Running through some of my tests with the `-3` flag in python2.7 I encountered some errors within py.test itself. This fixes those errors so we can use py.test in order to identify problems with Python 3.
Part two of https://github.com/pytest-dev/pytest/issues/1512. Update the format
of the truncation message to help make it clear that pytest truncates the
entire assertion output when verbosity < 2.
Part one of https://github.com/pytest-dev/pytest/issues/1512.
If verbosity=1, assertion explanations are truncated at 10 lines. In this
situation, it's more important to tell the user which dictionary items are
different than which are the same.
We used to have this when we where patching the real Python
AssertionError for use with reinterpret, but reinterpret is now
gone so we no longer need this as it is not used by rewrite.
The assertion reinterpretation is an old backwards compatibility
mode which was no longer being maintained on feature-parity with
the assertion rewriting mode. It was also responsible for some
dubious patching of builtins and test with side-effects would
suddenly start passing. Since re-writing has been the default for
a long time and plugins are now also re-written it is time to
retire reinterpretation.
Hook up the PEP 302 import hook very early in pytest startup so
that it gets installed before setuptools-installed plugins are
imported. Also iterate over all installed plugins and mark them
for rewriting. If an installed plugin is already imported then
a warning is issued, we can not break since that might break
existing plugins and the fallback will still be gracefull to
plain asserts.
Some existing tests are failing in this commit because of the new
warning triggered by inline pytest runs due to the hypothesis
plugin already being imported. The tests will be fixed in the next
commit.
When you don't get enough information with a test running on a CI, it's quite
frustrating, for various reasons:
- It's more likely to be a flaky test, so you might not be able to reproduce
the failure.
- Passing -vv is quite bothersome (creating a temporary commit and reverting
it)
For those reasons, if something goes wrong on CI, it's good to have as much
information as possible.
Sometimes the repr of an object can contain the "\n{" sequence which is
used as a formatting language, so they are escaped to "\\n{". But the
collapse-false code needs to look for the real "\n{" token instead of
simply "{" as otherwise it may get unbalanced braces from the object's
repr (sometimes caused by the collapsing of long reprs by saferepr).
Fixes issue #731.
--HG--
branch : pytest-2.7
for the already existing cleanup logic of the config object.
This simplifies lifecycle management as we don't keep two
layers of shutdown functions and also simplifies the pluginmanager
interface.
also add some docstrings.
--HG--
branch : plugin_no_pytest
From the python-dev thread it seemed like using
object.__getattribute__(self, 'name') is the cleanest way of
implementing a class wich uses .__getattr__() and should be
pickelable. That only works on new-style classes so this also turns
HookProxy into a new-style class on py2.
This also re-writes the test to not use cPickle so it runs on py3k.
User provided messages, or any valid expression given as second
argument to the assert statement, are now shown in addition to the
py.test introspection details. Formerly any user provided message
would entirely replace the introspection details.
Fixes issue549.
The assertion formatting mini-language depends on newlines being
escaped. Unfortunately if the repr of an object contained
newlines the rewriting module did not escape those, which is now
fixed.
Fixes issue453.
The result from the pytest_assertrepr_compare hook should not include
any newlines since that will confuse the mini-formatting language used
by assertion.util.format_explanation. So simply escape the included
newlines, this way hook writers do not have to worry about this at
all.
Fixes issue 453.
Such reference cycles unnecessarily cause Python interpreter not to garbage
collect the objects referenced in those cycles as soon they could be collected,
and in turn cause the tests to use more memory than is strictly necessary.
--HG--
branch : break_ExceptionInfo_reference_cycles
Prevent error on exit if some code messes with sys.meta_path and removes the
assertionrewrite hook (CaptureMock seems to do this):
File "/Users/marca/dev/hg-repos/pytest/_pytest/assertion/__init__.py", line 64, in pytest_unconfigure
sys.meta_path.remove(hook)
ValueError: list.remove(x): x not in list
--HG--
branch : sys_meta_path_remove_hook_only_if_present
If the compared text was in bytes and not actually valid text
(i.e. could not be encoded to text/unicode using the default encoding)
then the assertrepr would fail with an EncodingError. This ensures
that the internal string is always valid unicode, converting any bytes
safely to valid unicode. This is done using repr() which then needs
post-processing to fix the encompassing quotes and un-escape newlines.
This fixes issue 429.
The only remaining 'py.test' references are:
* those referring to the 'py.test' executable
* those in code explicitly testing py.test/pytest module compatibility
* those in old CHANGES documentation
* those in documentation generated based on external data
* those in seemingly unfinished & unmaintained Japanese documentation
Minor stylistic changes and typo corrections made to documentation next to
several applied py.test --> pytest content changes.
Starting with Python 3.3, NamespacePath passed to importlib hooks
seem to have lost the ability to be accessed by index.
We wrap the index access in a try..except and wrap the path in a
list if it happens.
Fixes#383.