root/django/branches/0.95-bugfixes/tests/doctest.py

# Module doctest.
# Released to the public domain 16-Jan-2001, by Tim Peters (tim@python.org).
# Major enhancements and refactoring by:
#     Jim Fulton
#     Edward Loper

# Provided as-is; use at your own risk; no warranty; no promises; enjoy!

r"""Module doctest -- a framework for running examples in docstrings.

In simplest use, end each module M to be tested with:

def _test():
    import doctest
    doctest.testmod()

if __name__ == "__main__":
    _test()

Then running the module as a script will cause the examples in the
docstrings to get executed and verified:

python M.py

This won't display anything unless an example fails, in which case the
failing example(s) and the cause(s) of the failure(s) are printed to stdout
(why not stderr? because stderr is a lame hack <0.2 wink>), and the final
line of output is "Test failed.".

Run it with the -v switch instead:

python M.py -v

and a detailed report of all examples tried is printed to stdout, along
with assorted summaries at the end.

You can force verbose mode by passing "verbose=True" to testmod, or prohibit
it by passing "verbose=False".  In either of those cases, sys.argv is not
examined by testmod.

There are a variety of other ways to run doctests, including integration
with the unittest framework, and support for running non-Python text
files containing doctests.  There are also many ways to override parts
of doctest's default behaviors.  See the Library Reference Manual for
details.
"""

__docformat__ = 'reStructuredText en'

__all__ = [
    # 0, Option Flags
    'register_optionflag',
    'DONT_ACCEPT_TRUE_FOR_1',
    'DONT_ACCEPT_BLANKLINE',
    'NORMALIZE_WHITESPACE',
    'ELLIPSIS',
    'IGNORE_EXCEPTION_DETAIL',
    'COMPARISON_FLAGS',
    'REPORT_UDIFF',
    'REPORT_CDIFF',
    'REPORT_NDIFF',
    'REPORT_ONLY_FIRST_FAILURE',
    'REPORTING_FLAGS',
    # 1. Utility Functions
    'is_private',
    # 2. Example & DocTest
    'Example',
    'DocTest',
    # 3. Doctest Parser
    'DocTestParser',
    # 4. Doctest Finder
    'DocTestFinder',
    # 5. Doctest Runner
    'DocTestRunner',
    'OutputChecker',
    'DocTestFailure',
    'UnexpectedException',
    'DebugRunner',
    # 6. Test Functions
    'testmod',
    'testfile',
    'run_docstring_examples',
    # 7. Tester
    'Tester',
    # 8. Unittest Support
    'DocTestSuite',
    'DocFileSuite',
    'set_unittest_reportflags',
    # 9. Debugging Support
    'script_from_examples',
    'testsource',
    'debug_src',
    'debug',
]

import __future__

import sys, traceback, inspect, linecache, os, re, types
import unittest, difflib, pdb, tempfile
import warnings
from StringIO import StringIO

# Don't whine about the deprecated is_private function in this
# module's tests.
warnings.filterwarnings("ignore", "is_private", DeprecationWarning,
                        __name__, 0)

# There are 4 basic classes:
#  - Example: a <source, want> pair, plus an intra-docstring line number.
#  - DocTest: a collection of examples, parsed from a docstring, plus
#    info about where the docstring came from (name, filename, lineno).
#  - DocTestFinder: extracts DocTests from a given object's docstring and
#    its contained objects' docstrings.
#  - DocTestRunner: runs DocTest cases, and accumulates statistics.
#
# So the basic picture is:
#
#                             list of:
# +------+                   +---------+                   +-------+
# |object| --DocTestFinder-> | DocTest | --DocTestRunner-> |results|
# +------+                   +---------+                   +-------+
#                            | Example |
#                            |   ...   |
#                            | Example |
#                            +---------+

# Option constants.

OPTIONFLAGS_BY_NAME = {}
def register_optionflag(name):
    flag = 1 << len(OPTIONFLAGS_BY_NAME)
    OPTIONFLAGS_BY_NAME[name] = flag
    return flag

DONT_ACCEPT_TRUE_FOR_1 = register_optionflag('DONT_ACCEPT_TRUE_FOR_1')
DONT_ACCEPT_BLANKLINE = register_optionflag('DONT_ACCEPT_BLANKLINE')
NORMALIZE_WHITESPACE = register_optionflag('NORMALIZE_WHITESPACE')
ELLIPSIS = register_optionflag('ELLIPSIS')
IGNORE_EXCEPTION_DETAIL = register_optionflag('IGNORE_EXCEPTION_DETAIL')

