mercurial/manifest.py

dirstate: avoid a race with multiple commits in the same process
(issue2264, issue2516)

The race happens when two commits in a row change the same file
without changing its size, *if* those two commits happen in the same
second in the same process while holding the same repo lock.  For
example:

  commit 1:
    M a
    M b
  commit 2:           # same process, same second, same repo lock
    M b               # modify b without changing its size
    M c

This first manifested in transplant, which is the most common way to
do multiple commits in the same process. But it can manifest in any
script or extension that does multiple commits under the same repo
lock. (Thus, the test script tests both transplant and a custom script.)

The problem was that dirstate.status() failed to notice the change to
b when localrepo is about to do the second commit, meaning that change
gets left in the working directory. In the context of transplant, that
means either a crash ("RuntimeError: nothing committed after
transplant") or a silently inaccurate transplant, depending on whether
any other files were modified by the second transplanted changeset.

The fix is to make status() work a little harder when we have
previously marked files as clean (state 'normal') in the same process.
Specifically, dirstate.normal() adds files to self._lastnormal, and
other state-changing methods remove them. Then dirstate.status() puts
any files in self._lastnormal into state 'lookup', which will make
localrepository.status() read file contents to see if it has really
changed.  So we pay a small performance penalty for the second (and
subsequent) commits in the same process, without affecting the common
case.  Anything that does lots of status updates and checks in the
same process could suffer a performance hit.

Incidentally, there is a simpler fix: call dirstate.normallookup() on
every file updated by commit() at the end of the commit.  The trouble
with that solution is that it imposes a performance penalty on the
common case: it means the next status-dependent hg command after every
"hg commit" will be a little bit slower.  The patch here is more
complex, but only affects performance for the uncommon case.

# manifest.py - manifest revision class for mercurial
#
# Copyright 2005-2007 Matt Mackall <[e-mail]>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.

from i18n import _
import mdiff, parsers, error, revlog
import array, struct

class manifestdict(dict):
    def __init__(self, mapping=None, flags=None):
        if mapping is None:
            mapping = {}
        if flags is None:
            flags = {}
        dict.__init__(self, mapping)
        self._flags = flags
    def flags(self, f):
        return self._flags.get(f, "")
    def set(self, f, flags):
        self._flags[f] = flags
    def copy(self):
        return manifestdict(self, dict.copy(self._flags))

class manifest(revlog.revlog):
    def __init__(self, opener):
        self._mancache = None
        revlog.revlog.__init__(self, opener, "00manifest.i")

    def parse(self, lines):
        mfdict = manifestdict()
        parsers.parse_manifest(mfdict, mfdict._flags, lines)
        return mfdict

    def readdelta(self, node):
        r = self.rev(node)
        return self.parse(mdiff.patchtext(self.revdiff(self.deltaparent(r), r)))

    def read(self, node):
        if node == revlog.nullid:
            return manifestdict() # don't upset local cache
        if self._mancache and self._mancache[0] == node:
            return self._mancache[1]
        text = self.revision(node)
        arraytext = array.array('c', text)
        mapping = self.parse(text)
        self._mancache = (node, mapping, arraytext)
        return mapping

    def _search(self, m, s, lo=0, hi=None):
        '''return a tuple (start, end) that says where to find s within m.

        If the string is found m[start:end] are the line containing
        that string.  If start == end the string was not found and
        they indicate the proper sorted insertion point.  This was
        taken from bisect_left, and modified to find line start/end as
        it goes along.

