# revset.py - revision set queries 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.

from __future__ import absolute_import

import heapq

import re

from .i18n import _

from . import (

    destutil,

    encoding,

    error,

    hbisect,

    match as matchmod,

    node,

    obsolete as obsmod,

    pathutil,

    phases,

    registrar,

    repoview,

    revsetlang,

    smartset,

    util,

)

# helpers for processing parsed tree

getsymbol = revsetlang.getsymbol

getstring = revsetlang.getstring

getinteger = revsetlang.getinteger

getboolean = revsetlang.getboolean

getlist = revsetlang.getlist

getrange = revsetlang.getrange

getargs = revsetlang.getargs

getargsdict = revsetlang.getargsdict

# constants used as an argument of match() and matchany()

anyorder = revsetlang.anyorder

defineorder = revsetlang.defineorder

followorder = revsetlang.followorder

baseset = smartset.baseset

generatorset = smartset.generatorset

spanset = smartset.spanset

fullreposet = smartset.fullreposet

def _revancestors(repo, revs, followfirst):

    """Like revlog.ancestors(), but supports followfirst."""

    if followfirst:

        cut = 1

    else:

        cut = None

    cl = repo.changelog

    def iterate():

        revs.sort(reverse=True)

        irevs = iter(revs)

        h = []

        inputrev = next(irevs, None)

        if inputrev is not None:

            heapq.heappush(h, -inputrev)

        seen = set()

        while h:

            current = -heapq.heappop(h)

            if current == inputrev:

                inputrev = next(irevs, None)

                if inputrev is not None:

                    heapq.heappush(h, -inputrev)

            if current not in seen:

                seen.add(current)

                yield current

                for parent in cl.parentrevs(current)[:cut]:

                    if parent != node.nullrev:

                        heapq.heappush(h, -parent)

    return generatorset(iterate(), iterasc=False)

def _revdescendants(repo, revs, followfirst):

    """Like revlog.descendants() but supports followfirst."""

    if followfirst:

        cut = 1

    else:

        cut = None

    def iterate():

        cl = repo.changelog

        # XXX this should be 'parentset.min()' assuming 'parentset' is a

        # smartset (and if it is not, it should.)

        first = min(revs)

        nullrev = node.nullrev

        if first == nullrev:

            # Are there nodes with a null first parent and a non-null

            # second one? Maybe. Do we care? Probably not.

            for i in cl:

                yield i

        else:

            seen = set(revs)

            for i in cl.revs(first + 1):

                for x in cl.parentrevs(i)[:cut]:

                    if x != nullrev and x in seen:

                        seen.add(i)

                        yield i

                        break

    return generatorset(iterate(), iterasc=True)

def _reachablerootspure(repo, minroot, roots, heads, includepath):

    """return (heads(::<roots> and ::<heads>))

    If includepath is True, return (<roots>::<heads>)."""

    if not roots:

        return []

    parentrevs = repo.changelog.parentrevs

    roots = set(roots)

    visit = list(heads)

    reachable = set()

    seen = {}

    # prefetch all the things! (because python is slow)

    reached = reachable.add

    dovisit = visit.append

    nextvisit = visit.pop

    # open-code the post-order traversal due to the tiny size of

    # sys.getrecursionlimit()

    while visit:

        rev = nextvisit()

        if rev in roots:

            reached(rev)

            if not includepath:

                continue

        parents = parentrevs(rev)

        seen[rev] = parents

        for parent in parents:

            if parent >= minroot and parent not in seen:

                dovisit(parent)

    if not reachable:

        return baseset()

    if not includepath:

        return reachable

    for rev in sorted(seen):

        for parent in seen[rev]:

            if parent in reachable:

                reached(rev)

    return reachable

def reachableroots(repo, roots, heads, includepath=False):

    """return (heads(::<roots> and ::<heads>))

    If includepath is True, return (<roots>::<heads>)."""

    if not roots:

        return baseset()

    minroot = roots.min()

    roots = list(roots)

    heads = list(heads)

    try:

        revs = repo.changelog.reachableroots(minroot, heads, roots, includepath)

    except AttributeError:

        revs = _reachablerootspure(repo, minroot, roots, heads, includepath)

    revs = baseset(revs)

    revs.sort()

    return revs

# helpers

def getset(repo, subset, x):

    if not x:

        raise error.ParseError(_("missing argument"))

    return methods[x[0]](repo, subset, *x[1:])

def _getrevsource(repo, r):

    extra = repo[r].extra()

    for label in ('source', 'transplant_source', 'rebase_source'):

        if label in extra:

            try:

                return repo[extra[label]].rev()

            except error.RepoLookupError:

                pass

    return None

# operator methods

def stringset(repo, subset, x):

    x = repo[x].rev()

    if (x in subset

        or x == node.nullrev and isinstance(subset, fullreposet)):

        return baseset([x])

    return baseset()

def rangeset(repo, subset, x, y, order):

    m = getset(repo, fullreposet(repo), x)

    n = getset(repo, fullreposet(repo), y)

    if not m or not n:

        return baseset()

    return _makerangeset(repo, subset, m.first(), n.last(), order)

def rangeall(repo, subset, x, order):

    assert x is None

    return _makerangeset(repo, subset, 0, len(repo) - 1, order)

def rangepre(repo, subset, y, order):

