Mercurial > hg
view mercurial/parser.py @ 42745:4d20b1fe8a72
rust-discovery: using from Python code
As previously done in other topics, the Rust version is used if it's been
built.
The version fully in Rust of the partialdiscovery class has the performance
advantage over the Python version (actually using the Rust MissingAncestor) if
the undecided set is big enough. Otherwise no sampling occurs, and the
discovery is reasonably fast anyway.
Note: it's hard to predict the size of the initial undecided set, it can
depend on the kind of topological changes between the local and remote graphs.
The point of the Rust version is to make the bad cases acceptable.
More specifically, the performance advantages are:
- faster sampling, especially takefullsample()
- much faster addmissings() in almost all cases (see commit message in
grandparent of the present changeset)
- no conversion cost of the undecided set at the interface between Rust and
Python
== Measurements with big undecided sets
For an extreme example, discovery between mozilla-try and mozilla-unified
(over one million undecided revisions, same case as in dbd0fcca6dfc), we
get roughly a x2.5/x3 better performance:
Growing sample size (5% starting with 200): time goes down from
210 to 72 seconds.
Constant sample size of 200: time down from 1853 to 659 seconds.
With a sample size computed from number of roots and heads of the
undecided set (`respectsize` is `False`), here are perfdiscovery results:
Before ! wall 9.358729 comb 9.360000 user 9.310000 sys 0.050000 (median of 50)
After ! wall 3.793819 comb 3.790000 user 3.750000 sys 0.040000 (median of 50)
In that later case, the sample sizes are routinely in the hundreds of
thousands of revisions. While still faster, the Rust iteration in
addmissings has less of an advantage than with smaller sample sizes, but
one sees addcommons becoming faster, probably a consequence of not having
to copy big sets back and forth.
This example is not a goal in itself, but it showcases several different
areas in which the process can become slow, due to different factors, and
how this full Rust version can help.
== Measurements with small undecided sets
In cases the undecided set is small enough than no sampling occurs,
the Rust version has a disadvantage at init if `targetheads` is really big
(some time is lost in the translation to Rust data structures),
and that is compensated by the faster `addmissings()`.
On a private repository with over one million commits, we still get a minor
improvement, of 6.8%:
Before ! wall 0.593585 comb 0.590000 user 0.550000 sys 0.040000 (median of 50)
After ! wall 0.553035 comb 0.550000 user 0.520000 sys 0.030000 (median of 50)
What's interesting in that case is the first addinfo() at 180ms for Rust and
233ms for Python+C, mostly due to add_missings and the children cache
computation being done in less than 0.2ms on the Rust side vs over 40ms on the
Python side.
The worst case we have on hand is with mozilla-try, prepared with
discovery-helper.sh for 10 heads and depth 10, time goes up 2.2% on the median.
In this case `targetheads` is really huge with 165842 server heads.
Before ! wall 0.823884 comb 0.810000 user 0.790000 sys 0.020000 (median of 50)
After ! wall 0.842607 comb 0.840000 user 0.800000 sys 0.040000 (median of 50)
If that would be considered a problem, more adjustments can be made, which are
prematurate at this stage: cooking special variants of methods of the inner
MissingAncestors object, retrieving local heads directly from Rust to avoid
the cost of conversion. Effort would probably be better spent at this point
improving the surroundings if needed.
