~bzr-pqm/bzr/bzr.dev

# Copyright (C) 2005 Canonical Ltd

#

# This program is free software; you can redistribute it and/or modify

# it under the terms of the GNU General Public License as published by

# the Free Software Foundation; either version 2 of the License, or

# (at your option) any later version.

#

# This program is distributed in the hope that it will be useful,

# but WITHOUT ANY WARRANTY; without even the implied warranty of

# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

# GNU General Public License for more details.

#

# You should have received a copy of the GNU General Public License

# along with this program; if not, write to the Free Software

# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA

# XXX: Can we do any better about making interrupted commits change

# nothing?  Perhaps the best approach is to integrate commit of

# AtomicFiles with releasing the lock on the Branch.

# TODO: Separate 'prepare' phase where we find a list of potentially

# committed files.  We then can then pause the commit to prompt for a

# commit message, knowing the summary will be the same as what's

# actually used for the commit.  (But perhaps simpler to simply get

# the tree status, then use that for a selective commit?)

# The newly committed revision is going to have a shape corresponding

# to that of the working inventory.  Files that are not in the

# working tree and that were in the predecessor are reported as

# removed --- this can include files that were either removed from the

# inventory or deleted in the working tree.  If they were only

# deleted from disk, they are removed from the working inventory.

# We then consider the remaining entries, which will be in the new

# version.  Directory entries are simply copied across.  File entries

# must be checked to see if a new version of the file should be

# recorded.  For each parent revision inventory, we check to see what

# version of the file was present.  If the file was present in at

# least one tree, and if it was the same version in all the trees,

# then we can just refer to that version.  Otherwise, a new version

# representing the merger of the file versions must be added.

# TODO: Update hashcache before and after - or does the WorkingTree

# look after that?

# This code requires all merge parents to be present in the branch.

# We could relax this but for the sake of simplicity the constraint is

# here for now.  It's not totally clear to me how we'd know which file

# need new text versions if some parents are absent.  -- mbp 20050915

# TODO: Rather than mashing together the ancestry and storing it back,

# perhaps the weave should have single method which does it all in one

# go, avoiding a lot of redundant work.

# TODO: Perhaps give a warning if one of the revisions marked as

# merged is already in the ancestry, and then don't record it as a

# distinct parent.

# TODO: If the file is newly merged but unchanged from the version it

# merges from, then it should still be reported as newly added

# relative to the basis revision.

import os

import re

import sys

import time

import pdb

from binascii import hexlify

from cStringIO import StringIO

from bzrlib.osutils import (local_time_offset, username,

                            rand_bytes, compact_date, user_email,

                            kind_marker, is_inside_any, quotefn,

                            sha_string, sha_strings, sha_file, isdir, isfile,

                            split_lines)

from bzrlib.branch import gen_file_id

from bzrlib.errors import (BzrError, PointlessCommit,

                           HistoryMissing,

                           )

from bzrlib.revision import Revision

from bzrlib.trace import mutter, note, warning

from bzrlib.xml5 import serializer_v5

from bzrlib.inventory import Inventory

from bzrlib.weave import Weave

from bzrlib.weavefile import read_weave, write_weave_v5

from bzrlib.atomicfile import AtomicFile

def commit(*args, **kwargs):

    """Commit a new revision to a branch.

    Function-style interface for convenience of old callers.

    New code should use the Commit class instead.

    """

    ## XXX: Remove this in favor of Branch.commit?

    Commit().commit(*args, **kwargs)

class NullCommitReporter(object):

    """I report on progress of a commit."""

    def added(self, path):

        pass

    def removed(self, path):

        pass

    def renamed(self, old_path, new_path):

        pass

class ReportCommitToLog(NullCommitReporter):

    def added(self, path):

        note('added %s', path)

    def removed(self, path):

        note('removed %s', path)

    def renamed(self, old_path, new_path):

        note('renamed %s => %s', old_path, new_path)

class Commit(object):

    """Task of committing a new revision.

    This is a MethodObject: it accumulates state as the commit is

    prepared, and then it is discarded.  It doesn't represent

    historical revisions, just the act of recording a new one.

            missing_ids

            Modified to hold a list of files that have been deleted from

            the working directory; these should be removed from the

            working inventory.

