tests/hghave.py

changelog: disable delta chains

This patch disables delta chains on changelogs. After this patch, new
entries on changelogs - including existing changelogs - will be stored
as the fulltext of that data (likely compressed). No delta computation
will be performed.

An overview of delta chains and data justifying this change follows.

Revlogs try to store entries as a delta against a previous entry (either
a parent revision in the case of generaldelta or the previous physical
revision when not using generaldelta). Most of the time this is the
correct thing to do: it frequently results in less CPU usage and smaller
storage.

Delta chains are most effective when the base revision being deltad
against is similar to the current data. This tends to occur naturally
for manifests and file data, since only small parts of each tend to
change with each revision. Changelogs, however, are a different story.

Changelog entries represent changesets/commits. And unless commits in a
repository are homogonous (same author, changing same files, similar
commit messages, etc), a delta from one entry to the next tends to be
relatively large compared to the size of the entry. This means that
delta chains tend to be short. How short? Here is the full vs delta
revision breakdown on some real world repos:

Repo             % Full    % Delta   Max Length
hg                45.8       54.2        6
mozilla-central   42.4       57.6        8
mozilla-unified   42.5       57.5       17
pypy              46.1       53.9        6
python-zstandard  46.1       53.9        3

(I threw in python-zstandard as an example of a repo that is homogonous.
It contains a small Python project with changes all from the same
author.)

Contrast this with the manifest revlog for these repos, where 99+% of
revisions are deltas and delta chains run into the thousands.

So delta chains aren't as useful on changelogs. But even a short delta
chain may provide benefits. Let's measure that.

Delta chains may require less CPU to read revisions if the CPU time
spent reading smaller deltas is less than the CPU time used to
decompress larger individual entries. We can measure this via
`hg perfrevlog -c -d 1` to iterate a revlog to resolve each revision's
fulltext. Here are the results of that command on a repo using delta
chains in its changelog and on a repo without delta chains:

hg (forward)
! wall 0.407008 comb 0.410000 user 0.410000 sys 0.000000 (best of 25)
! wall 0.390061 comb 0.390000 user 0.390000 sys 0.000000 (best of 26)

hg (reverse)
! wall 0.515221 comb 0.520000 user 0.520000 sys 0.000000 (best of 19)
! wall 0.400018 comb 0.400000 user 0.390000 sys 0.010000 (best of 25)

mozilla-central (forward)
! wall 4.508296 comb 4.490000 user 4.490000 sys 0.000000 (best of 3)
! wall 4.370222 comb 4.370000 user 4.350000 sys 0.020000 (best of 3)

mozilla-central (reverse)
! wall 5.758995 comb 5.760000 user 5.720000 sys 0.040000 (best of 3)
! wall 4.346503 comb 4.340000 user 4.320000 sys 0.020000 (best of 3)

mozilla-unified (forward)
! wall 4.957088 comb 4.950000 user 4.940000 sys 0.010000 (best of 3)
! wall 4.660528 comb 4.650000 user 4.630000 sys 0.020000 (best of 3)

mozilla-unified (reverse)
! wall 6.119827 comb 6.110000 user 6.090000 sys 0.020000 (best of 3)
! wall 4.675136 comb 4.670000 user 4.670000 sys 0.000000 (best of 3)

pypy (forward)
! wall 1.231122 comb 1.240000 user 1.230000 sys 0.010000 (best of 8)
! wall 1.164896 comb 1.160000 user 1.160000 sys 0.000000 (best of 9)

pypy (reverse)
! wall 1.467049 comb 1.460000 user 1.460000 sys 0.000000 (best of 7)
! wall 1.160200 comb 1.170000 user 1.160000 sys 0.010000 (best of 9)

The data clearly shows that it takes less wall and CPU time to resolve
revisions when there are no delta chains in the changelogs, regardless
of the direction of traversal. Furthermore, not using a delta chain
means that fulltext resolution in reverse is as fast as iterating
forward. So not using delta chains on the changelog is a clear CPU win
for reading operations.