COMPARISON_FLAGS = (DONT_ACCEPT_TRUE_FOR_1 |
                    DONT_ACCEPT_BLANKLINE |
                    NORMALIZE_WHITESPACE |
                    ELLIPSIS |
                    IGNORE_EXCEPTION_DETAIL)

REPORT_UDIFF = register_optionflag('REPORT_UDIFF')
REPORT_CDIFF = register_optionflag('REPORT_CDIFF')
REPORT_NDIFF = register_optionflag('REPORT_NDIFF')
REPORT_ONLY_FIRST_FAILURE = register_optionflag('REPORT_ONLY_FIRST_FAILURE')

REPORTING_FLAGS = (REPORT_UDIFF |
                   REPORT_CDIFF |
                   REPORT_NDIFF |
                   REPORT_ONLY_FIRST_FAILURE)

# Special string markers for use in `want` strings:
BLANKLINE_MARKER = '<BLANKLINE>'
ELLIPSIS_MARKER = '...'

######################################################################
## Table of Contents
######################################################################
#  1. Utility Functions
#  2. Example & DocTest -- store test cases
#  3. DocTest Parser -- extracts examples from strings
#  4. DocTest Finder -- extracts test cases from objects
#  5. DocTest Runner -- runs test cases
#  6. Test Functions -- convenient wrappers for testing
#  7. Tester Class -- for backwards compatibility
#  8. Unittest Support
#  9. Debugging Support
# 10. Example Usage

######################################################################
## 1. Utility Functions
######################################################################

def is_private(prefix, base):
    """prefix, base -> true iff name prefix + "." + base is "private".

    Prefix may be an empty string, and base does not contain a period.
    Prefix is ignored (although functions you write conforming to this
    protocol may make use of it).
    Return true iff base begins with an (at least one) underscore, but
    does not both begin and end with (at least) two underscores.

    >>> is_private("a.b", "my_func")
    False
    >>> is_private("____", "_my_func")
    True
    >>> is_private("someclass", "__init__")
    False
    >>> is_private("sometypo", "__init_")
    True
    >>> is_private("x.y.z", "_")
    True
    >>> is_private("_x.y.z", "__")
    False
    >>> is_private("", "")  # senseless but consistent
    False
    """
    warnings.warn("is_private is deprecated; it wasn't useful; "
                  "examine DocTestFinder.find() lists instead",
                  DeprecationWarning, stacklevel=2)
    return base[:1] == "_" and not base[:2] == "__" == base[-2:]

def _extract_future_flags(globs):
    """
    Return the compiler-flags associated with the future features that
    have been imported into the given namespace (globs).
    """
    flags = 0
    for fname in __future__.all_feature_names:
        feature = globs.get(fname, None)
        if feature is getattr(__future__, fname):
            flags |= feature.compiler_flag
    return flags

def _normalize_module(module, depth=2):
    """
    Return the module specified by `module`.  In particular:
      - If `module` is a module, then return module.
      - If `module` is a string, then import and return the
        module with that name.
      - If `module` is None, then return the calling module.
        The calling module is assumed to be the module of
        the stack frame at the given depth in the call stack.
    """
    if inspect.ismodule(module):
        return module
    elif isinstance(module, (str, unicode)):
        return __import__(module, globals(), locals(), ["*"])
    elif module is None:
        return sys.modules[sys._getframe(depth).f_globals['__name__']]
    else:
        raise TypeError("Expected a module, string, or None")

def _indent(s, indent=4):
    """
    Add the given number of space characters to the beginning every
    non-blank line in `s`, and return the result.
    """
    # This regexp matches the start of non-blank lines:
    return re.sub('(?m)^(?!$)', indent*' ', s)

def _exception_traceback(exc_info):
    """
    Return a string containing a traceback message for the given
    exc_info tuple (as returned by sys.exc_info()).
    """
    # Get a traceback message.
    excout = StringIO()
    exc_type, exc_val, exc_tb = exc_info
    traceback.print_exception(exc_type, exc_val, exc_tb, file=excout)
    return excout.getvalue()