        m should be a buffer or a string
        s is a string'''
        def advance(i, c):
            while i < lenm and m[i] != c:
                i += 1
            return i
        if not s:
            return (lo, lo)
        lenm = len(m)
        if not hi:
            hi = lenm
        while lo < hi:
            mid = (lo + hi) // 2
            start = mid
            while start > 0 and m[start - 1] != '\n':
                start -= 1
            end = advance(start, '\0')
            if m[start:end] < s:
                # we know that after the null there are 40 bytes of sha1
                # this translates to the bisect lo = mid + 1
                lo = advance(end + 40, '\n') + 1
            else:
                # this translates to the bisect hi = mid
                hi = start
        end = advance(lo, '\0')
        found = m[lo:end]
        if s == found:
            # we know that after the null there are 40 bytes of sha1
            end = advance(end + 40, '\n')
            return (lo, end + 1)
        else:
            return (lo, lo)

    def find(self, node, f):
        '''look up entry for a single file efficiently.
        return (node, flags) pair if found, (None, None) if not.'''
        if self._mancache and self._mancache[0] == node:
            return self._mancache[1].get(f), self._mancache[1].flags(f)
        text = self.revision(node)
        start, end = self._search(text, f)
        if start == end:
            return None, None
        l = text[start:end]
        f, n = l.split('\0')
        return revlog.bin(n[:40]), n[40:-1]

    def add(self, map, transaction, link, p1=None, p2=None,
            changed=None):
        # apply the changes collected during the bisect loop to our addlist
        # return a delta suitable for addrevision
        def addlistdelta(addlist, x):
            # start from the bottom up
            # so changes to the offsets don't mess things up.
            for start, end, content in reversed(x):
                if content:
                    addlist[start:end] = array.array('c', content)
                else:
                    del addlist[start:end]
            return "".join(struct.pack(">lll", start, end, len(content)) + content
                           for start, end, content in x)

        def checkforbidden(l):
            for f in l:
                if '\n' in f or '\r' in f:
                    raise error.RevlogError(
                        _("'\\n' and '\\r' disallowed in filenames: %r") % f)

        # if we're using the cache, make sure it is valid and
        # parented by the same node we're diffing against
        if not (changed and self._mancache and p1 and self._mancache[0] == p1):
            files = sorted(map)
            checkforbidden(files)

            # if this is changed to support newlines in filenames,
            # be sure to check the templates/ dir again (especially *-raw.tmpl)
            hex, flags = revlog.hex, map.flags
            text = ''.join("%s\000%s%s\n" % (f, hex(map[f]), flags(f))
                           for f in files)
            arraytext = array.array('c', text)
            cachedelta = None
        else:
            added, removed = changed
            addlist = self._mancache[2]

            checkforbidden(added)
            # combine the changed lists into one list for sorting
            work = [(x, False) for x in added]
            work.extend((x, True) for x in removed)
            # this could use heapq.merge() (from python2.6+) or equivalent
            # since the lists are already sorted
            work.sort()

            delta = []
            dstart = None
            dend = None
            dline = [""]
            start = 0
            # zero copy representation of addlist as a buffer
            addbuf = buffer(addlist)

            # start with a readonly loop that finds the offset of
            # each line and creates the deltas
            for f, todelete in work:
                # bs will either be the index of the item or the insert point
                start, end = self._search(addbuf, f, start)
                if not todelete:
                    l = "%s\000%s%s\n" % (f, revlog.hex(map[f]), map.flags(f))
                else:
                    if start == end:
                        # item we want to delete was not found, error out
                        raise AssertionError(
                                _("failed to remove %s from manifest") % f)
                    l = ""
                if dstart is not None and dstart <= start and dend >= start:
                    if dend < end:
                        dend = end
                    if l:
                        dline.append(l)
                else:
                    if dstart is not None:
                        delta.append([dstart, dend, "".join(dline)])
                    dstart = start
                    dend = end
                    dline = [l]

            if dstart is not None:
                delta.append([dstart, dend, "".join(dline)])
            # apply the delta to the addlist, and get a delta for addrevision
            cachedelta = (self.rev(p1), addlistdelta(addlist, delta))
            arraytext = addlist
            text = buffer(arraytext)

        n = self.addrevision(text, transaction, link, p1, p2, cachedelta)
        self._mancache = (n, map, arraytext)

        return n