view tests/test-manifest.py @ 40326:fed697fa1734

sqlitestore: file storage backend using SQLite

This commit provides an extension which uses SQLite to store file
data (as opposed to revlogs).

As the inline documentation describes, there are still several
aspects to the extension that are incomplete. But it's a start.
The extension does support basic clone, checkout, and commit
workflows, which makes it suitable for simple use cases.

One notable missing feature is support for "bundlerepos." This is
probably responsible for the most test failures when the extension
is activated as part of the test suite.

All revision data is stored in SQLite. Data is stored as zstd
compressed chunks (default if zstd is available), zlib compressed
chunks (default if zstd is not available), or raw chunks (if
configured or if a compressed delta is not smaller than the raw
delta). This makes things very similar to revlogs.

Unlike revlogs, the extension doesn't yet enforce a limit on delta
chain length. This is an obvious limitation and should be addressed.
This is somewhat mitigated by the use of zstd, which is much faster
than zlib to decompress.

There is a dedicated table for storing deltas. Deltas are stored
by the SHA-1 hash of their uncompressed content. The "fileindex" table
has columns that reference the delta for each revision and the base
delta that delta should be applied against. A recursive SQL query
is used to resolve the delta chain along with the delta data.

By storing deltas by hash, we are able to de-duplicate delta storage!
With revlogs, the same deltas in different revlogs would result in
duplicate storage of that delta. In this scheme, inserting the
duplicate delta is a no-op and delta chains simply reference the
existing delta.

When initially implementing this extension, I did not have
content-indexed deltas and deltas could be duplicated across files
(just like revlogs). When I implemented content-indexed deltas, the
size of the SQLite database for a full clone of mozilla-unified
dropped:

before: 2,554,261,504 bytes
after:  2,488,754,176 bytes

Surprisingly, this is still larger than the bytes size of revlog
files:

revlog files: 2,104,861,230 bytes
du -b:        2,254,381,614

I would have expected storage to be smaller since we're not limiting
delta chain length and since we're using zstd instead of zlib. I
suspect the SQLite indexes and per-column overhead account for the
bulk of the differences. (Keep in mind that revlog uses a 64-byte
packed struct for revision index data and deltas are stored without
padding. Aside from the 12 unused bytes in the 32 byte node field,
revlogs are pretty efficient.) Another source of overhead is file
name storage. With revlogs, file names are stored in the filesystem.
But with SQLite, we need to store file names in the database. This is
roughly equivalent to the size of the fncache file, which for the
mozilla-unified repository is ~34MB.

Since the SQLite database isn't append-only and since delta chains
can reference any delta, this opens some interesting possibilities.
For example, we could store deltas in reverse, such that fulltexts
are stored for newer revisions and deltas are applied to reconstruct
older revisions. This is likely a more optimal storage strategy for
version control, as new data tends to be more frequently accessed
than old data. We would obviously need wire protocol support for
transferring revision data from newest to oldest. And we would
probably need some kind of mechanism for "re-encoding" stores. But
it should be doable.

This extension is very much experimental quality. There are a handful
of features that don't work. It probably isn't suitable for day-to-day
use. But it could be used in limited cases (e.g. read-only checkouts
like in CI). And it is also a good proving ground for alternate
storage backends. As we continue to define interfaces for all things
storage, it will be useful to have a viable alternate storage backend
to see how things shake out in practice.

test-storage.py passes on Python 2 and introduces no new test failures on
Python 3. Having the storage-level unit tests has proved to be insanely
useful when developing this extension. Those tests caught numerous bugs
during development and I'm convinced this style of testing is the way
forward for ensuring alternate storage backends work as intended. Of
course, test coverage isn't close to what it needs to be. But it is
a start. And what coverage we have gives me confidence that basic store
functionality is implemented properly.