# Override some StringIO methods.
class _SpoofOut(StringIO):
    def getvalue(self):
        result = StringIO.getvalue(self)
        # If anything at all was written, make sure there's a trailing
        # newline.  There's no way for the expected output to indicate
        # that a trailing newline is missing.
        if result and not result.endswith("\n"):
            result += "\n"
        # Prevent softspace from screwing up the next test case, in
        # case they used print with a trailing comma in an example.
        if hasattr(self, "softspace"):
            del self.softspace
        return result

    def truncate(self,   size=None):
        StringIO.truncate(self, size)
        if hasattr(self, "softspace"):
            del self.softspace

# Worst-case linear-time ellipsis matching.
def _ellipsis_match(want, got):
    """
    Essentially the only subtle case:
    >>> _ellipsis_match('aa...aa', 'aaa')
    False
    """
    if ELLIPSIS_MARKER not in want:
        return want == got

    # Find "the real" strings.
    ws = want.split(ELLIPSIS_MARKER)
    assert len(ws) >= 2

    # Deal with exact matches possibly needed at one or both ends.
    startpos, endpos = 0, len(got)
    w = ws[0]
    if w:   # starts with exact match
        if got.startswith(w):
            startpos = len(w)
            del ws[0]
        else:
            return False
    w = ws[-1]
    if w:   # ends with exact match
        if got.endswith(w):
            endpos -= len(w)
            del ws[-1]
        else:
            return False

    if startpos > endpos:
        # Exact end matches required more characters than we have, as in
        # _ellipsis_match('aa...aa', 'aaa')
        return False

    # For the rest, we only need to find the leftmost non-overlapping
    # match for each piece.  If there's no overall match that way alone,
    # there's no overall match period.
    for w in ws:
        # w may be '' at times, if there are consecutive ellipses, or
        # due to an ellipsis at the start or end of `want`.  That's OK.
        # Search for an empty string succeeds, and doesn't change startpos.
        startpos = got.find(w, startpos, endpos)
        if startpos < 0:
            return False
        startpos += len(w)

    return True

def _comment_line(line):
    "Return a commented form of the given line"
    line = line.rstrip()
    if line:
        return '# '+line
    else:
        return '#'

class _OutputRedirectingPdb(pdb.Pdb):
    """
    A specialized version of the python debugger that redirects stdout
    to a given stream when interacting with the user.  Stdout is *not*
    redirected when traced code is executed.
    """
    def __init__(self, out):
        self.__out = out
        pdb.Pdb.__init__(self)

    def trace_dispatch(self, *args):
        # Redirect stdout to the given stream.
        save_stdout = sys.stdout
        sys.stdout = self.__out
        # Call Pdb's trace dispatch method.
        try:
            return pdb.Pdb.trace_dispatch(self, *args)
        finally:
            sys.stdout = save_stdout

# [XX] Normalize with respect to os.path.pardir?
def _module_relative_path(module, path):
    if not inspect.ismodule(module):
        raise TypeError, 'Expected a module: %r' % module
    if path.startswith('/'):
        raise ValueError, 'Module-relative files may not have absolute paths'

    # Find the base directory for the path.
    if hasattr(module, '__file__'):
        # A normal module/package
        basedir = os.path.split(module.__file__)[0]
    elif module.__name__ == '__main__':
        # An interactive session.
        if len(sys.argv)>0 and sys.argv[0] != '':
            basedir = os.path.split(sys.argv[0])[0]
        else:
            basedir = os.curdir
    else:
        # A module w/o __file__ (this includes builtins)
        raise ValueError("Can't resolve paths relative to the module " +
                         module + " (it has no __file__)")

    # Combine the base directory and the path.
    return os.path.join(basedir, *(path.split('/')))

######################################################################
## 2. Example & DocTest
######################################################################
## - An "example" is a <source, want> pair, where "source" is a
##   fragment of source code, and "want" is the expected output for
##   "source."  The Example class also includes information about
##   where the example was extracted from.
##
## - A "doctest" is a collection of examples, typically extracted from
##   a string (such as an object's docstring).  The DocTest class also
##   includes information about where the string was extracted from.

class Example:
    """
    A single doctest example, consisting of source code and expected
    output.  `Example` defines the following attributes:

      - source: A single Python statement, always ending with a newline.
        The constructor adds a newline if needed.

      - want: The expected output from running the source code (either
        from stdout, or a traceback in case of exception).  `want` ends
        with a newline unless it's empty, in which case it's an empty
        string.  The constructor adds a newline if needed.

