comparison mercurial/match.py @ 42098:5753e5949b51

py3: add b'' prefixes to new doctests in match.py

# skip-blame as just b'' prefixes

Differential Revision: https://phab.mercurial-scm.org/D6221

42097:f7fdb24d4178

    'subinclude:<path>' - a file of patterns to match against files under
                          the same directory
    '<something>' - a pattern of the specified default type

    Usually a patternmatcher is returned:
    >>> match('foo', '.', ['re:.*\.c$', 'path:foo/a', '*.py'])
    <patternmatcher patterns='.*\\.c$|foo/a(?:/|$)|[^/]*\\.py$'>

    Combining 'patterns' with 'include' (resp. 'exclude') gives an
    intersectionmatcher (resp. a differencematcher):
    >>> type(match('foo', '.', ['re:.*\.c$'], include=['path:lib']))
    <class 'mercurial.match.intersectionmatcher'>
    >>> type(match('foo', '.', ['re:.*\.c$'], exclude=['path:build']))
    <class 'mercurial.match.differencematcher'>

    Notice that, if 'patterns' is empty, an alwaysmatcher is returned:
    >>> match('foo', '.', [])
    <alwaysmatcher>

    The 'default' argument determines which kind of pattern is assumed if a
    pattern has no prefix:
    >>> match('foo', '.', ['.*\.c$'], default='re')
    <patternmatcher patterns='.*\\.c$'>
    >>> match('foo', '.', ['main.py'], default='relpath')
    <patternmatcher patterns='main\\.py(?:/|$)'>
    >>> match('foo', '.', ['main.py'], default='re')
    <patternmatcher patterns='main.py'>

    The primary use of matchers is to check whether a value (usually a file
    name) matches againset one of the patterns given at initialization. There
    are two ways of doing this check.

    >>> m = match('foo', '', ['re:.*\.c$', 'relpath:a'])

    1. Calling the matcher with a file name returns True if any pattern
    matches that file name:
    >>> m('a')
    True
    >>> m('main.c')
    True
    >>> m('test.py')
    False

    2. Using the exact() method only returns True if the file name matches one
    of the exact patterns (i.e. not re: or glob: patterns):
    >>> m.exact('a')
    True
    >>> m.exact('main.c')
    False
    """
    normalize = _donormalize
    if icasefs:
        dirstate = ctx.repo().dirstate

class patternmatcher(basematcher):
    """Matches a set of (kind, pat, source) against a 'root' directory.

    >>> kindpats = [
    ...     ('re', '.*\.c$', ''),
    ...     ('path', 'foo/a', ''),
    ...     ('relpath', 'b', ''),
    ...     ('glob', '*.h', ''),
    ... ]
    >>> m = patternmatcher('foo', kindpats)
    >>> m('main.c')  # matches re:.*\.c$
    True
    >>> m('b.txt')
    False
    >>> m('foo/a')  # matches path:foo/a
    True
    >>> m('a')  # does not match path:b, since 'root' is 'foo'
    False
    >>> m('b')  # matches relpath:b, since 'root' is 'foo'
    True
    >>> m('lib.h')  # matches glob:*.h
    True

    >>> m.files()
    ['.', 'foo/a', 'b', '.']
    >>> m.exact('foo/a')
    True
    >>> m.exact('b')
    True
    >>> m.exact('lib.h')  # exact matches are for (rel)path kinds
    False
    """

    def __init__(self, root, kindpats, badfn=None):
        super(patternmatcher, self).__init__(badfn)

class exactmatcher(basematcher):
    r'''Matches the input files exactly. They are interpreted as paths, not
    patterns (so no kind-prefixes).

    >>> m = exactmatcher(['a.txt', 're:.*\.c$'])
    >>> m('a.txt')
    True
    >>> m('b.txt')
    False

    Input files that would be matched are exactly those returned by .files()
    >>> m.files()
    ['a.txt', 're:.*\\.c$']

    So pattern 're:.*\.c$' is not considered as a regex, but as a file name
    >>> m('main.c')
    False
    >>> m('re:.*\.c$')
    True
    '''

    def __init__(self, files, badfn=None):
        super(exactmatcher, self).__init__(badfn)

        return ('<unionmatcher matchers=%r>' % self._matchers)

def patkind(pattern, default=None):
    '''If pattern is 'kind:pat' with a known kind, return kind.

