mercurial/revsetlang.py

43075:57875cf423c9

    parser,
    pycompat,
    smartset,
    util,
)
from .utils import (
    stringutil,
)

elements = {
    # token-type: binding-strength, primary, prefix, infix, suffix
    "(": (21, None, ("group", 1, ")"), ("func", 1, ")"), None),
    "[": (21, None, None, ("subscript", 1, "]"), None),
    "#": (21, None, None, ("relation", 21), None),
    "##": (20, None, None, ("_concat", 20), None),
    "~": (18, None, None, ("ancestor", 18), None),
    "^": (18, None, None, ("parent", 18), "parentpost"),
    "-": (5, None, ("negate", 19), ("minus", 5), None),
    "::": (17, "dagrangeall", ("dagrangepre", 17), ("dagrange", 17),
           "dagrangepost"),
    "..": (17, "dagrangeall", ("dagrangepre", 17), ("dagrange", 17),
           "dagrangepost"),
    ":": (15, "rangeall", ("rangepre", 15), ("range", 15), "rangepost"),
    "not": (10, None, ("not", 10), None, None),
    "!": (10, None, ("not", 10), None, None),
    "and": (5, None, None, ("and", 5), None),
    "&": (5, None, None, ("and", 5), None),

_quoteletters = {'"', "'"}
_simpleopletters = set(pycompat.iterbytestr("()[]#:=,-|&+!~^%"))

# default set of valid characters for the initial letter of symbols
_syminitletters = set(pycompat.iterbytestr(
    pycompat.sysbytes(string.ascii_letters) +
    pycompat.sysbytes(string.digits) +
    '._@')) | set(map(pycompat.bytechr, pycompat.xrange(128, 256)))

# default set of valid characters for non-initial letters of symbols
_symletters = _syminitletters | set(pycompat.iterbytestr('-/'))

def tokenize(program, lookup=None, syminitletters=None, symletters=None):
    '''
    Parse a revset statement into a stream of tokens

    >>> list(tokenize(b"@::"))
    [('symbol', '@', 0), ('::', None, 1), ('end', None, 3)]

    '''
    if not isinstance(program, bytes):
        raise error.ProgrammingError('revset statement must be bytes, got %r'
                                     % program)
    program = pycompat.bytestr(program)
    if syminitletters is None:
        syminitletters = _syminitletters
    if symletters is None:
        symletters = _symletters

            return

    pos, l = 0, len(program)
    while pos < l:
        c = program[pos]
        if c.isspace(): # skip inter-token whitespace
            pass
        elif c == ':' and program[pos:pos + 2] == '::': # look ahead carefully
            yield ('::', None, pos)
            pos += 1 # skip ahead
        elif c == '.' and program[pos:pos + 2] == '..': # look ahead carefully
            yield ('..', None, pos)
            pos += 1 # skip ahead
        elif c == '#' and program[pos:pos + 2] == '##': # look ahead carefully
            yield ('##', None, pos)
            pos += 1 # skip ahead
        elif c in _simpleopletters: # handle simple operators
            yield (c, None, pos)
        elif (c in _quoteletters or c == 'r' and
              program[pos:pos + 2] in ("r'", 'r"')): # handle quoted strings
            if c == 'r':
                pos += 1
                c = program[pos]
                decode = lambda x: x
            else:
                decode = parser.unescapestr
            pos += 1
            s = pos
            while pos < l: # find closing quote
                d = program[pos]
                if d == '\\': # skip over escaped characters
                    pos += 2
                    continue
                if d == c:
                    yield ('string', decode(program[s:pos]), s)
                    break

                raise error.ParseError(_("unterminated string"), s)
        # gather up a symbol/keyword
        elif c in syminitletters:
            s = pos
            pos += 1
            while pos < l: # find end of symbol
                d = program[pos]
                if d not in symletters:
                    break
                if d == '.' and program[pos - 1] == '.': # special case for ..
                    pos -= 1
                    break
                pos += 1
            sym = program[s:pos]
            if sym in keywords: # operator keywords
                yield (sym, None, s)
            elif '-' in sym:
                # some jerk gave us foo-bar-baz, try to check if it's a symbol
                if lookup and lookup(sym):
                    # looks like a real symbol
                    yield ('symbol', sym, s)
                else:
                    # looks like an expression
                    parts = sym.split('-')
                    for p in parts[:-1]:
                        if p: # possible consecutive -
                            yield ('symbol', p, s)
                        s += len(p)
                        yield ('-', None, s)
                        s += 1
                    if parts[-1]: # possible trailing -
                        yield ('symbol', parts[-1], s)
            else:
                yield ('symbol', sym, s)
            pos -= 1
        else:
            raise error.ParseError(_("syntax error in revset '%s'") %
                                   program, pos)
        pos += 1
    yield ('end', None, pos)