    """

    def __init__(self,

                 reporter=None):

        if reporter is not None:

            self.reporter = reporter

        else:

            self.reporter = NullCommitReporter()

    def commit(self,

               branch, message,

               timestamp=None,

               timezone=None,

               committer=None,

               specific_files=None,

               rev_id=None,

               allow_pointless=True,

               verbose=False):

        """Commit working copy as a new revision.

        timestamp -- if not None, seconds-since-epoch for a

             postdated/predated commit.

        specific_files -- If true, commit only those files.

        rev_id -- If set, use this as the new revision id.

            Useful for test or import commands that need to tightly

            control what revisions are assigned.  If you duplicate

            a revision id that exists elsewhere it is your own fault.

            If null (default), a time/random revision id is generated.

        allow_pointless -- If true (default), commit even if nothing

            has changed and no merges are recorded.

        """

        mutter('preparing to commit')

        self.branch = branch

        self.weave_store = branch.weave_store

        self.rev_id = rev_id

        self.specific_files = specific_files

        self.allow_pointless = allow_pointless

        if timestamp is None:

            self.timestamp = time.time()

        else:

            self.timestamp = long(timestamp)

        if rev_id is None:

            self.rev_id = _gen_revision_id(self.branch, self.timestamp)

        else:

            self.rev_id = rev_id

        if committer is None:

            self.committer = username(self.branch)

        else:

            assert isinstance(committer, basestring), type(committer)

            self.committer = committer

        if timezone is None:

            self.timezone = local_time_offset()

        else:

            self.timezone = int(timezone)

        assert isinstance(message, basestring), type(message)

        self.message = message

        self._escape_commit_message()

        self.branch.lock_write()

        try:

            self.work_tree = self.branch.working_tree()

            self.work_inv = self.work_tree.inventory

            self.basis_tree = self.branch.basis_tree()

            self.basis_inv = self.basis_tree.inventory

            self._gather_parents()

            if len(self.parents) > 1 and self.specific_files:

                raise NotImplementedError('selected-file commit of merges is not supported yet')

            self._check_parents_present()

            self._remove_deleted()

            self.new_inv = Inventory()

            self._store_entries()

            self._report_deletes()

            self._set_name_versions()

            if not (self.allow_pointless

                    or len(self.parents) > 1

                    or self.new_inv != self.basis_inv):

                raise PointlessCommit()

            self._record_inventory()

            self._record_ancestry()

            self._make_revision()

            note('committed r%d {%s}', (self.branch.revno() + 1),

                 self.rev_id)

            self.branch.append_revision(self.rev_id)

            self.branch.set_pending_merges([])

        finally:

            self.branch.unlock()

    def _record_inventory(self):

        """Store the inventory for the new revision."""

        inv_text = serializer_v5.write_inventory_to_string(self.new_inv)

        self.inv_sha1 = sha_string(inv_text)

        s = self.branch.control_weaves

        s.add_text('inventory', self.rev_id,

                   split_lines(inv_text), self.parents)

    def _escape_commit_message(self):

        """Replace xml-incompatible control characters."""

        # Python strings can include characters that can't be

        # represented in well-formed XML; escape characters that

        # aren't listed in the XML specification

        # (http://www.w3.org/TR/REC-xml/#NT-Char).

        if isinstance(self.message, unicode):

            char_pattern = u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]'

        else:

            # Use a regular 'str' as pattern to avoid having re.subn

            # return 'unicode' results.

            char_pattern = '[^x09\x0A\x0D\x20-\xFF]'

        self.message, escape_count = re.subn(

            char_pattern,

            lambda match: match.group(0).encode('unicode_escape'),

            self.message)

        if escape_count:

            note("replaced %d control characters in message", escape_count)

    def _record_ancestry(self):

        """Append merged revision ancestry to the ancestry file.

        This should be the merged ancestry of all parents, plus the

        new revision id."""

        s = self.branch.control_weaves

        w = s.get_weave_or_empty('ancestry')

        lines = self._make_ancestry(w)

        w.add(self.rev_id, self.parents, lines)

        s.put_weave('ancestry', w)

    def _make_ancestry(self, ancestry_weave):

        """Return merged ancestry lines.