Differential Revision: https://phab.mercurial-scm.org/D4928

from __future__ import absolute_import

import binascii
import itertools
import silenttestrunner
import unittest

from mercurial import (
    manifest as manifestmod,
    match as matchmod,
)

EMTPY_MANIFEST = b''

HASH_1 = b'1' * 40
BIN_HASH_1 = binascii.unhexlify(HASH_1)
HASH_2 = b'f' * 40
BIN_HASH_2 = binascii.unhexlify(HASH_2)
HASH_3 = b'1234567890abcdef0987654321deadbeef0fcafe'
BIN_HASH_3 = binascii.unhexlify(HASH_3)
A_SHORT_MANIFEST = (
    b'bar/baz/qux.py\0%(hash2)s%(flag2)s\n'
    b'foo\0%(hash1)s%(flag1)s\n'
    ) % {b'hash1': HASH_1,
         b'flag1': b'',
         b'hash2': HASH_2,
         b'flag2': b'l',
         }

A_DEEPER_MANIFEST = (
    b'a/b/c/bar.py\0%(hash3)s%(flag1)s\n'
    b'a/b/c/bar.txt\0%(hash1)s%(flag1)s\n'
    b'a/b/c/foo.py\0%(hash3)s%(flag1)s\n'
    b'a/b/c/foo.txt\0%(hash2)s%(flag2)s\n'
    b'a/b/d/baz.py\0%(hash3)s%(flag1)s\n'
    b'a/b/d/qux.py\0%(hash1)s%(flag2)s\n'
    b'a/b/d/ten.txt\0%(hash3)s%(flag2)s\n'
    b'a/b/dog.py\0%(hash3)s%(flag1)s\n'
    b'a/b/fish.py\0%(hash2)s%(flag1)s\n'
    b'a/c/london.py\0%(hash3)s%(flag2)s\n'
    b'a/c/paper.txt\0%(hash2)s%(flag2)s\n'
    b'a/c/paris.py\0%(hash2)s%(flag1)s\n'
    b'a/d/apple.py\0%(hash3)s%(flag1)s\n'
    b'a/d/pizza.py\0%(hash3)s%(flag2)s\n'
    b'a/green.py\0%(hash1)s%(flag2)s\n'
    b'a/purple.py\0%(hash2)s%(flag1)s\n'
    b'app.py\0%(hash3)s%(flag1)s\n'
    b'readme.txt\0%(hash2)s%(flag1)s\n'
    ) % {b'hash1': HASH_1,
         b'flag1': b'',
         b'hash2': HASH_2,
         b'flag2': b'l',
         b'hash3': HASH_3,
         }

HUGE_MANIFEST_ENTRIES = 200001

izip = getattr(itertools, 'izip', zip)
if 'xrange' not in globals():
    xrange = range

A_HUGE_MANIFEST = b''.join(sorted(
    b'file%d\0%s%s\n' % (i, h, f) for i, h, f in
    izip(xrange(200001),
         itertools.cycle((HASH_1, HASH_2)),
         itertools.cycle((b'', b'x', b'l')))))

class basemanifesttests(object):
    def parsemanifest(self, text):
        raise NotImplementedError('parsemanifest not implemented by test case')

    def testEmptyManifest(self):
        m = self.parsemanifest(EMTPY_MANIFEST)
        self.assertEqual(0, len(m))
        self.assertEqual([], list(m))

    def testManifest(self):
        m = self.parsemanifest(A_SHORT_MANIFEST)
        self.assertEqual([b'bar/baz/qux.py', b'foo'], list(m))
        self.assertEqual(BIN_HASH_2, m[b'bar/baz/qux.py'])
        self.assertEqual(b'l', m.flags(b'bar/baz/qux.py'))
        self.assertEqual(BIN_HASH_1, m[b'foo'])
        self.assertEqual(b'', m.flags(b'foo'))
        with self.assertRaises(KeyError):
            m[b'wat']

    def testSetItem(self):
        want = BIN_HASH_1

        m = self.parsemanifest(EMTPY_MANIFEST)
        m[b'a'] = want
        self.assertIn(b'a', m)
        self.assertEqual(want, m[b'a'])
        self.assertEqual(b'a\0' + HASH_1 + b'\n', m.text())

        m = self.parsemanifest(A_SHORT_MANIFEST)
        m[b'a'] = want
        self.assertEqual(want, m[b'a'])
        self.assertEqual(b'a\0' + HASH_1 + b'\n' + A_SHORT_MANIFEST,
                         m.text())

    def testSetFlag(self):
        want = b'x'