# helpers

_notset = object()

def getsymbol(x):
    if x and x[0] == 'symbol':
        return x[1]
    raise error.ParseError(_('not a symbol'))

def getstring(x, err):
    if x and (x[0] == 'string' or x[0] == 'symbol'):
        return x[1]
    raise error.ParseError(err)

def getinteger(x, err, default=_notset):
    if not x and default is not _notset:
        return default
    try:
        return int(getstring(x, err))
    except ValueError:
        raise error.ParseError(err)

def getboolean(x, err):
    value = stringutil.parsebool(getsymbol(x))
    if value is not None:
        return value
    raise error.ParseError(err)

def getlist(x):
    if not x:
        return []
    if x[0] == 'list':
        return list(x[1:])
    return [x]

def getrange(x, err):
    if not x:
        raise error.ParseError(err)
    op = x[0]

        return x[1], None
    elif op == 'rangeall':
        return None, None
    raise error.ParseError(err)

def getintrange(x, err1, err2, deffirst=_notset, deflast=_notset):
    """Get [first, last] integer range (both inclusive) from a parsed tree

    If any of the sides omitted, and if no default provided, ParseError will
    be raised.

        n = getinteger(x, err1)
        return n, n
    a, b = getrange(x, err1)
    return getinteger(a, err2, deffirst), getinteger(b, err2, deflast)

def getargs(x, min, max, err):
    l = getlist(x)
    if len(l) < min or (max >= 0 and len(l) > max):
        raise error.ParseError(err)
    return l

def getargsdict(x, funcname, keys):
    return parser.buildargsdict(getlist(x), funcname, parser.splitargspec(keys),
                                keyvaluenode='keyvalue', keynode='symbol')

# cache of {spec: raw parsed tree} built internally
_treecache = {}

def _cachedtree(spec):
    # thread safe because parse() is reentrant and dict.__setitem__() is atomic
    tree = _treecache.get(spec)
    if tree is None:
        _treecache[spec] = tree = parse(spec)
    return tree

def _build(tmplspec, *repls):
    """Create raw parsed tree from a template revset statement

    >>> _build(b'f(_) and _', (b'string', b'1'), (b'symbol', b'2'))
    ('and', ('func', ('symbol', 'f'), ('string', '1')), ('symbol', '2'))
    """
    template = _cachedtree(tmplspec)
    return parser.buildtree(template, ('symbol', '_'), *repls)

def _match(patspec, tree):
    """Test if a tree matches the given pattern statement; return the matches

    >>> _match(b'f(_)', parse(b'f()'))
    >>> _match(b'f(_)', parse(b'f(1)'))
    [('func', ('symbol', 'f'), ('symbol', '1')), ('symbol', '1')]
    >>> _match(b'f(_)', parse(b'f(1, 2)'))
    """
    pattern = _cachedtree(patspec)
    return parser.matchtree(pattern, tree, ('symbol', '_'),
                            {'keyvalue', 'list'})

def _matchonly(revs, bases):
    return _match('ancestors(_) and not ancestors(_)', ('and', revs, bases))

def _fixops(x):
    """Rewrite raw parsed tree to resolve ambiguous syntax which cannot be
    handled well by our simple top-down parser"""
    if not isinstance(x, tuple):

    elif op == 'subscript' and x[1][0] == 'relation':
        # x#y[z] ternary
        return _fixops(('relsubscript', x[1][1], x[1][2], x[2]))

    return (op,) + tuple(_fixops(y) for y in x[1:])

def _analyze(x):
    if x is None:
        return x

        return (op, None)
    elif op in {'or', 'not', 'rangepre', 'rangepost', 'parentpost'}:
        return (op, _analyze(x[1]))
    elif op == 'group':
        return _analyze(x[1])
    elif op in {'and', 'dagrange', 'range', 'parent', 'ancestor', 'relation',
                'subscript'}:
        ta = _analyze(x[1])
        tb = _analyze(x[2])
        return (op, ta, tb)
    elif op == 'relsubscript':
        ta = _analyze(x[1])

        return (op, x[1], _analyze(x[2]))
    elif op == 'func':
        return (op, x[1], _analyze(x[2]))
    raise ValueError('invalid operator %r' % op)

def analyze(x):
    """Transform raw parsed tree to evaluatable tree which can be fed to
    optimize() or getset()