Here's another data point with a smaller repository, pypy, where performance
is almost identical
Before ! wall 0.015121 comb 0.030000 user 0.020000 sys 0.010000 (median of 186)
After ! wall 0.015009 comb 0.010000 user 0.010000 sys 0.000000 (median of 184)
Differential Revision: https://phab.mercurial-scm.org/D6430
author | Georges Racinet <georges.racinet@octobus.net> |
---|---|
date | Wed, 20 Feb 2019 09:04:54 +0100 |
parents | 29798c9ba5c9 |
children | 2372284d9457 |
line wrap: on
line source
# parser.py - simple top-down operator precedence parser for mercurial # # Copyright 2010 Matt Mackall <mpm@selenic.com> # # This software may be used and distributed according to the terms of the # GNU General Public License version 2 or any later version. # see http://effbot.org/zone/simple-top-down-parsing.htm and # http://eli.thegreenplace.net/2010/01/02/top-down-operator-precedence-parsing/ # for background # takes a tokenizer and elements # tokenizer is an iterator that returns (type, value, pos) tuples # elements is a mapping of types to binding strength, primary, prefix, infix # and suffix actions # an action is a tree node name, a tree label, and an optional match # __call__(program) parses program into a labeled tree from __future__ import absolute_import, print_function from .i18n import _ from . import ( error, pycompat, util, ) from .utils import ( stringutil, ) class parser(object): def __init__(self, elements, methods=None): self._elements = elements self._methods = methods self.current = None def _advance(self): 'advance the tokenizer' t = self.current self.current = next(self._iter, None) return t def _hasnewterm(self): 'True if next token may start new term' return any(self._elements[self.current[0]][1:3]) def _match(self, m): 'make sure the tokenizer matches an end condition' if self.current[0] != m: raise error.ParseError(_("unexpected token: %s") % self.current[0], self.current[2]) self._advance() def _parseoperand(self, bind, m=None): 'gather right-hand-side operand until an end condition or binding met' if m and self.current[0] == m: expr = None else: expr = self._parse(bind) if m: self._match(m) return expr def _parse(self, bind=0): token, value, pos = self._advance() # handle prefix rules on current token, take as primary if unambiguous primary, prefix = self._elements[token][1:3] if primary and not (prefix and self._hasnewterm()): expr = (primary, value) elif prefix: expr = (prefix[0], self._parseoperand(*prefix[1:])) else: raise error.ParseError(_("not a prefix: %s") % token, pos) # gather tokens until we meet a lower binding strength while bind < self._elements[self.current[0]][0]: token, value, pos = self._advance() # handle infix rules, take as suffix if unambiguous infix, suffix = self._elements[token][3:] if suffix and not (infix and self._hasnewterm()): expr = (suffix, expr) elif infix: expr = (infix[0], expr, self._parseoperand(*infix[1:])) else: raise error.ParseError(_("not an infix: %s") % token, pos) return expr def parse(self, tokeniter): 'generate a parse tree from tokens' self._iter = tokeniter self._advance() res = self._parse() token, value, pos = self.current return res, pos def eval(self, tree): 'recursively evaluate a parse tree using node methods' if not isinstance(tree, tuple): return tree return self._methods[tree[0]](*[self.eval(t) for t in tree[1:]]) def __call__(self, tokeniter): 'parse tokens into a parse tree and evaluate if methods given' t = self.parse(tokeniter) if self._methods: return self.eval(t) return t def splitargspec(spec): """Parse spec of function arguments into (poskeys, varkey, keys, optkey) >>> splitargspec(b'') ([], None, [], None) >>> splitargspec(b'foo bar') ([], None, ['foo', 'bar'], None) >>> splitargspec(b'foo *bar baz **qux') (['foo'], 'bar', ['baz'], 'qux') >>> splitargspec(b'*foo') ([], 'foo', [], None) >>> splitargspec(b'**foo') ([], None, [], 'foo') """ optkey = None pre, sep, post = spec.partition('**') if sep: posts = post.split() if not posts: raise error.ProgrammingError('no **optkey name provided') if len(posts) > 1: raise error.ProgrammingError('excessive **optkey names provided') optkey = posts[0] pre, sep, post = pre.partition('*') pres = pre.split() posts = post.split() if sep: if not posts: raise error.ProgrammingError('no *varkey name provided') return pres, posts[0], posts[1:], optkey return [], None, pres, optkey def buildargsdict(trees, funcname, argspec, keyvaluenode, keynode): """Build dict from list containing positional and keyword arguments Arguments are specified by a tuple of ``(poskeys, varkey, keys, optkey)`` where - ``poskeys``: list of names of positional arguments - ``varkey``: optional argument name that takes up remainder - ``keys``: list of names that can be either positional or keyword arguments - ``optkey``: optional argument name that takes up excess keyword arguments If ``varkey`` specified, all ``keys`` must be given as keyword arguments. Invalid keywords, too few positional arguments, or too many positional arguments are rejected, but missing keyword arguments are just omitted. """ poskeys, varkey, keys, optkey = argspec kwstart = next((i for i, x in enumerate(trees) if x and x[0] == keyvaluenode), len(trees)) if kwstart < len(poskeys): raise error.ParseError(_("%(func)s takes at least %(nargs)d positional " "arguments") % {'func': funcname, 'nargs': len(poskeys)}) if not varkey and kwstart > len(poskeys) + len(keys): raise error.ParseError(_("%(func)s takes at most %(nargs)d positional " "arguments") % {'func': funcname, 'nargs': len(poskeys) + len(keys)}) args = util.sortdict() # consume positional arguments for k, x in zip(poskeys, trees[:kwstart]): args[k] = x if varkey: args[varkey] = trees[len(args):kwstart] else: for k, x in zip(keys, trees[len(args):kwstart]): args[k] = x # remainder should be keyword arguments if optkey: args[optkey] = util.sortdict() for x in trees[kwstart:]: if not x or x[0] != keyvaluenode or x[1][0] != keynode: raise error.ParseError(_("%(func)s got an invalid argument") % {'func': funcname}) k = x[1][1] if k in keys: d = args elif not optkey: raise error.ParseError(_("%(func)s got an unexpected keyword " "argument '%(key)s'") % {'func': funcname, 'key': k}) else: d = args[optkey] if k in d: raise error.ParseError(_("%(func)s got multiple values for keyword " "argument '%(key)s'") % {'func': funcname, 'key': k}) d[k] = x[2] return args def unescapestr(s): try: return stringutil.unescapestr(s) except ValueError as e: # mangle Python's exception into our format raise error.ParseError(pycompat.bytestr(e).lower()) def _prettyformat(tree, leafnodes, level, lines): if not isinstance(tree, tuple): lines.append((level, stringutil.pprint(tree))) elif tree[0] in leafnodes: rs = map(stringutil.pprint, tree[1:]) lines.append((level, '(%s %s)' % (tree[0], ' '.join(rs)))) else: lines.append((level, '(%s' % tree[0])) for s in tree[1:]: _prettyformat(s, leafnodes, level + 1, lines) lines[-1:] = [(lines[-1][0], lines[-1][1] + ')')] def prettyformat(tree, leafnodes): lines = [] _prettyformat(tree, leafnodes, 0, lines) output = '\n'.join((' ' * l + s) for l, s in lines) return output def simplifyinfixops(tree, targetnodes): """Flatten chained infix operations to reduce usage of Python stack >>> from . import pycompat >>> def f(tree): ... s = prettyformat(simplifyinfixops(tree, (b'or',)), (b'symbol',)) ... print(pycompat.sysstr(s)) >>> f((b'or', ... (b'or', ... (b'symbol', b'1'), ... (b'symbol', b'2')), ... (b'symbol', b'3'))) (or (symbol '1') (symbol '2') (symbol '3')) >>> f((b'func', ... (b'symbol', b'p1'), ... (b'or', ... (b'or', ... (b'func', ... (b'symbol', b'sort'), ... (b'list', ... (b'or', ... (b'or', ... (b'symbol', b'1'), ... (b'symbol', b'2')), ... (b'symbol', b'3')), ... (b'negate', ... (b'symbol', b'rev')))), ... (b'and', ... (b'symbol', b'4'), ... (b'group', ... (b'or', ... (b'or', ... (b'symbol', b'5'), ... (b'symbol', b'6')), ... (b'symbol', b'7'))))), ... (b'symbol', b'8')))) (func (symbol 'p1') (or (func (symbol 'sort') (list (or (symbol '1') (symbol '2') (symbol '3')) (negate (symbol 'rev')))) (and (symbol '4') (group (or (symbol '5') (symbol '6') (symbol '7')))) (symbol '8'))) """ if not isinstance(tree, tuple): return tree op = tree[0] if op not in targetnodes: return (op,) + tuple(simplifyinfixops(x, targetnodes) for x in tree[1:]) # walk down left nodes taking each right node. no recursion to