    >>> patkind('re:.*\.c$')
    're'
    >>> patkind('glob:*.c')
    'glob'
    >>> patkind('relpath:test.py')
    'relpath'
    >>> patkind('main.py')
    >>> patkind('main.py', default='re')
    're'
    '''
    return _patsplit(pattern, default)[0]

def _patsplit(pattern, default):

42098:5753e5949b51

    'subinclude:<path>' - a file of patterns to match against files under
                          the same directory
    '<something>' - a pattern of the specified default type

    Usually a patternmatcher is returned:
    >>> match(b'foo', b'.', [b're:.*\.c$', b'path:foo/a', b'*.py'])
    <patternmatcher patterns='.*\\.c$|foo/a(?:/|$)|[^/]*\\.py$'>

    Combining 'patterns' with 'include' (resp. 'exclude') gives an
    intersectionmatcher (resp. a differencematcher):
    >>> type(match(b'foo', b'.', [b're:.*\.c$'], include=[b'path:lib']))
    <class 'mercurial.match.intersectionmatcher'>
    >>> type(match(b'foo', b'.', [b're:.*\.c$'], exclude=[b'path:build']))
    <class 'mercurial.match.differencematcher'>

    Notice that, if 'patterns' is empty, an alwaysmatcher is returned:
    >>> match(b'foo', b'.', [])
    <alwaysmatcher>

    The 'default' argument determines which kind of pattern is assumed if a
    pattern has no prefix:
    >>> match(b'foo', b'.', [b'.*\.c$'], default=b're')
    <patternmatcher patterns='.*\\.c$'>
    >>> match(b'foo', b'.', [b'main.py'], default=b'relpath')
    <patternmatcher patterns='main\\.py(?:/|$)'>
    >>> match(b'foo', b'.', [b'main.py'], default=b're')
    <patternmatcher patterns='main.py'>

    The primary use of matchers is to check whether a value (usually a file
    name) matches againset one of the patterns given at initialization. There
    are two ways of doing this check.

    >>> m = match(b'foo', b'', [b're:.*\.c$', b'relpath:a'])

    1. Calling the matcher with a file name returns True if any pattern
    matches that file name:
    >>> m(b'a')
    True
    >>> m(b'main.c')
    True
    >>> m(b'test.py')
    False

    2. Using the exact() method only returns True if the file name matches one
    of the exact patterns (i.e. not re: or glob: patterns):
    >>> m.exact(b'a')
    True
    >>> m.exact(b'main.c')
    False
    """
    normalize = _donormalize
    if icasefs:
        dirstate = ctx.repo().dirstate

class patternmatcher(basematcher):
    """Matches a set of (kind, pat, source) against a 'root' directory.

    >>> kindpats = [
    ...     (b're', b'.*\.c$', b''),
    ...     (b'path', b'foo/a', b''),
    ...     (b'relpath', b'b', b''),
    ...     (b'glob', b'*.h', b''),
    ... ]
    >>> m = patternmatcher(b'foo', kindpats)
    >>> m(b'main.c')  # matches re:.*\.c$
    True
    >>> m(b'b.txt')
    False
    >>> m(b'foo/a')  # matches path:foo/a
    True
    >>> m(b'a')  # does not match path:b, since 'root' is 'foo'
    False
    >>> m(b'b')  # matches relpath:b, since 'root' is 'foo'
    True
    >>> m(b'lib.h')  # matches glob:*.h
    True

    >>> m.files()
    ['.', 'foo/a', 'b', '.']
    >>> m.exact(b'foo/a')
    True
    >>> m.exact(b'b')
    True
    >>> m.exact(b'lib.h')  # exact matches are for (rel)path kinds
    False
    """

    def __init__(self, root, kindpats, badfn=None):
        super(patternmatcher, self).__init__(badfn)

class exactmatcher(basematcher):
    r'''Matches the input files exactly. They are interpreted as paths, not
    patterns (so no kind-prefixes).

    >>> m = exactmatcher([b'a.txt', b're:.*\.c$'])
    >>> m(b'a.txt')
    True
    >>> m(b'b.txt')
    False

    Input files that would be matched are exactly those returned by .files()
    >>> m.files()
    ['a.txt', 're:.*\\.c$']

    So pattern 're:.*\.c$' is not considered as a regex, but as a file name
    >>> m(b'main.c')
    False
    >>> m(b're:.*\.c$')
    True
    '''

    def __init__(self, files, badfn=None):
        super(exactmatcher, self).__init__(badfn)

        return ('<unionmatcher matchers=%r>' % self._matchers)

def patkind(pattern, default=None):
    '''If pattern is 'kind:pat' with a known kind, return kind.

    >>> patkind(b're:.*\.c$')
    're'
    >>> patkind(b'glob:*.c')
    'glob'
    >>> patkind(b'relpath:test.py')
    'relpath'
    >>> patkind(b'main.py')
    >>> patkind(b'main.py', default=b're')
    're'
    '''
    return _patsplit(pattern, default)[0]

def _patsplit(pattern, default):