An example of a user-visible operation showing this speed-up is revset
evaluation. Here are results for
`hg perfrevset 'author(gps) or author(mpm)'`:

hg
! wall 1.655506 comb 1.660000 user 1.650000 sys 0.010000 (best of 6)
! wall 1.612723 comb 1.610000 user 1.600000 sys 0.010000 (best of 7)

mozilla-central
! wall 17.629826 comb 17.640000 user 17.600000 sys 0.040000 (best of 3)
! wall 17.311033 comb 17.300000 user 17.260000 sys 0.040000 (best of 3)

What about 00changelog.i size?

Repo                Delta Chains     No Delta Chains
hg                    7,033,250         6,976,771
mozilla-central      82,978,748        81,574,623
mozilla-unified      88,112,349        86,702,162
pypy                 20,740,699        20,659,741

The data shows that removing delta chains from the changelog makes the
changelog smaller.

Delta chains are also used during changegroup generation. This
operation essentially converts a series of revisions to one large
delta chain. And changegroup generation is smart: if the delta in
the revlog matches what the changegroup is emitting, it will reuse
the delta instead of recalculating it. We can measure the impact
removing changelog delta chains has on changegroup generation via
`hg perfchangegroupchangelog`:

hg
! wall 1.589245 comb 1.590000 user 1.590000 sys 0.000000 (best of 7)
! wall 1.788060 comb 1.790000 user 1.790000 sys 0.000000 (best of 6)

mozilla-central
! wall 17.382585 comb 17.380000 user 17.340000 sys 0.040000 (best of 3)
! wall 20.161357 comb 20.160000 user 20.120000 sys 0.040000 (best of 3)

mozilla-unified
! wall 18.722839 comb 18.720000 user 18.680000 sys 0.040000 (best of 3)
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)

pypy
! wall 4.828317 comb 4.830000 user 4.820000 sys 0.010000 (best of 3)
! wall 5.415455 comb 5.420000 user 5.410000 sys 0.010000 (best of 3)

The data shows eliminating delta chains makes the changelog part of
changegroup generation slower. This is expected since we now have to
compute deltas for revisions where we could recycle the delta before.

It is worth putting this regression into context of overall changegroup
times. Here is the rough total CPU time spent in changegroup generation
for various repos while using delta chains on the changelog:

Repo              CPU Time (s)    CPU Time w/ compression
hg                  4.50              7.05
mozilla-central   111.1             222.0
pypy               28.68             75.5

Before compression, removing delta chains from the changegroup adds
~4.4% overhead to hg changegroup generation, 1.3% to mozilla-central,
and 2.0% to pypy. When you factor in zlib compression, these percentages
are roughly divided by 2.

While the increased CPU usage for changegroup generation is unfortunate,
I think it is acceptable because the percentage is small, server
operators (those likely impacted most by this) have other mechanisms
to mitigate CPU consumption (namely reducing zlib compression level and
pre-generated clone bundles), and because there is room to optimize this
in the future. For example, we could use the nullid as the base revision,
effectively encoding the full revision for each entry in the changegroup.
When doing this, `hg perfchangegroupchangelog` nearly halves:

mozilla-unified
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
! wall 11.196461 comb 11.200000 user 11.190000 sys 0.010000 (best of 3)

This looks very promising as a future optimization opportunity.

It's worth that the changes in test-acl.t to the changegroup part size.
This is because revision 6 in the changegroup had a delta chain of
length 2 before and after this patch the base revision is nullrev.
When the base revision is nullrev, cg2packer.deltaparent() hardcodes
the *previous* revision from the changegroup as the delta parent.
This caused the delta in the changegroup to switch base revisions,
the delta to change, and the size to change accordingly. While the
size increased in this case, I think sizes will remain the same
on average, as the delta base for changelog revisions doesn't matter
too much (as this patch shows). So, I don't consider this a regression.

from __future__ import absolute_import

import errno
import os
import re
import socket
import stat
import subprocess
import sys
import tempfile

tempprefix = 'hg-hghave-'

checks = {
    "true": (lambda: True, "yak shaving"),
    "false": (lambda: False, "nail clipper"),
}

def check(name, desc):
    """Registers a check function for a feature."""
    def decorator(func):
        checks[name] = (func, desc)
        return func
    return decorator

def checkvers(name, desc, vers):
    """Registers a check function for each of a series of versions.