        m = self.parsemanifest(EMTPY_MANIFEST)
        # first add a file; a file-less flag makes no sense
        m[b'a'] = BIN_HASH_1
        m.setflag(b'a', want)
        self.assertEqual(want, m.flags(b'a'))
        self.assertEqual(b'a\0' + HASH_1 + want + b'\n', m.text())

        m = self.parsemanifest(A_SHORT_MANIFEST)
        # first add a file; a file-less flag makes no sense
        m[b'a'] = BIN_HASH_1
        m.setflag(b'a', want)
        self.assertEqual(want, m.flags(b'a'))
        self.assertEqual(b'a\0' + HASH_1 + want + b'\n' + A_SHORT_MANIFEST,
                         m.text())

    def testCopy(self):
        m = self.parsemanifest(A_SHORT_MANIFEST)
        m[b'a'] = BIN_HASH_1
        m2 = m.copy()
        del m
        del m2 # make sure we don't double free() anything

    def testCompaction(self):
        unhex = binascii.unhexlify
        h1, h2 = unhex(HASH_1), unhex(HASH_2)
        m = self.parsemanifest(A_SHORT_MANIFEST)
        m[b'alpha'] = h1
        m[b'beta'] = h2
        del m[b'foo']
        want = b'alpha\0%s\nbar/baz/qux.py\0%sl\nbeta\0%s\n' % (
            HASH_1, HASH_2, HASH_2)
        self.assertEqual(want, m.text())
        self.assertEqual(3, len(m))
        self.assertEqual([b'alpha', b'bar/baz/qux.py', b'beta'], list(m))
        self.assertEqual(h1, m[b'alpha'])
        self.assertEqual(h2, m[b'bar/baz/qux.py'])
        self.assertEqual(h2, m[b'beta'])
        self.assertEqual(b'', m.flags(b'alpha'))
        self.assertEqual(b'l', m.flags(b'bar/baz/qux.py'))
        self.assertEqual(b'', m.flags(b'beta'))
        with self.assertRaises(KeyError):
            m[b'foo']

    def testSetGetNodeSuffix(self):
        clean = self.parsemanifest(A_SHORT_MANIFEST)
        m = self.parsemanifest(A_SHORT_MANIFEST)
        h = m[b'foo']
        f = m.flags(b'foo')
        want = h + b'a'
        # Merge code wants to set 21-byte fake hashes at times
        m[b'foo'] = want
        self.assertEqual(want, m[b'foo'])
        self.assertEqual([(b'bar/baz/qux.py', BIN_HASH_2),
                          (b'foo', BIN_HASH_1 + b'a')],
                         list(m.items()))
        # Sometimes it even tries a 22-byte fake hash, but we can
        # return 21 and it'll work out
        m[b'foo'] = want + b'+'
        self.assertEqual(want, m[b'foo'])
        # make sure the suffix survives a copy
        match = matchmod.match(b'', b'', [b're:foo'])
        m2 = m.matches(match)
        self.assertEqual(want, m2[b'foo'])
        self.assertEqual(1, len(m2))
        m2 = m.copy()
        self.assertEqual(want, m2[b'foo'])
        # suffix with iteration
        self.assertEqual([(b'bar/baz/qux.py', BIN_HASH_2),
                          (b'foo', want)],
                         list(m.items()))

        # shows up in diff
        self.assertEqual({b'foo': ((want, f), (h, b''))}, m.diff(clean))
        self.assertEqual({b'foo': ((h, b''), (want, f))}, clean.diff(m))

    def testMatchException(self):
        m = self.parsemanifest(A_SHORT_MANIFEST)
        match = matchmod.match(b'', b'', [b're:.*'])
        def filt(path):
            if path == b'foo':
                assert False
            return True
        match.matchfn = filt
        with self.assertRaises(AssertionError):
            m.matches(match)

    def testRemoveItem(self):
        m = self.parsemanifest(A_SHORT_MANIFEST)
        del m[b'foo']
        with self.assertRaises(KeyError):
            m[b'foo']
        self.assertEqual(1, len(m))
        self.assertEqual(1, len(list(m)))
        # now restore and make sure everything works right
        m[b'foo'] = b'a' * 20
        self.assertEqual(2, len(m))
        self.assertEqual(2, len(list(m)))

