#

# Copyright 2005-2007 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 struct

# import stuff from node for others to import from revlog

    nullrev,

)

    REVIDX_ISCENSORED,

    REVIDX_RAWTEXT_CHANGING_FLAGS,

)

    attr,

)

    error,

    mdiff,

)

RevlogError = error.RevlogError

CensoredNodeError = error.CensoredNodeError

class _testrevlog(object):

    """minimalist fake revlog to use in doctests"""

    def __init__(self, data, density=0.5, mingap=0):

        """data is an list of revision payload boundaries"""

        self._data = data

        self._srdensitythreshold = density

        self._srmingapsize = mingap

    def start(self, rev):

        if rev == 0:

            return 0

        return self._data[rev - 1]

    def end(self, rev):

        return self._data[rev]

    def length(self, rev):

        return self.end(rev) - self.start(rev)

    def __len__(self):

        return len(self._data)

    """slice revs to reduce the amount of unrelated data to be read from disk.

    ``revs`` is sliced into groups that should be read in one time.

    Assume that revs are sorted.

    The initial chunk is sliced until the overall density (payload/chunks-span

    ratio) is above `revlog._srdensitythreshold`. No gap smaller than

    `revlog._srmingapsize` is skipped.

    If `targetsize` is set, no chunk larger than `targetsize` will be yield.

    For consistency with other slicing choice, this limit won't go lower than

    `revlog._srmingapsize`.

    If individual revisions chunk are larger than this limit, they will still

    be raised individually.

    >>> revlog = _testrevlog([

    ...  5,  #00 (5)

    ...  10, #01 (5)

    ...  12, #02 (2)

    ...  12, #03 (empty)

    ...  27, #04 (15)

    ...  31, #05 (4)

    ...  31, #06 (empty)

    ...  42, #07 (11)

    ...  47, #08 (5)

    ...  47, #09 (empty)

    ...  48, #10 (1)

    ...  51, #11 (3)

    ...  74, #12 (23)

    ...  85, #13 (11)

    ...  86, #14 (1)

    ...  91, #15 (5)

    ... ])

    [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]]

    [[0], [15]]

    [[0], [11], [15]]

    [[0], [11, 13, 15]]

    [[1, 2], [5, 8, 10, 11], [14]]

    Slicing with a maximum chunk size

    [[0], [11], [13], [15]]

    [[0], [11], [13, 15]]

    """

    if targetsize is not None:

        targetsize = max(targetsize, revlog._srmingapsize)

    # targetsize should not be specified when evaluating delta candidates:

    # * targetsize is used to ensure we stay within specification when reading,

    # * deltainfo is used to pick are good delta chain when writing.

    if not (deltainfo is None or targetsize is None):

        msg = 'cannot use `targetsize` with a `deltainfo`'

        raise error.ProgrammingError(msg)

    for chunk in _slicechunktodensity(revlog, revs,

                                      deltainfo,

                                      revlog._srdensitythreshold,

                                      revlog._srmingapsize):

        for subchunk in _slicechunktosize(revlog, chunk, targetsize):

            yield subchunk

def _slicechunktosize(revlog, revs, targetsize=None):

    """slice revs to match the target size

    This is intended to be used on chunk that density slicing selected by that

    are still too large compared to the read garantee of revlog. This might

    happens when "minimal gap size" interrupted the slicing or when chain are

    built in a way that create large blocks next to each other.

    >>> revlog = _testrevlog([

    ...  3,  #0 (3)

    ...  5,  #1 (2)

    ...  6,  #2 (1)

    ...  8,  #3 (2)

    ...  8,  #4 (empty)

    ...  11, #5 (3)

    ...  12, #6 (1)

    ...  13, #7 (1)

    ...  14, #8 (1)

    ... ])

    Cases where chunk is already small enough

    >>> list(_slicechunktosize(revlog, [0], 3))

    [[0]]

    >>> list(_slicechunktosize(revlog, [6, 7], 3))

    [[6, 7]]

    >>> list(_slicechunktosize(revlog, [0], None))

    [[0]]

    >>> list(_slicechunktosize(revlog, [6, 7], None))

    [[6, 7]]

    cases where we need actual slicing

    >>> list(_slicechunktosize(revlog, [0, 1], 3))

    [[0], [1]]

    >>> list(_slicechunktosize(revlog, [1, 3], 3))

    [[1], [3]]

    >>> list(_slicechunktosize(revlog, [1, 2, 3], 3))

    [[1, 2], [3]]

    >>> list(_slicechunktosize(revlog, [3, 5], 3))

    [[3], [5]]

    >>> list(_slicechunktosize(revlog, [3, 4, 5], 3))

    [[3], [5]]

    >>> list(_slicechunktosize(revlog, [5, 6, 7, 8], 3))

    [[5], [6, 7, 8]]

    >>> list(_slicechunktosize(revlog, [0, 1, 2, 3, 4, 5, 6, 7, 8], 3))

    [[0], [1, 2], [3], [5], [6, 7, 8]]

    Case with too large individual chunk (must return valid chunk)

    >>> list(_slicechunktosize(revlog, [0, 1], 2))

    [[0], [1]]