    # ':y' can't be rewritten to '0:y' since '0' may be hidden

    n = getset(repo, fullreposet(repo), y)

    if not n:

        return baseset()

    return _makerangeset(repo, subset, 0, n.last(), order)

def rangepost(repo, subset, x, order):

    m = getset(repo, fullreposet(repo), x)

    if not m:

        return baseset()

    return _makerangeset(repo, subset, m.first(), len(repo) - 1, order)

def _makerangeset(repo, subset, m, n, order):

    if m == n:

        r = baseset([m])

    elif n == node.wdirrev:

        r = spanset(repo, m, len(repo)) + baseset([n])

    elif m == node.wdirrev:

        r = baseset([m]) + spanset(repo, len(repo) - 1, n - 1)

    elif m < n:

        r = spanset(repo, m, n + 1)

    else:

        r = spanset(repo, m, n - 1)

    if order == defineorder:

        return r & subset

    else:

        # carrying the sorting over when possible would be more efficient

        return subset & r

def dagrange(repo, subset, x, y, order):

    r = fullreposet(repo)

    xs = reachableroots(repo, getset(repo, r, x), getset(repo, r, y),

                         includepath=True)

    return subset & xs

def andset(repo, subset, x, y, order):

    return getset(repo, getset(repo, subset, x), y)

def differenceset(repo, subset, x, y, order):

    return getset(repo, subset, x) - getset(repo, subset, y)

def _orsetlist(repo, subset, xs):

    assert xs

    if len(xs) == 1:

        return getset(repo, subset, xs[0])

    p = len(xs) // 2

    a = _orsetlist(repo, subset, xs[:p])

    b = _orsetlist(repo, subset, xs[p:])

    return a + b

def orset(repo, subset, x, order):

    xs = getlist(x)

    if order == followorder:

        # slow path to take the subset order

        return subset & _orsetlist(repo, fullreposet(repo), xs)

    else:

        return _orsetlist(repo, subset, xs)

def notset(repo, subset, x, order):

    return subset - getset(repo, subset, x)

def listset(repo, subset, *xs):

    raise error.ParseError(_("can't use a list in this context"),

                           hint=_('see hg help "revsets.x or y"'))

def keyvaluepair(repo, subset, k, v):

    raise error.ParseError(_("can't use a key-value pair in this context"))

def func(repo, subset, a, b, order):

    f = getsymbol(a)

    if f in symbols:

        func = symbols[f]

        if getattr(func, '_takeorder', False):

            return func(repo, subset, b, order)

        return func(repo, subset, b)

    keep = lambda fn: getattr(fn, '__doc__', None) is not None

    syms = [s for (s, fn) in symbols.items() if keep(fn)]

    raise error.UnknownIdentifier(f, syms)

# functions

# symbols are callables like:

#   fn(repo, subset, x)

# with:

#   repo - current repository instance

#   subset - of revisions to be examined

#   x - argument in tree form

symbols = {}

# symbols which can't be used for a DoS attack for any given input

# (e.g. those which accept regexes as plain strings shouldn't be included)

# functions that just return a lot of changesets (like all) don't count here

safesymbols = set()

predicate = registrar.revsetpredicate()

@predicate('_destupdate')

def _destupdate(repo, subset, x):

    # experimental revset for update destination

    args = getargsdict(x, 'limit', 'clean')

    return subset & baseset([destutil.destupdate(repo, **args)[0]])

@predicate('_destmerge')

def _destmerge(repo, subset, x):

    # experimental revset for merge destination

    sourceset = None

    if x is not None:

        sourceset = getset(repo, fullreposet(repo), x)

    return subset & baseset([destutil.destmerge(repo, sourceset=sourceset)])

@predicate('adds(pattern)', safe=True)

def adds(repo, subset, x):

    """Changesets that add a file matching pattern.

    The pattern without explicit kind like ``glob:`` is expected to be

    relative to the current directory and match against a file or a

    directory.

    """

    # i18n: "adds" is a keyword

    pat = getstring(x, _("adds requires a pattern"))

    return checkstatus(repo, subset, pat, 1)

@predicate('ancestor(*changeset)', safe=True)

def ancestor(repo, subset, x):

    """A greatest common ancestor of the changesets.

    Accepts 0 or more changesets.

    Will return empty list when passed no args.

    Greatest common ancestor of a single changeset is that changeset.

    """

    # i18n: "ancestor" is a keyword

    l = getlist(x)

    rl = fullreposet(repo)

    anc = None

    # (getset(repo, rl, i) for i in l) generates a list of lists

    for revs in (getset(repo, rl, i) for i in l):

        for r in revs:

            if anc is None:

                anc = repo[r]

            else:

                anc = anc.ancestor(repo[r])

    if anc is not None and anc.rev() in subset:

        return baseset([anc.rev()])

    return baseset()

def _ancestors(repo, subset, x, followfirst=False):

    heads = getset(repo, fullreposet(repo), x)

    if not heads:

        return baseset()

    s = _revancestors(repo, heads, followfirst)

    return subset & s

@predicate('ancestors(set)', safe=True)

def ancestors(repo, subset, x):

    """Changesets that are ancestors of a changeset in set.

    """

    return _ancestors(repo, subset, x)

@predicate('_firstancestors', safe=True)

def _firstancestors(repo, subset, x):