    vers can be a list or an iterator"""
    def decorator(func):
        def funcv(v):
            def f():
                return func(v)
            return f
        for v in vers:
            v = str(v)
            f = funcv(v)
            checks['%s%s' % (name, v.replace('.', ''))] = (f, desc % v)
        return func
    return decorator

def checkfeatures(features):
    result = {
        'error': [],
        'missing': [],
        'skipped': [],
    }

    for feature in features:
        negate = feature.startswith('no-')
        if negate:
            feature = feature[3:]

        if feature not in checks:
            result['missing'].append(feature)
            continue

        check, desc = checks[feature]
        try:
            available = check()
        except Exception:
            result['error'].append('hghave check failed: %s' % feature)
            continue

        if not negate and not available:
            result['skipped'].append('missing feature: %s' % desc)
        elif negate and available:
            result['skipped'].append('system supports %s' % desc)

    return result

def require(features):
    """Require that features are available, exiting if not."""
    result = checkfeatures(features)

    for missing in result['missing']:
        sys.stderr.write('skipped: unknown feature: %s\n' % missing)
    for msg in result['skipped']:
        sys.stderr.write('skipped: %s\n' % msg)
    for msg in result['error']:
        sys.stderr.write('%s\n' % msg)

    if result['missing']:
        sys.exit(2)

    if result['skipped'] or result['error']:
        sys.exit(1)

def matchoutput(cmd, regexp, ignorestatus=False):
    """Return the match object if cmd executes successfully and its output
    is matched by the supplied regular expression.
    """
    r = re.compile(regexp)
    try:
        p = subprocess.Popen(
            cmd, shell=True, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
    except OSError as e:
        if e.errno != errno.ENOENT:
            raise
        ret = -1
    ret = p.wait()
    s = p.stdout.read()
    return (ignorestatus or not ret) and r.search(s)

@check("baz", "GNU Arch baz client")
def has_baz():
    return matchoutput('baz --version 2>&1', br'baz Bazaar version')

@check("bzr", "Canonical's Bazaar client")
def has_bzr():
    try:
        import bzrlib
        import bzrlib.bzrdir
        import bzrlib.errors
        import bzrlib.revision
        import bzrlib.revisionspec
        bzrlib.revisionspec.RevisionSpec
        return bzrlib.__doc__ is not None
    except (AttributeError, ImportError):
        return False

@checkvers("bzr", "Canonical's Bazaar client >= %s", (1.14,))
def has_bzr_range(v):
    major, minor = v.split('.')[0:2]
    try:
        import bzrlib
        return (bzrlib.__doc__ is not None
                and bzrlib.version_info[:2] >= (int(major), int(minor)))
    except ImportError:
        return False

@check("chg", "running with chg")
def has_chg():
    return 'CHGHG' in os.environ

@check("cvs", "cvs client/server")
def has_cvs():
    re = br'Concurrent Versions System.*?server'
    return matchoutput('cvs --version 2>&1', re) and not has_msys()

@check("cvs112", "cvs client/server 1.12.* (not cvsnt)")
def has_cvs112():
    re = br'Concurrent Versions System \(CVS\) 1.12.*?server'
    return matchoutput('cvs --version 2>&1', re) and not has_msys()

@check("cvsnt", "cvsnt client/server")
def has_cvsnt():
    re = br'Concurrent Versions System \(CVSNT\) (\d+).(\d+).*\(client/server\)'
    return matchoutput('cvsnt --version 2>&1', re)

@check("darcs", "darcs client")
def has_darcs():
    return matchoutput('darcs --version', br'2\.[2-9]', True)

@check("mtn", "monotone client (>= 1.0)")
def has_mtn():
    return matchoutput('mtn --version', br'monotone', True) and not matchoutput(
        'mtn --version', br'monotone 0\.', True)

@check("eol-in-paths", "end-of-lines in paths")
def has_eol_in_paths():
    try:
        fd, path = tempfile.mkstemp(dir='.', prefix=tempprefix, suffix='\n\r')
        os.close(fd)
        os.remove(path)
        return True
    except (IOError, OSError):
        return False