    >>> list(_slicechunktosize(revlog, [1, 3], 1))

    [[1], [3]]

    >>> list(_slicechunktosize(revlog, [3, 4, 5], 2))

    [[3], [5]]

    """

    assert targetsize is None or 0 <= targetsize

        yield revs

        return

    startrevidx = 0

    startdata = revlog.start(revs[0])

    endrevidx = 0

    iterrevs = enumerate(revs)

    next(iterrevs) # skip first rev.

    for idx, r in iterrevs:

        span = revlog.end(r) - startdata

        if span <= targetsize:

            endrevidx = idx

        else:

            chunk = _trimchunk(revlog, revs, startrevidx, endrevidx + 1)

            if chunk:

                yield chunk

            startrevidx = idx

            startdata = revlog.start(r)

            endrevidx = idx

    yield _trimchunk(revlog, revs, startrevidx)

def _slicechunktodensity(revlog, revs, deltainfo=None, targetdensity=0.5,

                         mingapsize=0):

    """slice revs to reduce the amount of unrelated data to be read from disk.

    ``revs`` is sliced into groups that should be read in one time.

    Assume that revs are sorted.

    ``deltainfo`` is a _deltainfo instance of a revision that we would append

    to the top of the revlog.

    The initial chunk is sliced until the overall density (payload/chunks-span

    ratio) is above `targetdensity`. No gap smaller than `mingapsize` is

    skipped.

    >>> revlog = _testrevlog([

    ...  5,  #00 (5)

    ...  10, #01 (5)

    ...  12, #02 (2)

    ...  12, #03 (empty)

    ...  27, #04 (15)

    ...  31, #05 (4)

    ...  31, #06 (empty)

    ...  42, #07 (11)

    ...  47, #08 (5)

    ...  47, #09 (empty)

    ...  48, #10 (1)

    ...  51, #11 (3)

    ...  74, #12 (23)

    ...  85, #13 (11)

    ...  86, #14 (1)

    ...  91, #15 (5)

    ... ])

    >>> list(_slicechunktodensity(revlog, list(range(16))))

    [[0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]]

    >>> list(_slicechunktodensity(revlog, [0, 15]))

    [[0], [15]]

    >>> list(_slicechunktodensity(revlog, [0, 11, 15]))

    [[0], [11], [15]]

    >>> list(_slicechunktodensity(revlog, [0, 11, 13, 15]))

    [[0], [11, 13, 15]]

    >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14]))

    [[1, 2], [5, 8, 10, 11], [14]]

    >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14],

    ...                           mingapsize=20))

    [[1, 2, 3, 5, 8, 10, 11], [14]]

    >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14],

    ...                           targetdensity=0.95))

    [[1, 2], [5], [8, 10, 11], [14]]

    >>> list(_slicechunktodensity(revlog, [1, 2, 3, 5, 8, 10, 11, 14],

    ...                           targetdensity=0.95, mingapsize=12))

    [[1, 2], [5, 8, 10, 11], [14]]

    """

    start = revlog.start

    length = revlog.length

    if len(revs) <= 1:

        yield revs

        return

    nextrev = len(revlog)

    nextoffset = revlog.end(nextrev - 1)

    if deltainfo is None:

        chainpayload = sum(length(r) for r in revs)

    else:

        deltachainspan = deltainfo.distance

        chainpayload = deltainfo.compresseddeltalen

    if deltachainspan < mingapsize:

        yield revs

        return

    readdata = deltachainspan

    if deltachainspan:

        density = chainpayload / float(deltachainspan)

    else:

        density = 1.0

    if density >= targetdensity:

        yield revs

        return

    if deltainfo is not None and deltainfo.deltalen:

        revs = list(revs)

        revs.append(nextrev)

    # Store the gaps in a heap to have them sorted by decreasing size

    gapsheap = []

    heapq.heapify(gapsheap)

    prevend = None

    for i, rev in enumerate(revs):

        if rev < nextrev:

            revstart = start(rev)

            revlen = length(rev)

        else:

            revstart = nextoffset

            revlen = deltainfo.deltalen

        # Skip empty revisions to form larger holes

        if revlen == 0:

            continue

        if prevend is not None:

            gapsize = revstart - prevend

            # only consider holes that are large enough

            if gapsize > mingapsize:

                heapq.heappush(gapsheap, (-gapsize, i))

        prevend = revstart + revlen

    # Collect the indices of the largest holes until the density is acceptable

    indicesheap = []

    heapq.heapify(indicesheap)

    while gapsheap and density < targetdensity:

        oppgapsize, gapidx = heapq.heappop(gapsheap)

        heapq.heappush(indicesheap, gapidx)

        # the gap sizes are stored as negatives to be sorted decreasingly

        # by the heap

        readdata -= (-oppgapsize)

        if readdata > 0:

            density = chainpayload / float(readdata)

        else:

            density = 1.0

    # Cut the revs at collected indices

    previdx = 0

    while indicesheap:

        idx = heapq.heappop(indicesheap)

        chunk = _trimchunk(revlog, revs, previdx, idx)

        if chunk:

            yield chunk

        previdx = idx

    chunk = _trimchunk(revlog, revs, previdx)

    if chunk:

        yield chunk