left nodes # because infix operators are left-associative, i.e. left tree is deep. # e.g. '1 + 2 + 3' -> (+ (+ 1 2) 3) -> (+ 1 2 3) simplified = [] x = tree while x[0] == op: l, r = x[1:] simplified.append(simplifyinfixops(r, targetnodes)) x = l simplified.append(simplifyinfixops(x, targetnodes)) simplified.append(op) return tuple(reversed(simplified)) def _buildtree(template, placeholder, replstack): if template == placeholder: return replstack.pop() if not isinstance(template, tuple): return template return tuple(_buildtree(x, placeholder, replstack) for x in template) def buildtree(template, placeholder, *repls): """Create new tree by substituting placeholders by replacements >>> _ = (b'symbol', b'_') >>> def f(template, *repls): ... return buildtree(template, _, *repls) >>> f((b'func', (b'symbol', b'only'), (b'list', _, _)), ... ('symbol', '1'), ('symbol', '2')) ('func', ('symbol', 'only'), ('list', ('symbol', '1'), ('symbol', '2'))) >>> f((b'and', _, (b'not', _)), (b'symbol', b'1'), (b'symbol', b'2')) ('and', ('symbol', '1'), ('not', ('symbol', '2'))) """ if not isinstance(placeholder, tuple): raise error.ProgrammingError('placeholder must be a node tuple') replstack = list(reversed(repls)) r = _buildtree(template, placeholder, replstack) if replstack: raise error.ProgrammingError('too many replacements') return r def _matchtree(pattern, tree, placeholder, incompletenodes, matches): if pattern == tree: return True if not isinstance(pattern, tuple) or not isinstance(tree, tuple): return False if pattern == placeholder and tree[0] not in incompletenodes: matches.append(tree) return True if len(pattern) != len(tree): return False return all(_matchtree(p, x, placeholder, incompletenodes, matches) for p, x in zip(pattern, tree)) def matchtree(pattern, tree, placeholder=None, incompletenodes=()): """If a tree matches the pattern, return a list of the tree and nodes matched with the placeholder; Otherwise None >>> def f(pattern, tree): ... m = matchtree(pattern, tree, _, {b'keyvalue', b'list'}) ... if m: ... return m[1:] >>> _ = (b'symbol', b'_') >>> f((b'func', (b'symbol', b'ancestors'), _), ... (b'func', (b'symbol', b'ancestors'), (b'symbol', b'1'))) [('symbol', '1')] >>> f((b'func', (b'symbol', b'ancestors'), _), ... (b'func', (b'symbol', b'ancestors'), None)) >>> f((b'range', (b'dagrange', _, _), _), ... (b'range', ... (b'dagrange', (b'symbol', b'1'), (b'symbol', b'2')), ... (b'symbol', b'3'))) [('symbol', '1'), ('symbol', '2'), ('symbol', '3')] The placeholder does not match the specified incomplete nodes because an incomplete node (e.g. argument list) cannot construct an expression. >>> f((b'func', (b'symbol', b'ancestors'), _), ... (b'func', (b'symbol', b'ancestors'), ... (b'list', (b'symbol', b'1'), (b'symbol', b'2')))) The placeholder may be omitted, but which shouldn't match a None node. >>> _ = None >>> f((b'func', (b'symbol', b'ancestors'), None), ... (b'func', (b'symbol', b'ancestors'), (b'symbol', b'0'))) """ if placeholder is not None and not isinstance(placeholder, tuple): raise error.ProgrammingError('placeholder must be a node tuple') matches = [tree] if _matchtree(pattern, tree, placeholder, incompletenodes, matches): return matches def parseerrordetail(inst): """Compose error message from specified ParseError object """ if len(inst.args) > 1: return _('at %d: %s') % (inst.args[1], inst.args[0]) else: return inst.args[0] class alias(object): """Parsed result of alias""" def __init__(self, name, args, err, replacement): self.name = name self.args = args self.error = err self.replacement = replacement # whether own `error` information is already shown or not. # this avoids showing same warning multiple times at each # `expandaliases`. self.warned = False class basealiasrules(object): """Parsing and expansion rule set of aliases This is a helper for fileset/revset/template aliases. A concrete rule set should be made by sub-classing this and implementing class/static methods. It supports alias expansion of symbol and function-call styles:: # decl = defn h = heads(default) b($1) = ancestors($1) - ancestors(default) """ # typically a config section, which will be included in error messages _section = None # tag of symbol node _symbolnode = 'symbol' def __new__(cls): raise TypeError("'%s' is not instantiatable" % cls.