    def testManifestDiff(self):
        MISSING = (None, b'')
        addl = b'z-only-in-left\0' + HASH_1 + b'\n'
        addr = b'z-only-in-right\0' + HASH_2 + b'x\n'
        left = self.parsemanifest(
            A_SHORT_MANIFEST.replace(HASH_1, HASH_3 + b'x') + addl)
        right = self.parsemanifest(A_SHORT_MANIFEST + addr)
        want = {
            b'foo': ((BIN_HASH_3, b'x'),
                     (BIN_HASH_1, b'')),
            b'z-only-in-left': ((BIN_HASH_1, b''), MISSING),
            b'z-only-in-right': (MISSING, (BIN_HASH_2, b'x')),
            }
        self.assertEqual(want, left.diff(right))

        want = {
            b'bar/baz/qux.py': (MISSING, (BIN_HASH_2, b'l')),
            b'foo': (MISSING, (BIN_HASH_3, b'x')),
            b'z-only-in-left': (MISSING, (BIN_HASH_1, b'')),
            }
        self.assertEqual(want, self.parsemanifest(EMTPY_MANIFEST).diff(left))

        want = {
            b'bar/baz/qux.py': ((BIN_HASH_2, b'l'), MISSING),
            b'foo': ((BIN_HASH_3, b'x'), MISSING),
            b'z-only-in-left': ((BIN_HASH_1, b''), MISSING),
            }
        self.assertEqual(want, left.diff(self.parsemanifest(EMTPY_MANIFEST)))
        copy = right.copy()
        del copy[b'z-only-in-right']
        del right[b'foo']
        want = {
            b'foo': (MISSING, (BIN_HASH_1, b'')),
            b'z-only-in-right': ((BIN_HASH_2, b'x'), MISSING),
            }
        self.assertEqual(want, right.diff(copy))

        short = self.parsemanifest(A_SHORT_MANIFEST)
        pruned = short.copy()
        del pruned[b'foo']
        want = {
            b'foo': ((BIN_HASH_1, b''), MISSING),
            }
        self.assertEqual(want, short.diff(pruned))
        want = {
            b'foo': (MISSING, (BIN_HASH_1, b'')),
            }
        self.assertEqual(want, pruned.diff(short))
        want = {
            b'bar/baz/qux.py': None,
            b'foo': (MISSING, (BIN_HASH_1, b'')),
            }
        self.assertEqual(want, pruned.diff(short, clean=True))

    def testReversedLines(self):
        backwards = b''.join(
            l + b'\n' for l in reversed(A_SHORT_MANIFEST.split(b'\n')) if l)
        try:
            self.parsemanifest(backwards)
            self.fail('Should have raised ValueError')
        except ValueError as v:
            self.assertIn('Manifest lines not in sorted order.', str(v))

    def testNoTerminalNewline(self):
        try:
            self.parsemanifest(A_SHORT_MANIFEST + b'wat')
            self.fail('Should have raised ValueError')
        except ValueError as v:
            self.assertIn('Manifest did not end in a newline.', str(v))

    def testNoNewLineAtAll(self):
        try:
            self.parsemanifest(b'wat')
            self.fail('Should have raised ValueError')
        except ValueError as v:
            self.assertIn('Manifest did not end in a newline.', str(v))

    def testHugeManifest(self):
        m = self.parsemanifest(A_HUGE_MANIFEST)
        self.assertEqual(HUGE_MANIFEST_ENTRIES, len(m))
        self.assertEqual(len(m), len(list(m)))

    def testMatchesMetadata(self):
        '''Tests matches() for a few specific files to make sure that both
        the set of files as well as their flags and nodeids are correct in
        the resulting manifest.'''
        m = self.parsemanifest(A_HUGE_MANIFEST)

        match = matchmod.match(b'/', b'',
                [b'file1', b'file200', b'file300'], exact=True)
        m2 = m.matches(match)

        w = (b'file1\0%sx\n'
             b'file200\0%sl\n'
             b'file300\0%s\n') % (HASH_2, HASH_1, HASH_1)
        self.assertEqual(w, m2.text())