    All pseudo operations should be mapped to real operations or functions
    defined in methods or symbols table respectively.
    """
    return _analyze(x)

def _optimize(x):
    if x is None:
        return 0, x

    op = x[0]
    if op in ('string', 'symbol', 'smartset'):
        return 0.5, x # single revisions are small
    elif op == 'and':
        wa, ta = _optimize(x[1])
        wb, tb = _optimize(x[2])
        w = min(wa, wb)

        return w, (op, ta, tb)
    elif op == 'or':
        # fast path for machine-generated expression, that is likely to have
        # lots of trivial revisions: 'a + b + c()' to '_list(a b) + c()'
        ws, ts, ss = [], [], []
        def flushss():
            if not ss:
                return
            if len(ss) == 1:
                w, t = ss[0]

                y = _build('_list(_)', ('string', s))
                w, t = _optimize(y)
            ws.append(w)
            ts.append(t)
            del ss[:]
        for y in getlist(x[1]):
            w, t = _optimize(y)
            if t is not None and (t[0] == 'string' or t[0] == 'symbol'):
                ss.append((w, t))
                continue
            flushss()
            ws.append(w)
            ts.append(t)
        flushss()
        if len(ts) == 1:
            return ws[0], ts[0] # 'or' operation is fully optimized out
        return max(ws), (op, ('list',) + tuple(ts))
    elif op == 'not':
        # Optimize not public() to _notpublic() because we have a fast version
        if _match('public()', x[1]):
            o = _optimize(_build('_notpublic()'))

            return w + wa, _build('_commonancestorheads(_)', m[1])

        return w + wa, (op, x[1], ta)
    raise ValueError('invalid operator %r' % op)

def optimize(tree):
    """Optimize evaluatable tree

    All pseudo operations should be transformed beforehand.
    """
    _weight, newtree = _optimize(tree)
    return newtree

# the set of valid characters for the initial letter of symbols in
# alias declarations and definitions
_aliassyminitletters = _syminitletters | {'$'}

def _parsewith(spec, lookup=None, syminitletters=None):
    """Generate a parse tree of given spec with given tokenizing options

    >>> _parsewith(b'foo($1)', syminitletters=_aliassyminitletters)

    ParseError: ('invalid token', 4)
    """
    if lookup and spec.startswith('revset(') and spec.endswith(')'):
        lookup = None
    p = parser.parser(elements)
    tree, pos = p.parse(tokenize(spec, lookup=lookup,
                                 syminitletters=syminitletters))
    if pos != len(spec):
        raise error.ParseError(_('invalid token'), pos)
    return _fixops(parser.simplifyinfixops(tree, ('list', 'or')))

class _aliasrules(parser.basealiasrules):
    """Parsing and expansion rule set of revset aliases"""
    _section = _('revset alias')

    @staticmethod
    def _parse(spec):
        """Parse alias declaration/definition ``spec``

    @staticmethod
    def _trygetfunc(tree):
        if tree[0] == 'func' and tree[1][0] == 'symbol':
            return tree[1][1], getlist(tree[2])

def expandaliases(tree, aliases, warn=None):
    """Expand aliases in a tree, aliases is a list of (name, value) tuples"""
    aliases = _aliasrules.buildmap(aliases)
    tree = _aliasrules.expand(aliases, tree)

            if alias.error and not alias.warned:
                warn(_('warning: %s\n') % (alias.error))
                alias.warned = True
    return tree

def foldconcat(tree):
    """Fold elements to be concatenated by `##`
    """
    if (not isinstance(tree, tuple)
        or tree[0] in ('string', 'symbol', 'smartset')):
        return tree
    if tree[0] == '_concat':
        pending = [tree]
        l = []
        while pending:

                raise error.ParseError(msg)
        return ('string', ''.join(l))
    else:
        return tuple(foldconcat(t) for t in tree)

def parse(spec, lookup=None):
    try:
        return _parsewith(spec, lookup=lookup)
    except error.ParseError as inst:
        if len(inst.args) > 1:  # has location

            # start. Therefore, we print "loc + 1" spaces (instead of "loc")
            # to line up the caret with the location of the error.
            inst.hint = spec + '\n' + ' ' * (loc + 1) + '^ ' + _('here')
        raise

def _quote(s):
    r"""Quote a value in order to make it safe for the revset engine.