@check("execbit", "executable bit")
def has_executablebit():
    try:
        EXECFLAGS = stat.S_IXUSR | stat.S_IXGRP | stat.S_IXOTH
        fh, fn = tempfile.mkstemp(dir='.', prefix=tempprefix)
        try:
            os.close(fh)
            m = os.stat(fn).st_mode & 0o777
            new_file_has_exec = m & EXECFLAGS
            os.chmod(fn, m ^ EXECFLAGS)
            exec_flags_cannot_flip = ((os.stat(fn).st_mode & 0o777) == m)
        finally:
            os.unlink(fn)
    except (IOError, OSError):
        # we don't care, the user probably won't be able to commit anyway
        return False
    return not (new_file_has_exec or exec_flags_cannot_flip)

@check("icasefs", "case insensitive file system")
def has_icasefs():
    # Stolen from mercurial.util
    fd, path = tempfile.mkstemp(dir='.', prefix=tempprefix)
    os.close(fd)
    try:
        s1 = os.stat(path)
        d, b = os.path.split(path)
        p2 = os.path.join(d, b.upper())
        if path == p2:
            p2 = os.path.join(d, b.lower())
        try:
            s2 = os.stat(p2)
            return s2 == s1
        except OSError:
            return False
    finally:
        os.remove(path)

@check("fifo", "named pipes")
def has_fifo():
    if getattr(os, "mkfifo", None) is None:
        return False
    name = tempfile.mktemp(dir='.', prefix=tempprefix)
    try:
        os.mkfifo(name)
        os.unlink(name)
        return True
    except OSError:
        return False

@check("killdaemons", 'killdaemons.py support')
def has_killdaemons():
    return True

@check("cacheable", "cacheable filesystem")
def has_cacheable_fs():
    from mercurial import util

    fd, path = tempfile.mkstemp(dir='.', prefix=tempprefix)
    os.close(fd)
    try:
        return util.cachestat(path).cacheable()
    finally:
        os.remove(path)

@check("lsprof", "python lsprof module")
def has_lsprof():
    try:
        import _lsprof
        _lsprof.Profiler # silence unused import warning
        return True
    except ImportError:
        return False

def gethgversion():
    m = matchoutput('hg --version --quiet 2>&1', br'(\d+)\.(\d+)')
    if not m:
        return (0, 0)
    return (int(m.group(1)), int(m.group(2)))

@checkvers("hg", "Mercurial >= %s",
            list([(1.0 * x) / 10 for x in range(9, 40)]))
def has_hg_range(v):
    major, minor = v.split('.')[0:2]
    return gethgversion() >= (int(major), int(minor))

@check("hg08", "Mercurial >= 0.8")
def has_hg08():
    if checks["hg09"][0]():
        return True
    return matchoutput('hg help annotate 2>&1', '--date')

@check("hg07", "Mercurial >= 0.7")
def has_hg07():
    if checks["hg08"][0]():
        return True
    return matchoutput('hg --version --quiet 2>&1', 'Mercurial Distributed SCM')

@check("hg06", "Mercurial >= 0.6")
def has_hg06():
    if checks["hg07"][0]():
        return True
    return matchoutput('hg --version --quiet 2>&1', 'Mercurial version')

@check("gettext", "GNU Gettext (msgfmt)")
def has_gettext():
    return matchoutput('msgfmt --version', br'GNU gettext-tools')

@check("git", "git command line client")
def has_git():
    return matchoutput('git --version 2>&1', br'^git version')

@check("docutils", "Docutils text processing library")
def has_docutils():
    try:
        import docutils.core
        docutils.core.publish_cmdline # silence unused import
        return True
    except ImportError:
        return False

def getsvnversion():
    m = matchoutput('svn --version --quiet 2>&1', br'^(\d+)\.(\d+)')
    if not m:
        return (0, 0)
    return (int(m.group(1)), int(m.group(2)))

@checkvers("svn", "subversion client and admin tools >= %s", (1.3, 1.5))
def has_svn_range(v):
    major, minor = v.split('.')[0:2]
    return getsvnversion() >= (int(major), int(minor))

@check("svn", "subversion client and admin tools")
def has_svn():
    return matchoutput('svn --version 2>&1', br'^svn, version') and \
        matchoutput('svnadmin --version 2>&1', br'^svnadmin, version')

@check("svn-bindings", "subversion python bindings")
def has_svn_bindings():
    try:
        import svn.core
        version = svn.core.SVN_VER_MAJOR, svn.core.SVN_VER_MINOR
        if version < (1, 4):
            return False
        return True
    except ImportError:
        return False