        The lines are revision-ids followed by newlines."""

        parent_ancestries = [ancestry_weave.get(p) for p in self.parents]

        new_lines = merge_ancestry_lines(self.rev_id, parent_ancestries)

        mutter('merged ancestry of {%s}:\n%s', self.rev_id, ''.join(new_lines))

        return new_lines

    def _gather_parents(self):

        pending_merges = self.branch.pending_merges()

        self.parents = []

        self.parent_trees = []

        precursor_id = self.branch.last_revision()

        if precursor_id:

            self.parents.append(precursor_id)

            self.parent_trees.append(self.basis_tree)

        self.parents += pending_merges

        self.parent_trees.extend(map(self.branch.revision_tree, pending_merges))

    def _check_parents_present(self):

        for parent_id in self.parents:

            mutter('commit parent revision {%s}', parent_id)

            if not self.branch.has_revision(parent_id):

                warning("can't commit a merge from an absent parent")

                raise HistoryMissing(self.branch, 'revision', parent_id)

    def _make_revision(self):

        """Record a new revision object for this commit."""

        self.rev = Revision(timestamp=self.timestamp,

                            timezone=self.timezone,

                            committer=self.committer,

                            message=self.message,

                            inventory_sha1=self.inv_sha1,

                            revision_id=self.rev_id)

        self.rev.parent_ids = self.parents

        rev_tmp = StringIO()

        serializer_v5.write_revision(self.rev, rev_tmp)

        rev_tmp.seek(0)

        self.branch.revision_store.add(rev_tmp, self.rev_id, compressed=False)

        mutter('new revision_id is {%s}', self.rev_id)

    def _remove_deleted(self):

        """Remove deleted files from the working inventories.

        This is done prior to taking the working inventory as the

        basis for the new committed inventory.

        This returns true if any files

        *that existed in the basis inventory* were deleted.

        Files that were added and deleted

        in the working copy don't matter.

        """

        specific = self.specific_files

        deleted_ids = []

        for path, ie in self.work_inv.iter_entries():

            if specific and not is_inside_any(specific, path):

                continue

            if not self.work_tree.has_filename(path):

                note('missing %s', path)

                deleted_ids.append(ie.file_id)

        if deleted_ids:

            for file_id in deleted_ids:

                del self.work_inv[file_id]

            self.branch._write_inventory(self.work_inv)

    def _find_file_parents(self, file_id):

        """Return the text versions and hashes for all file parents.

        Returned as a map from text version to inventory entry.

        This is a set containing the file versions in all parents

        revisions containing the file.  If the file is new, the set

        will be empty."""

        r = {}

        for tree in self.parent_trees:

            if file_id in tree.inventory:

                ie = tree.inventory[file_id]

                assert ie.kind == 'file'

                assert ie.file_id == file_id

                if ie.text_version in r:

                    assert r[ie.text_version] == ie

                else:

                    r[ie.text_version] = ie

        return r

    def _set_name_versions(self):

        """Pass over inventory and mark new entry version as needed.

        Files get a new name version when they are new, have a

        different parent, or a different name from in the

        basis inventory, or if the file is in a different place

        to any of the parents."""

        # XXX: Need to think more here about when the user has

        # made a specific decision on a particular value -- c.f.

        # mark-merge.  

        for path, ie in self.new_inv.iter_entries():

            old_version = None

            file_id = ie.file_id

            for parent_tree in self.parent_trees:

                parent_inv = parent_tree.inventory

                if file_id not in parent_inv:

                    continue

                parent_ie = parent_inv[file_id]

                if parent_ie.parent_id != ie.parent_id:

                    old_version = None

                    break

                elif parent_ie.name != ie.name:

                    old_version = None

                    break

                elif old_version is None:

                    old_version = parent_ie.name_version

                elif old_version != parent_ie.name_version:

                    old_version = None

                    break

                else:

                    pass                # so far so good

            if old_version is None:

                mutter('new name_version for {%s}', file_id)

                ie.name_version = self.rev_id

            else:

                mutter('name_version for {%s} inherited as {%s}',

                       file_id, old_version)

                ie.name_version = old_version

    def _store_entries(self):

        """Build revision inventory and store modified files.