__name__) @staticmethod def _parse(spec): """Parse an alias name, arguments and definition""" raise NotImplementedError @staticmethod def _trygetfunc(tree): """Return (name, args) if tree is a function; otherwise None""" raise NotImplementedError @classmethod def _builddecl(cls, decl): """Parse an alias declaration into ``(name, args, errorstr)`` This function analyzes the parsed tree. The parsing rule is provided by ``_parse()``. - ``name``: of declared alias (may be ``decl`` itself at error) - ``args``: list of argument names (or None for symbol declaration) - ``errorstr``: detail about detected error (or None) >>> sym = lambda x: (b'symbol', x) >>> symlist = lambda *xs: (b'list',) + tuple(sym(x) for x in xs) >>> func = lambda n, a: (b'func', sym(n), a) >>> parsemap = { ... b'foo': sym(b'foo'), ... b'$foo': sym(b'$foo'), ... b'foo::bar': (b'dagrange', sym(b'foo'), sym(b'bar')), ... b'foo()': func(b'foo', None), ... b'$foo()': func(b'$foo', None), ... b'foo($1, $2)': func(b'foo', symlist(b'$1', b'$2')), ... b'foo(bar_bar, baz.baz)': ... func(b'foo', symlist(b'bar_bar', b'baz.baz')), ... b'foo(bar($1, $2))': ... func(b'foo', func(b'bar', symlist(b'$1', b'$2'))), ... b'foo($1, $2, nested($1, $2))': ... func(b'foo', (symlist(b'$1', b'$2') + ... (func(b'nested', symlist(b'$1', b'$2')),))), ... b'foo("bar")': func(b'foo', (b'string', b'bar')), ... b'foo($1, $2': error.ParseError(b'unexpected token: end', 10), ... b'foo("bar': error.ParseError(b'unterminated string', 5), ... b'foo($1, $2, $1)': func(b'foo', symlist(b'$1', b'$2', b'$1')), ... } >>> def parse(expr): ... x = parsemap[expr] ... if isinstance(x, Exception): ... raise x ... return x >>> def trygetfunc(tree): ... if not tree or tree[0] != b'func' or tree[1][0] != b'symbol': ... return None ... if not tree[2]: ... return tree[1][1], [] ... if tree[2][0] == b'list': ... return tree[1][1], list(tree[2][1:]) ... return tree[1][1], [tree[2]] >>> class aliasrules(basealiasrules): ... _parse = staticmethod(parse) ... _trygetfunc = staticmethod(trygetfunc) >>> builddecl = aliasrules._builddecl >>> builddecl(b'foo') ('foo', None, None) >>> builddecl(b'$foo') ('$foo', None, "invalid symbol '$foo'") >>> builddecl(b'foo::bar') ('foo::bar', None, 'invalid format') >>> builddecl(b'foo()') ('foo', [], None) >>> builddecl(b'$foo()') ('$foo()', None, "invalid function '$foo'") >>> builddecl(b'foo($1, $2)') ('foo', ['$1', '$2'], None) >>> builddecl(b'foo(bar_bar, baz.baz)') ('foo', ['bar_bar', 'baz.baz'], None) >>> builddecl(b'foo($1, $2, nested($1, $2))') ('foo($1, $2, nested($1, $2))', None, 'invalid argument list') >>> builddecl(b'foo(bar($1, $2))') ('foo(bar($1, $2))', None, 'invalid argument list') >>> builddecl(b'foo("bar")') ('foo("bar")', None, 'invalid argument list') >>> builddecl(b'foo($1, $2') ('foo($1, $2', None, 'at 10: unexpected token: end') >>> builddecl(b'foo("bar') ('foo("bar', None, 'at 5: unterminated string') >>> builddecl(b'foo($1, $2, $1)') ('foo', None, 'argument names collide with each other') """ try: tree = cls._parse(decl) except error.ParseError as inst: return (decl, None, parseerrordetail(inst)) if tree[0] == cls._symbolnode: # "name = ...." style name = tree[1] if name.startswith('$'): return (decl, None, _("invalid symbol '%s'") % name) return (name, None, None) func = cls._trygetfunc(tree) if func: # "name(arg, ....) = ...." style name, args = func if name.startswith('$'): return (decl, None, _("invalid function '%s'") % name) if any(t[0] != cls._symbolnode for t in args): return (decl, None, _("invalid argument list")) if len(args) != len(set(args)): return (name, None, _("argument names collide with each other")) return (name, [t[1] for t in args], None) return (decl, None, _("invalid format")) @classmethod def _relabelargs(cls, tree, args): """Mark alias arguments as ``_aliasarg``""" if not isinstance(tree, tuple): return tree op = tree[0] if op != cls._symbolnode: return (op,) + tuple(cls._relabelargs(x, args) for x in tree[1:]) assert len(tree) == 2 sym = tree[1] if sym in args: op = '_aliasarg' elif sym.startswith('$'): raise