    >>> _quote(b'asdf')
    "'asdf'"

    >>> _quote(1)
    "'1'"
    """
    return "'%s'" % stringutil.escapestr(pycompat.bytestr(s))

def _formatargtype(c, arg):
    if c == 'd':
        return '_rev(%d)' % int(arg)
    elif c == 's':
        return _quote(arg)
    elif c == 'r':
        if not isinstance(arg, bytes):
            raise TypeError
        parse(arg) # make sure syntax errors are confined
        return '(%s)' % arg
    elif c == 'n':
        return _quote(node.hex(arg))
    elif c == 'b':
        try:
            return _quote(arg.branch())
        except AttributeError:
            raise TypeError
    raise error.ParseError(_('unexpected revspec format character %s') % c)

def _formatlistexp(s, t):
    l = len(s)
    if l == 0:
        return "_list('')"

            raise TypeError

    m = l // 2
    return '(%s or %s)' % (_formatlistexp(s[:m], t), _formatlistexp(s[m:], t))

def _formatintlist(data):
    try:
        l = len(data)
        if l == 0:
            return "_list('')"

            return _formatargtype('d', data[0])
        return "_intlist('%s')" % "\0".join('%d' % int(a) for a in data)
    except (TypeError, ValueError):
        raise error.ParseError(_('invalid argument for revspec'))

def _formatparamexp(args, t):
    return ', '.join(_formatargtype(t, a) for a in args)

_formatlistfuncs = {
    'l': _formatlistexp,
    'p': _formatparamexp,
}

def formatspec(expr, *args):
    '''
    This is a convenience function for using revsets internally, and
    escapes arguments appropriately. Aliases are intentionally ignored

            ret.append(_formatintlist(list(arg)))
        else:
            raise error.ProgrammingError("unknown revspec item type: %r" % t)
    return b''.join(ret)

def spectree(expr, *args):
    """similar to formatspec but return a parsed and optimized tree"""
    parsed = _parseargs(expr, args)
    ret = []
    inputs = []

    tree = parser.buildtree(tree, ('symbol', '$'), *inputs)
    tree = foldconcat(tree)
    tree = analyze(tree)
    tree = optimize(tree)
    return tree

def _parseargs(expr, args):
    """parse the expression and replace all inexpensive args

    return a list of tuple [(arg-type, arg-value)]

            raise error.ParseError(_('missing argument for revspec'))
        f = _formatlistfuncs.get(d)
        if f:
            # a list of some type, might be expensive, do not replace
            pos += 1
            islist = (d == 'l')
            try:
                d = expr[pos]
            except IndexError:
                raise error.ParseError(_('incomplete revspec format character'))
            if islist and d == 'd' and arg:

        raise error.ParseError(_('too many revspec arguments specified'))
    except StopIteration:
        pass
    return ret

def prettyformat(tree):
    return parser.prettyformat(tree, ('string', 'symbol'))

def depth(tree):
    if isinstance(tree, tuple):
        return max(map(depth, tree)) + 1
    else:
        return 0

def funcsused(tree):
    if not isinstance(tree, tuple) or tree[0] in ('string', 'symbol'):
        return set()
    else:

            funcs |= funcsused(s)
        if tree[0] == 'func':
            funcs.add(tree[1][1])
        return funcs

_hashre = util.re.compile('[0-9a-fA-F]{1,40}$')

def _ishashlikesymbol(symbol):
    """returns true if the symbol looks like a hash"""
    return _hashre.match(symbol)

def gethashlikesymbols(tree):
    """returns the list of symbols of the tree that look like hashes

    >>> gethashlikesymbols(parse(b'3::abe3ff'))