      - exc_msg: The exception message generated by the example, if
        the example is expected to generate an exception; or `None` if
        it is not expected to generate an exception.  This exception
        message is compared against the return value of
        `traceback.format_exception_only()`.  `exc_msg` ends with a
        newline unless it's `None`.  The constructor adds a newline
        if needed.

      - lineno: The line number within the DocTest string containing
        this Example where the Example begins.  This line number is
        zero-based, with respect to the beginning of the DocTest.

      - indent: The example's indentation in the DocTest string.
        I.e., the number of space characters that preceed the
        example's first prompt.

      - options: A dictionary mapping from option flags to True or
        False, which is used to override default options for this
        example.  Any option flags not contained in this dictionary
        are left at their default value (as specified by the
        DocTestRunner's optionflags).  By default, no options are set.
    """
    def __init__(self, source, want, exc_msg=None, lineno=0, indent=0,
                 options=None):
        # Normalize inputs.
        if not source.endswith('\n'):
            source += '\n'
        if want and not want.endswith('\n'):
            want += '\n'
        if exc_msg is not None and not exc_msg.endswith('\n'):
            exc_msg += '\n'
        # Store properties.
        self.source = source
        self.want = want
        self.lineno = lineno
        self.indent = indent
        if options is None: options = {}
        self.options = options
        self.exc_msg = exc_msg

class DocTest:
    """
    A collection of doctest examples that should be run in a single
    namespace.  Each `DocTest` defines the following attributes:

      - examples: the list of examples.

      - globs: The namespace (aka globals) that the examples should
        be run in.

      - name: A name identifying the DocTest (typically, the name of
        the object whose docstring this DocTest was extracted from).

      - filename: The name of the file that this DocTest was extracted
        from, or `None` if the filename is unknown.

      - lineno: The line number within filename where this DocTest
        begins, or `None` if the line number is unavailable.  This
        line number is zero-based, with respect to the beginning of
        the file.

      - docstring: The string that the examples were extracted from,
        or `None` if the string is unavailable.
    """
    def __init__(self, examples, globs, name, filename, lineno, docstring):
        """
        Create a new DocTest containing the given examples.  The
        DocTest's globals are initialized with a copy of `globs`.
        """
        assert not isinstance(examples, basestring), \
               "DocTest no longer accepts str; use DocTestParser instead"
        self.examples = examples
        self.docstring = docstring
        self.globs = globs.copy()
        self.name = name
        self.filename = filename
        self.lineno = lineno

    def __repr__(self):
        if len(self.examples) == 0:
            examples = 'no examples'
        elif len(self.examples) == 1:
            examples = '1 example'
        else:
            examples = '%d examples' % len(self.examples)
        return ('<DocTest %s from %s:%s (%s)>' %
                (self.name, self.filename, self.lineno, examples))


    # This lets us sort tests by name:
    def __cmp__(self, other):
        if not isinstance(other, DocTest):
            return -1
        return cmp((self.name, self.filename, self.lineno, id(self)),
                   (other.name, other.filename, other.lineno, id(other)))

######################################################################
## 3. DocTestParser
######################################################################

class DocTestParser:
    """
    A class used to parse strings containing doctest examples.
    """
    # This regular expression is used to find doctest examples in a
    # string.  It defines three groups: `source` is the source code
    # (including leading indentation and prompts); `indent` is the
    # indentation of the first (PS1) line of the source code; and
    # `want` is the expected output (including leading indentation).
    _EXAMPLE_RE = re.compile(r'''
        # Source consists of a PS1 line followed by zero or more PS2 lines.
        (?P<source>
            (?:^(?P<indent> [ ]*) >>>    .*)    # PS1 line
            (?:\n           [ ]*  \.\.\. .*)*)  # PS2 lines
        \n?
        # Want consists of any non-blank lines that do not start with PS1.
        (?P<want> (?:(?![ ]*$)    # Not a blank line
                     (?![ ]*>>>)  # Not a line starting with PS1
                     .*$\n?       # But any other line
                  )*)
        ''', re.MULTILINE | re.VERBOSE)