        This is called with new_inv a new empty inventory.  Depending on

        which files are selected for commit, and which ones have

        been modified or merged, new inventory entries are built

        based on the working and parent inventories.

        As a side-effect this stores new text versions for committed

        files with text changes or merges.

        Each entry can have one of several things happen:

        carry_file -- carried from the previous version (if not

            selected for commit)

        commit_nonfile -- no text to worry about

        commit_old_text -- same text, may have moved

        commit_file -- new text version

        """

        for path, new_ie in self.work_inv.iter_entries():

            file_id = new_ie.file_id

            mutter('check %s {%s}', path, new_ie.file_id)

            if self.specific_files:

                if not is_inside_any(self.specific_files, path):

                    mutter('%s not selected for commit', path)

                    self._carry_file(file_id)

                    continue

            if new_ie.kind != 'file':

                self._commit_nonfile(file_id)

                continue

            file_parents = self._find_file_parents(file_id)

            mutter('parents of %s are %r', path, file_parents)

            if len(file_parents) == 1:

                parent_ie = file_parents.values()[0]

                wc_sha1 = self.work_tree.get_file_sha1(file_id)

                if parent_ie.text_sha1 == wc_sha1:

                    # text not changed or merged

                    self._commit_old_text(file_id, parent_ie)

                    continue

            # file is either new, or a file merge; need to record

            # a new version

            if len(file_parents) > 1:

                note('merged %s', path)

            elif len(file_parents) == 0:

                note('added %s', path)

            else:

                note('modified %s', path)

            self._commit_file(new_ie, file_id, file_parents)

    def _commit_nonfile(self, file_id):

        self.new_inv.add(self.work_inv[file_id].copy())

    def _carry_file(self, file_id):

        """Carry the file unchanged from the basis revision."""

        if self.basis_inv.has_id(file_id):

            self.new_inv.add(self.basis_inv[file_id].copy())

    def _commit_old_text(self, file_id, parent_ie):

        """Keep the same text as last time, but possibly a different name."""

        ie = self.work_inv[file_id].copy()

        ie.text_version = parent_ie.text_version

        ie.text_size = parent_ie.text_size

        ie.text_sha1 = parent_ie.text_sha1

        self.new_inv.add(ie)

    def _report_deletes(self):

        for file_id in self.basis_inv:

            if file_id not in self.new_inv:

                note('deleted %s', self.basis_inv.id2path(file_id))

    def _commit_file(self, new_ie, file_id, file_parents):                    

        mutter('store new text for {%s} in revision {%s}',

               file_id, self.rev_id)

        new_lines = self.work_tree.get_file(file_id).readlines()

        self._add_text_to_weave(file_id, new_lines, file_parents)

        new_ie.text_version = self.rev_id

        new_ie.text_sha1 = sha_strings(new_lines)

        new_ie.text_size = sum(map(len, new_lines))

        self.new_inv.add(new_ie)

    def _add_text_to_weave(self, file_id, new_lines, parents):

        self.weave_store.add_text(file_id, self.rev_id, new_lines, parents)

def _gen_revision_id(branch, when):

    """Return new revision-id."""

    s = '%s-%s-' % (user_email(branch), compact_date(when))

    s += hexlify(rand_bytes(8))

    return s

def merge_ancestry_lines(rev_id, ancestries):

    """Return merged ancestry lines.

    rev_id -- id of the new revision

    ancestries -- a sequence of ancestries for parent revisions,

        as newline-terminated line lists.

    """

    if len(ancestries) == 0:

        return [rev_id + '\n']

    seen = set(ancestries[0])

    ancs = ancestries[0][:]    

    for parent_ancestry in ancestries[1:]:

        for line in parent_ancestry:

            assert line[-1] == '\n'

            if line not in seen:

                ancs.append(line)

                seen.add(line)

    r = rev_id + '\n'

    assert r not in seen

    ancs.append(r)

    return ancs