43076:2372284d9457

    parser,
    pycompat,
    smartset,
    util,
)
from .utils import stringutil

elements = {
    # token-type: binding-strength, primary, prefix, infix, suffix
    "(": (21, None, ("group", 1, ")"), ("func", 1, ")"), None),
    "[": (21, None, None, ("subscript", 1, "]"), None),
    "#": (21, None, None, ("relation", 21), None),
    "##": (20, None, None, ("_concat", 20), None),
    "~": (18, None, None, ("ancestor", 18), None),
    "^": (18, None, None, ("parent", 18), "parentpost"),
    "-": (5, None, ("negate", 19), ("minus", 5), None),
    "::": (
        17,
        "dagrangeall",
        ("dagrangepre", 17),
        ("dagrange", 17),
        "dagrangepost",
    ),
    "..": (
        17,
        "dagrangeall",
        ("dagrangepre", 17),
        ("dagrange", 17),
        "dagrangepost",
    ),
    ":": (15, "rangeall", ("rangepre", 15), ("range", 15), "rangepost"),
    "not": (10, None, ("not", 10), None, None),
    "!": (10, None, ("not", 10), None, None),
    "and": (5, None, None, ("and", 5), None),
    "&": (5, None, None, ("and", 5), None),

_quoteletters = {'"', "'"}
_simpleopletters = set(pycompat.iterbytestr("()[]#:=,-|&+!~^%"))

# default set of valid characters for the initial letter of symbols
_syminitletters = set(
    pycompat.iterbytestr(
        pycompat.sysbytes(string.ascii_letters)
        + pycompat.sysbytes(string.digits)
        + '._@'
    )
) | set(map(pycompat.bytechr, pycompat.xrange(128, 256)))

# default set of valid characters for non-initial letters of symbols
_symletters = _syminitletters | set(pycompat.iterbytestr('-/'))


def tokenize(program, lookup=None, syminitletters=None, symletters=None):
    '''
    Parse a revset statement into a stream of tokens

    >>> list(tokenize(b"@::"))
    [('symbol', '@', 0), ('::', None, 1), ('end', None, 3)]

    '''
    if not isinstance(program, bytes):
        raise error.ProgrammingError(
            'revset statement must be bytes, got %r' % program
        )
    program = pycompat.bytestr(program)
    if syminitletters is None:
        syminitletters = _syminitletters
    if symletters is None:
        symletters = _symletters

            return

    pos, l = 0, len(program)
    while pos < l:
        c = program[pos]
        if c.isspace():  # skip inter-token whitespace
            pass
        elif (
            c == ':' and program[pos : pos + 2] == '::'
        ):  # look ahead carefully
            yield ('::', None, pos)
            pos += 1  # skip ahead
        elif (
            c == '.' and program[pos : pos + 2] == '..'
        ):  # look ahead carefully
            yield ('..', None, pos)
            pos += 1  # skip ahead
        elif (
            c == '#' and program[pos : pos + 2] == '##'
        ):  # look ahead carefully
            yield ('##', None, pos)
            pos += 1  # skip ahead
        elif c in _simpleopletters:  # handle simple operators
            yield (c, None, pos)
        elif (
            c in _quoteletters
            or c == 'r'
            and program[pos : pos + 2] in ("r'", 'r"')
        ):  # handle quoted strings
            if c == 'r':
                pos += 1
                c = program[pos]
                decode = lambda x: x
            else:
                decode = parser.unescapestr
            pos += 1
            s = pos
            while pos < l:  # find closing quote
                d = program[pos]
                if d == '\\':  # skip over escaped characters
                    pos += 2
                    continue
                if d == c:
                    yield ('string', decode(program[s:pos]), s)
                    break

                raise error.ParseError(_("unterminated string"), s)
        # gather up a symbol/keyword
        elif c in syminitletters:
            s = pos
            pos += 1
            while pos < l:  # find end of symbol
                d = program[pos]
                if d not in symletters:
                    break
                if d == '.' and program[pos - 1] == '.':  # special case for ..
                    pos -= 1
                    break
                pos += 1
            sym = program[s:pos]
            if sym in keywords:  # operator keywords
                yield (sym, None, s)
            elif '-' in sym:
                # some jerk gave us foo-bar-baz, try to check if it's a symbol
                if lookup and lookup(sym):
                    # looks like a real symbol
                    yield ('symbol', sym, s)
                else:
                    # looks like an expression
                    parts = sym.split('-')
                    for p in parts[:-1]:
                        if p:  # possible consecutive -
                            yield ('symbol', p, s)
                        s += len(p)
                        yield ('-', None, s)
                        s += 1
                    if parts[-1]:  # possible trailing -
                        yield ('symbol', parts[-1], s)
            else:
                yield ('symbol', sym, s)
            pos -= 1
        else:
            raise error.ParseError(
                _("syntax error in revset '%s'") % program, pos
            )
        pos += 1
    yield ('end', None, pos)