    def testMatchesNonexistentFile(self):
        '''Tests matches() for a small set of specific files, including one
        nonexistent file to make sure in only matches against existing files.
        '''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'',
                [b'a/b/c/bar.txt', b'a/b/d/qux.py',
                 b'readme.txt', b'nonexistent'],
                exact=True)
        m2 = m.matches(match)

        self.assertEqual(
                [b'a/b/c/bar.txt', b'a/b/d/qux.py', b'readme.txt'],
                m2.keys())

    def testMatchesNonexistentDirectory(self):
        '''Tests matches() for a relpath match on a directory that doesn't
        actually exist.'''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'', [b'a/f'], default=b'relpath')
        m2 = m.matches(match)

        self.assertEqual([], m2.keys())

    def testMatchesExactLarge(self):
        '''Tests matches() for files matching a large list of exact files.
        '''
        m = self.parsemanifest(A_HUGE_MANIFEST)

        flist = m.keys()[80:300]
        match = matchmod.match(b'/', b'', flist, exact=True)
        m2 = m.matches(match)

        self.assertEqual(flist, m2.keys())

    def testMatchesFull(self):
        '''Tests matches() for what should be a full match.'''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'', [b''])
        m2 = m.matches(match)

        self.assertEqual(m.keys(), m2.keys())

    def testMatchesDirectory(self):
        '''Tests matches() on a relpath match on a directory, which should
        match against all files within said directory.'''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'', [b'a/b'], default=b'relpath')
        m2 = m.matches(match)

        self.assertEqual([
            b'a/b/c/bar.py', b'a/b/c/bar.txt', b'a/b/c/foo.py',
            b'a/b/c/foo.txt',
            b'a/b/d/baz.py', b'a/b/d/qux.py', b'a/b/d/ten.txt', b'a/b/dog.py',
            b'a/b/fish.py'], m2.keys())

    def testMatchesExactPath(self):
        '''Tests matches() on an exact match on a directory, which should
        result in an empty manifest because you can't perform an exact match
        against a directory.'''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'', [b'a/b'], exact=True)
        m2 = m.matches(match)

        self.assertEqual([], m2.keys())

    def testMatchesCwd(self):
        '''Tests matches() on a relpath match with the current directory ('.')
        when not in the root directory.'''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'a/b', [b'.'], default=b'relpath')
        m2 = m.matches(match)

        self.assertEqual([
            b'a/b/c/bar.py', b'a/b/c/bar.txt', b'a/b/c/foo.py',
            b'a/b/c/foo.txt', b'a/b/d/baz.py', b'a/b/d/qux.py',
            b'a/b/d/ten.txt', b'a/b/dog.py', b'a/b/fish.py'], m2.keys())

    def testMatchesWithPattern(self):
        '''Tests matches() for files matching a pattern that reside
        deeper than the specified directory.'''
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        match = matchmod.match(b'/', b'', [b'a/b/*/*.txt'])
        m2 = m.matches(match)

        self.assertEqual(
                [b'a/b/c/bar.txt', b'a/b/c/foo.txt', b'a/b/d/ten.txt'],
                m2.keys())

class testmanifestdict(unittest.TestCase, basemanifesttests):
    def parsemanifest(self, text):
        return manifestmod.manifestdict(text)

class testtreemanifest(unittest.TestCase, basemanifesttests):
    def parsemanifest(self, text):
        return manifestmod.treemanifest(b'', text)

    def testWalkSubtrees(self):
        m = self.parsemanifest(A_DEEPER_MANIFEST)

        dirs = [s._dir for s in m.walksubtrees()]
        self.assertEqual(
            sorted([
                b'', b'a/', b'a/c/', b'a/d/', b'a/b/', b'a/b/c/', b'a/b/d/']),
            sorted(dirs)
        )

        match = matchmod.match(b'/', b'', [b'path:a/b/'])
        dirs = [s._dir for s in m.walksubtrees(matcher=match)]
        self.assertEqual(
            sorted([b'a/b/', b'a/b/c/', b'a/b/d/']),
            sorted(dirs)
        )

if __name__ == '__main__':
    silenttestrunner.main(__name__)