail, but don't want to
 prevent the remainder of the test code from being executed.
 
+Exception formatting
+--------------------
+
+Testtools ``TestCase`` instances format their own exceptions. The attribute
+``__testtools_tb_locals__`` controls whether to include local variables in the
+formatted exceptions.
+
 Test placeholders
 =================
 

doc/for-test-authors.rst

@@ -72,10 +72,12 @@ installed or have Python 2.7 or later, and then run::
 
   $ python -m testtools.run discover packagecontainingtests
 
-For more information see the Python 2.7 unittest documentation, or::
+For more information see the Python unittest documentation, and::
 
     python -m testtools.run --help
 
+which describes the options available to ``testtools.run``.
+
 As your testing needs grow and evolve, you will probably want to use a more
 sophisticated test runner.  There are many of these for Python, and almost all
 of them will happily run testtools tests.  In particular:

doc/hacking.rst

@@ -169,20 +169,16 @@ Tasks
 +++++
 
 #. Choose a version number, say X.Y.Z
-#. In trunk, ensure __init__ has version ``(X, Y, Z, 'final', 0)``
 #. Under NEXT in NEWS add a heading with the version number X.Y.Z.
 #. Possibly write a blurb into NEWS.
 #. Commit the changes.
-#. Tag the release, ``git tag -s testtools-X.Y.Z``
+#. Tag the release, ``git tag -s X.Y.Z -m "Releasing X.Y.Z"``
 #. Run 'make release', this:
    #. Creates a source distribution and uploads to PyPI
    #. Ensures all Fix Committed bugs are in the release milestone
    #. Makes a release on Launchpad and uploads the tarball
    #. Marks all the Fix Committed bugs as Fix Released
    #. Creates a new milestone
-#. Change __version__ in __init__.py to the probable next version.
-   e.g. to ``(X, Y, Z+1, 'dev', 0)``.
-#. Commit 'Opening X.Y.Z+1 for development.'
 #. If a new series has been created (e.g. 0.10.0), make the series on Launchpad.
 #. Push trunk to Github, ``git push --tags origin master``
 

requirements.txt

@@ -0,0 +1,7 @@
+pbr>=0.11
+extras
+# 'mimeparse' has not been uploaded by the maintainer with Python3 compat
+# but someone kindly uploaded a fixed version as 'python-mimeparse'.
+python-mimeparse
+unittest2>=1.0.0
+traceback2

setup.cfg

@@ -1,7 +1,16 @@
-[test]
-test_module = testtools.tests
-buffer=1
-catch=1
+[metadata]
+name = testtools
+summary = Extensions to the Python standard library unit testing framework
+home-page = https://github.com/testing-cabal/testtools
+description-file = doc/overview.rst
+author = Jonathan M. Lange
+author-email = jml+testtools@mumak.net
+classifier = 
+    License :: OSI Approved :: MIT License
+    Programming Language :: Python :: 3
+
+[files]
+packages = testtools
 
 [bdist_wheel]
 universal = 1

setup.py

@@ -1,92 +1,16 @@
 #!/usr/bin/env python
-"""Distutils installer for testtools."""
+import setuptools
 
-from setuptools import setup
-from distutils.command.build_py import build_py
-import email
-import os
-import sys
+try:
+    import testtools
+    cmd_class = {}
+    if getattr(testtools, 'TestCommand', None) is not None:
+        cmd_class['test'] = testtools.TestCommand
+except:
+    cmd_class = None
 
-import testtools
-cmd_class = {}
-if getattr(testtools, 'TestCommand', None) is not None:
-    cmd_class['test'] = testtools.TestCommand
 
-
-class testtools_build_py(build_py):
-    def build_module(self, module, module_file, package):
-        if sys.version_info >= (3,) and module == '_compat2x':
-            return
-        return build_py.build_module(self, module, module_file, package)
-cmd_class['build_py'] = testtools_build_py
-
-
-def get_version_from_pkg_info():
-    """Get the version from PKG-INFO file if we can."""
-    pkg_info_path = os.path.join(os.path.dirname(__file__), 'PKG-INFO')
-    try:
-        pkg_info_file = open(pkg_info_path, 'r')
-    except (IOError, OSError):
-        return None
-    try:
-        pkg_info = email.message_from_file(pkg_info_file)
-    except email.MessageError:
-        return None
-    return pkg_info.get('Version', None)
-
-
-def get_version():
-    """Return the version of testtools that we are building."""
-    version = '.'.join(
-        str(component) for component in testtools.__version__[0:3])
-    phase = testtools.__version__[3]
-    if phase == 'final':
-        return version
-    pkg_info_version = get_version_from_pkg_info()
-    if pkg_info_version:
-        return pkg_info_version
-    # Apparently if we just say "snapshot" then distribute won't accept it
-    # as satisfying versioned dependencies. This is a problem for the
-    # daily build version.
-    return "%s.0dev0" % (version,)
-
-
-def get_long_description():
-    manual_path = os.path.join(
-        os.path.dirname(__file__), 'doc/overview.rst')
-    return open(manual_path).read()
-
-# Since we import testtools in setup.py, our setup requirements are our install
-# requirements.
-deps = [
-    'extras',
-    # 'mimeparse' has not been uploaded by the maintainer with Python3 compat
-    # but someone kindly uploaded a fixed version as 'python-mimeparse'.
-    'python-mimeparse',
-    'unittest2>=0.8.0',
-    ]
-
-
-setup(name='testtools',
-      author='Jonathan M. Lange',
-      author_email='jml+testtools@mumak.net',
-      url='https://github.com/testing-cabal/testtools',
-      description=('Extensions to the Python standard library unit testing '
-                   'framework'),
-      long_description=get_long_description(),
-      version=get_version(),
-      classifiers=["License :: OSI Approved :: MIT License",
-        "Programming Language :: Python :: 3",
-        ],
-      packages=[
-        'testtools',
-        'testtools.matchers',
-        'testtools.testresult',
-        'testtools.tests',
-        'testtools.tests.matchers',
-        ],
-      cmdclass=cmd_class,
-      zip_safe=False,
-      install_requires=deps,
-      setup_requires=deps,
-      )
+setuptools.setup(
+    cmdclass=cmd_class,
+    setup_requires=['pbr'],
+    pbr=True)

testtools/__init__.py

@@ -121,5 +121,7 @@
 # established at this point, and setup.py will use a version of next-$(revno).
 # If the releaselevel is 'final', then the tarball will be major.minor.micro.
 # Otherwise it is major.minor.micro~$(revno).
-
-__version__ = (1, 4, 0, 'final', 0)
+from pbr.version import VersionInfo
+_version = VersionInfo('testtools')
+__version__ = _version.semantic_version().version_tuple()
+version = _version.release_string()