# helpers

_notset = object()


def getsymbol(x):
    if x and x[0] == 'symbol':
        return x[1]
    raise error.ParseError(_('not a symbol'))


def getstring(x, err):
    if x and (x[0] == 'string' or x[0] == 'symbol'):
        return x[1]
    raise error.ParseError(err)


def getinteger(x, err, default=_notset):
    if not x and default is not _notset:
        return default
    try:
        return int(getstring(x, err))
    except ValueError:
        raise error.ParseError(err)


def getboolean(x, err):
    value = stringutil.parsebool(getsymbol(x))
    if value is not None:
        return value
    raise error.ParseError(err)


def getlist(x):
    if not x:
        return []
    if x[0] == 'list':
        return list(x[1:])
    return [x]


def getrange(x, err):
    if not x:
        raise error.ParseError(err)
    op = x[0]

        return x[1], None
    elif op == 'rangeall':
        return None, None
    raise error.ParseError(err)


def getintrange(x, err1, err2, deffirst=_notset, deflast=_notset):
    """Get [first, last] integer range (both inclusive) from a parsed tree

    If any of the sides omitted, and if no default provided, ParseError will
    be raised.

        n = getinteger(x, err1)
        return n, n
    a, b = getrange(x, err1)
    return getinteger(a, err2, deffirst), getinteger(b, err2, deflast)


def getargs(x, min, max, err):
    l = getlist(x)
    if len(l) < min or (max >= 0 and len(l) > max):
        raise error.ParseError(err)
    return l


def getargsdict(x, funcname, keys):
    return parser.buildargsdict(
        getlist(x),
        funcname,
        parser.splitargspec(keys),
        keyvaluenode='keyvalue',
        keynode='symbol',
    )


# cache of {spec: raw parsed tree} built internally
_treecache = {}


def _cachedtree(spec):
    # thread safe because parse() is reentrant and dict.__setitem__() is atomic
    tree = _treecache.get(spec)
    if tree is None:
        _treecache[spec] = tree = parse(spec)
    return tree


def _build(tmplspec, *repls):
    """Create raw parsed tree from a template revset statement

    >>> _build(b'f(_) and _', (b'string', b'1'), (b'symbol', b'2'))
    ('and', ('func', ('symbol', 'f'), ('string', '1')), ('symbol', '2'))
    """
    template = _cachedtree(tmplspec)
    return parser.buildtree(template, ('symbol', '_'), *repls)


def _match(patspec, tree):
    """Test if a tree matches the given pattern statement; return the matches

    >>> _match(b'f(_)', parse(b'f()'))
    >>> _match(b'f(_)', parse(b'f(1)'))
    [('func', ('symbol', 'f'), ('symbol', '1')), ('symbol', '1')]
    >>> _match(b'f(_)', parse(b'f(1, 2)'))
    """
    pattern = _cachedtree(patspec)
    return parser.matchtree(
        pattern, tree, ('symbol', '_'), {'keyvalue', 'list'}
    )


def _matchonly(revs, bases):
    return _match('ancestors(_) and not ancestors(_)', ('and', revs, bases))


def _fixops(x):
    """Rewrite raw parsed tree to resolve ambiguous syntax which cannot be
    handled well by our simple top-down parser"""
    if not isinstance(x, tuple):

    elif op == 'subscript' and x[1][0] == 'relation':
        # x#y[z] ternary
        return _fixops(('relsubscript', x[1][1], x[1][2], x[2]))

    return (op,) + tuple(_fixops(y) for y in x[1:])


def _analyze(x):
    if x is None:
        return x

        return (op, None)
    elif op in {'or', 'not', 'rangepre', 'rangepost', 'parentpost'}:
        return (op, _analyze(x[1]))
    elif op == 'group':
        return _analyze(x[1])
    elif op in {
        'and',
        'dagrange',
        'range',
        'parent',
        'ancestor',
        'relation',
        'subscript',
    }:
        ta = _analyze(x[1])
        tb = _analyze(x[2])
        return (op, ta, tb)
    elif op == 'relsubscript':
        ta = _analyze(x[1])

        return (op, x[1], _analyze(x[2]))
    elif op == 'func':
        return (op, x[1], _analyze(x[2]))
    raise ValueError('invalid operator %r' % op)


def analyze(x):
    """Transform raw parsed tree to evaluatable tree which can be fed to
    optimize() or getset()