error.ParseError(_("invalid symbol '%s'") % sym) return (op, sym) @classmethod def _builddefn(cls, defn, args): """Parse an alias definition into a tree and marks substitutions This function marks alias argument references as ``_aliasarg``. The parsing rule is provided by ``_parse()``. ``args`` is a list of alias argument names, or None if the alias is declared as a symbol. >>> from . import pycompat >>> parsemap = { ... b'$1 or foo': (b'or', (b'symbol', b'$1'), (b'symbol', b'foo')), ... b'$1 or $bar': ... (b'or', (b'symbol', b'$1'), (b'symbol', b'$bar')), ... b'$10 or baz': ... (b'or', (b'symbol', b'$10'), (b'symbol', b'baz')), ... b'"$1" or "foo"': ... (b'or', (b'string', b'$1'), (b'string', b'foo')), ... } >>> class aliasrules(basealiasrules): ... _parse = staticmethod(parsemap.__getitem__) ... _trygetfunc = staticmethod(lambda x: None) >>> builddefn = aliasrules._builddefn >>> def pprint(tree): ... s = prettyformat(tree, (b'_aliasarg', b'string', b'symbol')) ... print(pycompat.sysstr(s)) >>> args = [b'$1', b'$2', b'foo'] >>> pprint(builddefn(b'$1 or foo', args)) (or (_aliasarg '$1') (_aliasarg 'foo')) >>> try: ... builddefn(b'$1 or $bar', args) ... except error.ParseError as inst: ... print(pycompat.sysstr(parseerrordetail(inst))) invalid symbol '$bar' >>> args = [b'$1', b'$10', b'foo'] >>> pprint(builddefn(b'$10 or baz', args)) (or (_aliasarg '$10') (symbol 'baz')) >>> pprint(builddefn(b'"$1" or "foo"', args)) (or (string '$1') (string 'foo')) """ tree = cls._parse(defn) if args: args = set(args) else: args = set() return cls._relabelargs(tree, args) @classmethod def build(cls, decl, defn): """Parse an alias declaration and definition into an alias object""" repl = efmt = None name, args, err = cls._builddecl(decl) if err: efmt = _('bad declaration of %(section)s "%(name)s": %(error)s') else: try: repl = cls._builddefn(defn, args) except error.ParseError as inst: err = parseerrordetail(inst) efmt = _('bad definition of %(section)s "%(name)s": %(error)s') if err: err = efmt % {'section': cls._section, 'name': name, 'error': err} return alias(name, args, err, repl) @classmethod def buildmap(cls, items): """Parse a list of alias (name, replacement) pairs into a dict of alias objects""" aliases = {} for decl, defn in items: a = cls.build(decl, defn) aliases[a.name] = a return aliases @classmethod def _getalias(cls, aliases, tree): """If tree looks like an unexpanded alias, return (alias, pattern-args) pair. Return None otherwise. """ if not isinstance(tree, tuple): return None if tree[0] == cls._symbolnode: name = tree[1] a = aliases.get(name) if a and a.args is None: return a, None func = cls._trygetfunc(tree) if func: name, args = func a = aliases.get(name) if a and a.args is not None: return a, args return None @classmethod def _expandargs(cls, tree, args): """Replace _aliasarg instances with the substitution value of the same name in args, recursively. """ if not isinstance(tree, tuple): return tree if tree[0] == '_aliasarg': sym = tree[1] return args[sym] return tuple(cls._expandargs(t, args) for t in tree) @classmethod def _expand(cls, aliases, tree, expanding, cache): if not isinstance(tree, tuple): return tree r = cls._getalias(aliases, tree) if r is None: return tuple(cls._expand(aliases, t, expanding, cache) for t in tree) a, l = r if a.error: raise error.Abort(a.error) if a in expanding: raise error.ParseError(_('infinite expansion of %(section)s ' '"%(name)s" detected') % {'section': cls._section, 'name': a.name}) # get cacheable replacement tree by expanding aliases recursively expanding.append(a) if a.name not in cache: cache[a.name] = cls._expand(aliases, a.replacement, expanding, cache) result = cache[a.name] expanding.pop() if a.args is None: return result # substitute function arguments in replacement tree if len(l) != len(a.args): raise error.ParseError(_('invalid number of arguments: %d') % len(l)) l = [cls._expand(aliases, t, [], cache) for t in l] return cls._expandargs(result, dict(zip(a.args, l))) @classmethod def expand(cls, aliases, tree): """Expand aliases in tree, recursively. 'aliases' is a dictionary mapping user defined aliases to alias objects. """ return cls._expand(aliases, tree, [], {})