@check("p4", "Perforce server and client")
def has_p4():
    return (matchoutput('p4 -V', br'Rev\. P4/') and
            matchoutput('p4d -V', br'Rev\. P4D/'))

@check("symlink", "symbolic links")
def has_symlink():
    if getattr(os, "symlink", None) is None:
        return False
    name = tempfile.mktemp(dir='.', prefix=tempprefix)
    try:
        os.symlink(".", name)
        os.unlink(name)
        return True
    except (OSError, AttributeError):
        return False

@check("hardlink", "hardlinks")
def has_hardlink():
    from mercurial import util
    fh, fn = tempfile.mkstemp(dir='.', prefix=tempprefix)
    os.close(fh)
    name = tempfile.mktemp(dir='.', prefix=tempprefix)
    try:
        util.oslink(fn, name)
        os.unlink(name)
        return True
    except OSError:
        return False
    finally:
        os.unlink(fn)

@check("tla", "GNU Arch tla client")
def has_tla():
    return matchoutput('tla --version 2>&1', br'The GNU Arch Revision')

@check("gpg", "gpg client")
def has_gpg():
    return matchoutput('gpg --version 2>&1', br'GnuPG')

@check("gpg2", "gpg client v2")
def has_gpg2():
    return matchoutput('gpg --version 2>&1', br'GnuPG[^0-9]+2\.')

@check("gpg21", "gpg client v2.1+")
def has_gpg21():
    return matchoutput('gpg --version 2>&1', br'GnuPG[^0-9]+2\.(?!0)')

@check("unix-permissions", "unix-style permissions")
def has_unix_permissions():
    d = tempfile.mkdtemp(dir='.', prefix=tempprefix)
    try:
        fname = os.path.join(d, 'foo')
        for umask in (0o77, 0o07, 0o22):
            os.umask(umask)
            f = open(fname, 'w')
            f.close()
            mode = os.stat(fname).st_mode
            os.unlink(fname)
            if mode & 0o777 != ~umask & 0o666:
                return False
        return True
    finally:
        os.rmdir(d)

@check("unix-socket", "AF_UNIX socket family")
def has_unix_socket():
    return getattr(socket, 'AF_UNIX', None) is not None

@check("root", "root permissions")
def has_root():
    return getattr(os, 'geteuid', None) and os.geteuid() == 0

@check("pyflakes", "Pyflakes python linter")
def has_pyflakes():
    return matchoutput("sh -c \"echo 'import re' 2>&1 | pyflakes\"",
                       br"<stdin>:1: 're' imported but unused",
                       True)

@check("pygments", "Pygments source highlighting library")
def has_pygments():
    try:
        import pygments
        pygments.highlight # silence unused import warning
        return True
    except ImportError:
        return False

@check("outer-repo", "outer repo")
def has_outer_repo():
    # failing for other reasons than 'no repo' imply that there is a repo
    return not matchoutput('hg root 2>&1',
                           br'abort: no repository found', True)

@check("ssl", "ssl module available")
def has_ssl():
    try:
        import ssl
        ssl.CERT_NONE
        return True
    except ImportError:
        return False

@check("sslcontext", "python >= 2.7.9 ssl")
def has_sslcontext():
    try:
        import ssl
        ssl.SSLContext
        return True
    except (ImportError, AttributeError):
        return False

@check("defaultcacerts", "can verify SSL certs by system's CA certs store")
def has_defaultcacerts():
    from mercurial import sslutil, ui as uimod
    ui = uimod.ui()
    return sslutil._defaultcacerts(ui) or sslutil._canloaddefaultcerts

@check("defaultcacertsloaded", "detected presence of loaded system CA certs")
def has_defaultcacertsloaded():
    import ssl
    from mercurial import sslutil, ui as uimod

    if not has_defaultcacerts():
        return False
    if not has_sslcontext():
        return False

    ui = uimod.ui()
    cafile = sslutil._defaultcacerts(ui)
    ctx = ssl.create_default_context()
    if cafile:
        ctx.load_verify_locations(cafile=cafile)
    else:
        ctx.load_default_certs()

    return len(ctx.get_ca_certs()) > 0

@check("tls1.2", "TLS 1.2 protocol support")
def has_tls1_2():
    from mercurial import sslutil
    return 'tls1.2' in sslutil.supportedprotocols