    All pseudo operations should be mapped to real operations or functions
    defined in methods or symbols table respectively.
    """
    return _analyze(x)


def _optimize(x):
    if x is None:
        return 0, x

    op = x[0]
    if op in ('string', 'symbol', 'smartset'):
        return 0.5, x  # single revisions are small
    elif op == 'and':
        wa, ta = _optimize(x[1])
        wb, tb = _optimize(x[2])
        w = min(wa, wb)

        return w, (op, ta, tb)
    elif op == 'or':
        # fast path for machine-generated expression, that is likely to have
        # lots of trivial revisions: 'a + b + c()' to '_list(a b) + c()'
        ws, ts, ss = [], [], []

        def flushss():
            if not ss:
                return
            if len(ss) == 1:
                w, t = ss[0]

                y = _build('_list(_)', ('string', s))
                w, t = _optimize(y)
            ws.append(w)
            ts.append(t)
            del ss[:]

        for y in getlist(x[1]):
            w, t = _optimize(y)
            if t is not None and (t[0] == 'string' or t[0] == 'symbol'):
                ss.append((w, t))
                continue
            flushss()
            ws.append(w)
            ts.append(t)
        flushss()
        if len(ts) == 1:
            return ws[0], ts[0]  # 'or' operation is fully optimized out
        return max(ws), (op, ('list',) + tuple(ts))
    elif op == 'not':
        # Optimize not public() to _notpublic() because we have a fast version
        if _match('public()', x[1]):
            o = _optimize(_build('_notpublic()'))

            return w + wa, _build('_commonancestorheads(_)', m[1])

        return w + wa, (op, x[1], ta)
    raise ValueError('invalid operator %r' % op)


def optimize(tree):
    """Optimize evaluatable tree

    All pseudo operations should be transformed beforehand.
    """
    _weight, newtree = _optimize(tree)
    return newtree


# the set of valid characters for the initial letter of symbols in
# alias declarations and definitions
_aliassyminitletters = _syminitletters | {'$'}


def _parsewith(spec, lookup=None, syminitletters=None):
    """Generate a parse tree of given spec with given tokenizing options

    >>> _parsewith(b'foo($1)', syminitletters=_aliassyminitletters)

    ParseError: ('invalid token', 4)
    """
    if lookup and spec.startswith('revset(') and spec.endswith(')'):
        lookup = None
    p = parser.parser(elements)
    tree, pos = p.parse(
        tokenize(spec, lookup=lookup, syminitletters=syminitletters)
    )
    if pos != len(spec):
        raise error.ParseError(_('invalid token'), pos)
    return _fixops(parser.simplifyinfixops(tree, ('list', 'or')))


class _aliasrules(parser.basealiasrules):
    """Parsing and expansion rule set of revset aliases"""

    _section = _('revset alias')

    @staticmethod
    def _parse(spec):
        """Parse alias declaration/definition ``spec``

    @staticmethod
    def _trygetfunc(tree):
        if tree[0] == 'func' and tree[1][0] == 'symbol':
            return tree[1][1], getlist(tree[2])


def expandaliases(tree, aliases, warn=None):
    """Expand aliases in a tree, aliases is a list of (name, value) tuples"""
    aliases = _aliasrules.buildmap(aliases)
    tree = _aliasrules.expand(aliases, tree)

            if alias.error and not alias.warned:
                warn(_('warning: %s\n') % (alias.error))
                alias.warned = True
    return tree


def foldconcat(tree):
    """Fold elements to be concatenated by `##`
    """
    if not isinstance(tree, tuple) or tree[0] in (
        'string',
        'symbol',
        'smartset',
    ):
        return tree
    if tree[0] == '_concat':
        pending = [tree]
        l = []
        while pending:

                raise error.ParseError(msg)
        return ('string', ''.join(l))
    else:
        return tuple(foldconcat(t) for t in tree)


def parse(spec, lookup=None):
    try:
        return _parsewith(spec, lookup=lookup)
    except error.ParseError as inst:
        if len(inst.args) > 1:  # has location

