comparison contrib/shrink-revlog.py @ 10213:9e6848f352b0

contrib: use ui to write in shrink-revlog.py

10212:3ab391dd5ec5

import sys, os, tempfile
import optparse
from mercurial import ui as ui_, hg, revlog, transaction, node, util
from mercurial import changegroup

def toposort(rl):
    write = sys.stdout.write

    children = {}
    root = []
    # build children and roots
    write('reading %d revs ' % len(rl))
    try:
        for i in rl:
            children[i] = []
            parents = [p for p in rl.parentrevs(i) if p != node.nullrev]
            # in case of duplicate parents

            if len(parents) == 0:
                root.append(i)

            if i % 1000 == 0:
                write('.')
    finally:
        write('\n')

    # XXX this is a reimplementation of the 'branchsort' topo sort
    # algorithm in hgext.convert.convcmd... would be nice not to duplicate
    # the algorithm
    write('sorting ...')
    visit = root
    ret = []
    while visit:
        i = visit.pop(0)
        ret.append(i)

            parents_unseen = [p for p in rl.parentrevs(c)
                              if p != node.nullrev and p in children]
            if len(parents_unseen) == 0:
                next.append(c)
        visit = next + visit
    write('\n')
    return ret

def writerevs(r1, r2, order, tr):
    write = sys.stdout.write
    write('writing %d revs ' % len(order))

    count = [0]
    def progress(*args):
        if count[0] % 1000 == 0:
            write('.')
        count[0] += 1

    order = [r1.node(r) for r in order]

    # this is a bit ugly, but it works

    try:
        group = util.chunkbuffer(r1.group(order, lookup, progress))
        chunkiter = changegroup.chunkiter(group)
        r2.addgroup(chunkiter, unlookup, tr)
    finally:
        write('\n')

def report(olddatafn, newdatafn):
    oldsize = float(os.stat(olddatafn).st_size)
    newsize = float(os.stat(newdatafn).st_size)

    # argh: have to pass an int to %d, because a float >= 2^32
    # blows up under Python 2.5 or earlier
    sys.stdout.write('old file size: %12d bytes (%6.1f MiB)\n'
                     % (int(oldsize), oldsize/1024/1024))
    sys.stdout.write('new file size: %12d bytes (%6.1f MiB)\n'
                     % (int(newsize), newsize/1024/1024))

    shrink_percent = (oldsize - newsize) / oldsize * 100
    shrink_factor = oldsize / newsize
    sys.stdout.write('shrinkage: %.1f%% (%.1fx)\n'
                     % (shrink_percent, shrink_factor))

def main():

    # Unbuffer stdout for nice progress output.
    sys.stdout = os.fdopen(sys.stdout.fileno(), 'w', 0)
    write = sys.stdout.write

    parser = optparse.OptionParser(description=__doc__)
    parser.add_option('-R', '--repository',
                      default=os.path.curdir,
                      metavar='REPO',

                     '  %s\n'
                     'exists from a previous run; please clean up before '
                     'running again'
                     % (oldindexfn, olddatafn))

    write('shrinking %s\n' % indexfn)
    prefix = os.path.basename(indexfn)[:-1]
    (tmpfd, tmpindexfn) = tempfile.mkstemp(dir=os.path.dirname(indexfn),
                                           prefix=prefix,
                                           suffix='.i')
    tmpdatafn = tmpindexfn[:-2] + '.d'

                                 open,
                                 repo.sjoin('journal'))

    try:
        try:
            order = toposort(r1)
            writerevs(r1, r2, order, tr)
            report(datafn, tmpdatafn)
            tr.close()
        except:
            # Abort transaction first, so we truncate the files before
            # deleting them.
            tr.abort()

    os.link(indexfn, oldindexfn)
    os.link(datafn, olddatafn)
    os.rename(tmpindexfn, indexfn)
    os.rename(tmpdatafn, datafn)
    write('note: old revlog saved in:\n'
          '  %s\n'
          '  %s\n'
          '(You can delete those files when you are satisfied that your\n'
          'repository is still sane.  '
          'Running \'hg verify\' is strongly recommended.)\n'