@check("windows", "Windows")
def has_windows():
    return os.name == 'nt'

@check("system-sh", "system() uses sh")
def has_system_sh():
    return os.name != 'nt'

@check("serve", "platform and python can manage 'hg serve -d'")
def has_serve():
    return os.name != 'nt' # gross approximation

@check("test-repo", "running tests from repository")
def has_test_repo():
    t = os.environ["TESTDIR"]
    return os.path.isdir(os.path.join(t, "..", ".hg"))

@check("tic", "terminfo compiler and curses module")
def has_tic():
    try:
        import curses
        curses.COLOR_BLUE
        return matchoutput('test -x "`which tic`"', br'')
    except ImportError:
        return False

@check("msys", "Windows with MSYS")
def has_msys():
    return os.getenv('MSYSTEM')

@check("aix", "AIX")
def has_aix():
    return sys.platform.startswith("aix")

@check("osx", "OS X")
def has_osx():
    return sys.platform == 'darwin'

@check("osxpackaging", "OS X packaging tools")
def has_osxpackaging():
    try:
        return (matchoutput('pkgbuild', br'Usage: pkgbuild ', ignorestatus=1)
                and matchoutput(
                    'productbuild', br'Usage: productbuild ',
                    ignorestatus=1)
                and matchoutput('lsbom', br'Usage: lsbom', ignorestatus=1)
                and matchoutput(
                    'xar --help', br'Usage: xar', ignorestatus=1))
    except ImportError:
        return False

@check("docker", "docker support")
def has_docker():
    pat = br'A self-sufficient runtime for'
    if matchoutput('docker --help', pat):
        if 'linux' not in sys.platform:
            # TODO: in theory we should be able to test docker-based
            # package creation on non-linux using boot2docker, but in
            # practice that requires extra coordination to make sure
            # $TESTTEMP is going to be visible at the same path to the
            # boot2docker VM. If we figure out how to verify that, we
            # can use the following instead of just saying False:
            # return 'DOCKER_HOST' in os.environ
            return False

        return True
    return False

@check("debhelper", "debian packaging tools")
def has_debhelper():
    dpkg = matchoutput('dpkg --version',
                       br"Debian `dpkg' package management program")
    dh = matchoutput('dh --help',
                     br'dh is a part of debhelper.', ignorestatus=True)
    dh_py2 = matchoutput('dh_python2 --help',
                         br'other supported Python versions')
    return dpkg and dh and dh_py2

@check("demandimport", "demandimport enabled")
def has_demandimport():
    return os.environ.get('HGDEMANDIMPORT') != 'disable'

@check("absimport", "absolute_import in __future__")
def has_absimport():
    import __future__
    from mercurial import util
    return util.safehasattr(__future__, "absolute_import")

@check("py27+", "running with Python 2.7+")
def has_python27ornewer():
    return sys.version_info[0:2] >= (2, 7)

@check("py3k", "running with Python 3.x")
def has_py3k():
    return 3 == sys.version_info[0]

@check("py3exe", "a Python 3.x interpreter is available")
def has_python3exe():
    return 'PYTHON3' in os.environ

@check("py3pygments", "Pygments available on Python 3.x")
def has_py3pygments():
    if has_py3k():
        return has_pygments()
    elif has_python3exe():
        # just check exit status (ignoring output)
        py3 = os.environ['PYTHON3']
        return matchoutput('%s -c "import pygments"' % py3, br'')
    return False

@check("pure", "running with pure Python code")
def has_pure():
    return any([
        os.environ.get("HGMODULEPOLICY") == "py",
        os.environ.get("HGTEST_RUN_TESTS_PURE") == "--pure",
    ])

@check("slow", "allow slow tests")
def has_slow():
    return os.environ.get('HGTEST_SLOW') == 'slow'

@check("hypothesis", "Hypothesis automated test generation")
def has_hypothesis():
    try:
        import hypothesis
        hypothesis.given
        return True
    except ImportError:
        return False

@check("unziplinks", "unzip(1) understands and extracts symlinks")
def unzip_understands_symlinks():
    return matchoutput('unzip --help', br'Info-ZIP')