            # start. Therefore, we print "loc + 1" spaces (instead of "loc")
            # to line up the caret with the location of the error.
            inst.hint = spec + '\n' + ' ' * (loc + 1) + '^ ' + _('here')
        raise


def _quote(s):
    r"""Quote a value in order to make it safe for the revset engine.

    >>> _quote(b'asdf')
    "'asdf'"

    >>> _quote(1)
    "'1'"
    """
    return "'%s'" % stringutil.escapestr(pycompat.bytestr(s))


def _formatargtype(c, arg):
    if c == 'd':
        return '_rev(%d)' % int(arg)
    elif c == 's':
        return _quote(arg)
    elif c == 'r':
        if not isinstance(arg, bytes):
            raise TypeError
        parse(arg)  # make sure syntax errors are confined
        return '(%s)' % arg
    elif c == 'n':
        return _quote(node.hex(arg))
    elif c == 'b':
        try:
            return _quote(arg.branch())
        except AttributeError:
            raise TypeError
    raise error.ParseError(_('unexpected revspec format character %s') % c)


def _formatlistexp(s, t):
    l = len(s)
    if l == 0:
        return "_list('')"

            raise TypeError

    m = l // 2
    return '(%s or %s)' % (_formatlistexp(s[:m], t), _formatlistexp(s[m:], t))


def _formatintlist(data):
    try:
        l = len(data)
        if l == 0:
            return "_list('')"

            return _formatargtype('d', data[0])
        return "_intlist('%s')" % "\0".join('%d' % int(a) for a in data)
    except (TypeError, ValueError):
        raise error.ParseError(_('invalid argument for revspec'))


def _formatparamexp(args, t):
    return ', '.join(_formatargtype(t, a) for a in args)


_formatlistfuncs = {
    'l': _formatlistexp,
    'p': _formatparamexp,
}


def formatspec(expr, *args):
    '''
    This is a convenience function for using revsets internally, and
    escapes arguments appropriately. Aliases are intentionally ignored

            ret.append(_formatintlist(list(arg)))
        else:
            raise error.ProgrammingError("unknown revspec item type: %r" % t)
    return b''.join(ret)


def spectree(expr, *args):
    """similar to formatspec but return a parsed and optimized tree"""
    parsed = _parseargs(expr, args)
    ret = []
    inputs = []

    tree = parser.buildtree(tree, ('symbol', '$'), *inputs)
    tree = foldconcat(tree)
    tree = analyze(tree)
    tree = optimize(tree)
    return tree


def _parseargs(expr, args):
    """parse the expression and replace all inexpensive args

    return a list of tuple [(arg-type, arg-value)]

            raise error.ParseError(_('missing argument for revspec'))
        f = _formatlistfuncs.get(d)
        if f:
            # a list of some type, might be expensive, do not replace
            pos += 1
            islist = d == 'l'
            try:
                d = expr[pos]
            except IndexError:
                raise error.ParseError(_('incomplete revspec format character'))
            if islist and d == 'd' and arg:

        raise error.ParseError(_('too many revspec arguments specified'))
    except StopIteration:
        pass
    return ret


def prettyformat(tree):
    return parser.prettyformat(tree, ('string', 'symbol'))


def depth(tree):
    if isinstance(tree, tuple):
        return max(map(depth, tree)) + 1
    else:
        return 0


def funcsused(tree):
    if not isinstance(tree, tuple) or tree[0] in ('string', 'symbol'):
        return set()
    else:

            funcs |= funcsused(s)
        if tree[0] == 'func':
            funcs.add(tree[1][1])
        return funcs


_hashre = util.re.compile('[0-9a-fA-F]{1,40}$')


def _ishashlikesymbol(symbol):
    """returns true if the symbol looks like a hash"""
    return _hashre.match(symbol)


def gethashlikesymbols(tree):
    """returns the list of symbols of the tree that look like hashes

    >>> gethashlikesymbols(parse(b'3::abe3ff'))