    # A regular expression for handling `want` strings that contain
    # expected exceptions.  It divides `want` into three pieces:
    #    - the traceback header line (`hdr`)
    #    - the traceback stack (`stack`)
    #    - the exception message (`msg`), as generated by
    #      traceback.format_exception_only()
    # `msg` may have multiple lines.  We assume/require that the
    # exception message is the first non-indented line starting with a word
    # character following the traceback header line.
    _EXCEPTION_RE = re.compile(r"""
        # Grab the traceback header.  Different versions of Python have
        # said different things on the first traceback line.
        ^(?P<hdr> Traceback\ \(
            (?: most\ recent\ call\ last
            |   innermost\ last
            ) \) :
        )
        \s* $                # toss trailing whitespace on the header.
        (?P<stack> .*?)      # don't blink: absorb stuff until...
        ^ (?P<msg> \w+ .*)   #     a line *starts* with alphanum.
        """, re.VERBOSE | re.MULTILINE | re.DOTALL)

    # A callable returning a true value iff its argument is a blank line
    # or contains a single comment.
    _IS_BLANK_OR_COMMENT = re.compile(r'^[ ]*(#.*)?$').match

    def parse(self, string, name='<string>'):
        """
        Divide the given string into examples and intervening text,
        and return them as a list of alternating Examples and strings.
        Line numbers for the Examples are 0-based.  The optional
        argument `name` is a name identifying this string, and is only
        used for error messages.
        """
        string = string.expandtabs()
        # If all lines begin with the same indentation, then strip it.
        min_indent = self._min_indent(string)
        if min_indent > 0:
            string = '\n'.join([l[min_indent:] for l in string.split('\n')])

        output = []
        charno, lineno = 0, 0
        # Find all doctest examples in the string:
        for m in self._EXAMPLE_RE.finditer(string):
            # Add the pre-example text to `output`.
            output.append(string[charno:m.start()])
            # Update lineno (lines before this example)
            lineno += string.count('\n', charno, m.start())
            # Extract info from the regexp match.
            (source, options, want, exc_msg) = \
                     self._parse_example(m, name, lineno)
            # Create an Example, and add it to the list.
            if not self._IS_BLANK_OR_COMMENT(source):
                output.append( Example(source, want, exc_msg,
                                    lineno=lineno,
                                    indent=min_indent+len(m.group('indent')),
                                    options=options) )
            # Update lineno (lines inside this example)
            lineno += string.count('\n', m.start(), m.end())
            # Update charno.
            charno = m.end()
        # Add any remaining post-example text to `output`.
        output.append(string[charno:])
        return output

    def get_doctest(self, string, globs, name, filename, lineno):
        """
        Extract all doctest examples from the given string, and
        collect them into a `DocTest` object.

        `globs`, `name`, `filename`, and `lineno` are attributes for
        the new `DocTest` object.  See the documentation for `DocTest`
        for more information.
        """
        return DocTest(self.get_examples(string, name), globs,
                       name, filename, lineno, string)

    def get_examples(self, string, name='<string>'):
        """
        Extract all doctest examples from the given string, and return
        them as a list of `Example` objects.  Line numbers are
        0-based, because it's most common in doctests that nothing
        interesting appears on the same line as opening triple-quote,
        and so the first interesting line is called \"line 1\" then.

        The optional argument `name` is a name identifying this
        string, and is only used for error messages.
        """
        return [x for x in self.parse(string, name)
                if isinstance(x, Example)]

    def _parse_example(self, m, name, lineno):
        """
        Given a regular expression match from `_EXAMPLE_RE` (`m`),
        return a pair `(source, want)`, where `source` is the matched
        example's source code (with prompts and indentation stripped);
        and `want` is the example's expected output (with indentation
        stripped).

        `name` is the string's name, and `lineno` is the line number
        where the example starts; both are used for error messages.
        """
        # Get the example's indentation level.
        indent = len(m.group('indent'))

        # Divide source into lines; check that they're properly
        # indented; and then strip their indentation & prompts.
        source_lines = m.group('source').split('\n')
        self._check_prompt_blank(source_lines, indent, name, lineno)
        self._check_prefix(source_lines[1:], ' '*indent + '.', name, lineno)
        source = '\n'.join([sl[indent+4:] for sl in source_lines])

        # Divide want into lines; check that it's properly indented; and
        # then strip the indentation.  Spaces before the last newline should
        # be preserved, so plain rstrip() isn't good enough.
        want = m.group('want')
        want_lines = want.split('\n')
        if len(want_lines) > 1 and re.match(r' *$', want_lines[-1]):
            del want_lines[-1]  # forget final newline & spaces after it
        self._check_prefix(want_lines, ' '*