10213:9e6848f352b0

import sys, os, tempfile
import optparse
from mercurial import ui as ui_, hg, revlog, transaction, node, util
from mercurial import changegroup

def toposort(ui, rl):

    children = {}
    root = []
    # build children and roots
    ui.write('reading %d revs ' % len(rl))
    try:
        for i in rl:
            children[i] = []
            parents = [p for p in rl.parentrevs(i) if p != node.nullrev]
            # in case of duplicate parents

            if len(parents) == 0:
                root.append(i)

            if i % 1000 == 0:
                ui.write('.')
    finally:
        ui.write('\n')

    # XXX this is a reimplementation of the 'branchsort' topo sort
    # algorithm in hgext.convert.convcmd... would be nice not to duplicate
    # the algorithm
    ui.write('sorting ...')
    visit = root
    ret = []
    while visit:
        i = visit.pop(0)
        ret.append(i)

            parents_unseen = [p for p in rl.parentrevs(c)
                              if p != node.nullrev and p in children]
            if len(parents_unseen) == 0:
                next.append(c)
        visit = next + visit
    ui.write('\n')
    return ret

def writerevs(ui, r1, r2, order, tr):

    ui.write('writing %d revs ' % len(order))
    count = [0]
    def progress(*args):
        if count[0] % 1000 == 0:
            ui.write('.')
        count[0] += 1

    order = [r1.node(r) for r in order]

    # this is a bit ugly, but it works

    try:
        group = util.chunkbuffer(r1.group(order, lookup, progress))
        chunkiter = changegroup.chunkiter(group)
        r2.addgroup(chunkiter, unlookup, tr)
    finally:
        ui.write('\n')

def report(ui, olddatafn, newdatafn):
    oldsize = float(os.stat(olddatafn).st_size)
    newsize = float(os.stat(newdatafn).st_size)

    # argh: have to pass an int to %d, because a float >= 2^32
    # blows up under Python 2.5 or earlier
    ui.write('old file size: %12d bytes (%6.1f MiB)\n'
             % (int(oldsize), oldsize/1024/1024))
    ui.write('new file size: %12d bytes (%6.1f MiB)\n'
             % (int(newsize), newsize/1024/1024))

    shrink_percent = (oldsize - newsize) / oldsize * 100
    shrink_factor = oldsize / newsize
    ui.write('shrinkage: %.1f%% (%.1fx)\n' % (shrink_percent, shrink_factor))

def main():

    # Unbuffer stdout for nice progress output.
    sys.stdout = os.fdopen(sys.stdout.fileno(), 'w', 0)

    parser = optparse.OptionParser(description=__doc__)
    parser.add_option('-R', '--repository',
                      default=os.path.curdir,
                      metavar='REPO',

                     '  %s\n'
                     'exists from a previous run; please clean up before '
                     'running again'
                     % (oldindexfn, olddatafn))

    ui.write('shrinking %s\n' % indexfn)
    prefix = os.path.basename(indexfn)[:-1]
    (tmpfd, tmpindexfn) = tempfile.mkstemp(dir=os.path.dirname(indexfn),
                                           prefix=prefix,
                                           suffix='.i')
    tmpdatafn = tmpindexfn[:-2] + '.d'

                                 open,
                                 repo.sjoin('journal'))

    try:
        try:
            order = toposort(ui, r1)
            writerevs(ui, r1, r2, order, tr)
            report(ui, datafn, tmpdatafn)
            tr.close()
        except:
            # Abort transaction first, so we truncate the files before
            # deleting them.
            tr.abort()

    os.link(indexfn, oldindexfn)
    os.link(datafn, olddatafn)
    os.rename(tmpindexfn, indexfn)
    os.rename(tmpdatafn, datafn)
    ui.write('note: old revlog saved in:\n'
          '  %s\n'
          '  %s\n'
          '(You can delete those files when you are satisfied that your\n'
          'repository is still sane.  '
          'Running \'hg verify\' is strongly recommended.)\n'