~bzr-pqm/bzr/bzr.dev

#! /usr/bin/env python

# -*- coding: UTF-8 -*-

# 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

from sets import Set

from trace import mutter

from errors import BzrError

def diff_trees(old_tree, new_tree):

    """Compute diff between two trees.

    They may be in different branches and may be working or historical

    trees.

    Yields a sequence of (state, id, old_name, new_name, kind).

    Each filename and each id is listed only once.

    """

    ## TODO: Compare files before diffing; only mention those that have changed

    ## TODO: Set nice names in the headers, maybe include diffstat

    ## TODO: Perhaps make this a generator rather than using

    ## a callback object?

    ## TODO: Allow specifying a list of files to compare, rather than

    ## doing the whole tree?  (Not urgent.)

    ## TODO: Allow diffing any two inventories, not just the

    ## current one against one.  We mgiht need to specify two

    ## stores to look for the files if diffing two branches.  That

    ## might imply this shouldn't be primarily a Branch method.

    ## XXX: This doesn't report on unknown files; that can be done

    ## from a separate method.

    old_it = old_tree.list_files()

    new_it = new_tree.list_files()

    def next(it):

        try:

            return it.next()

        except StopIteration:

            return None

    old_item = next(old_it)

    new_item = next(new_it)

    # We step through the two sorted iterators in parallel, trying to

    # keep them lined up.

    while (old_item != None) or (new_item != None):

        # OK, we still have some remaining on both, but they may be

        # out of step.        

        if old_item != None:

            old_name, old_class, old_kind, old_id = old_item

        else:

            old_name = None

        if new_item != None:

            new_name, new_class, new_kind, new_id = new_item

        else:

            new_name = None

        mutter("   diff pairwise %r" % (old_item,))

        mutter("                 %r" % (new_item,))

        if old_item:

            # can't handle the old tree being a WorkingTree

            assert old_class == 'V'

        if new_item and (new_class != 'V'):

            yield new_class, None, None, new_name, new_kind

            new_item = next(new_it)

        elif (not new_item) or (old_item and (old_name < new_name)):

            mutter("     extra entry in old-tree sequence")

            if new_tree.has_id(old_id):

                # will be mentioned as renamed under new name

                pass

            else:

                yield 'D', old_id, old_name, None, old_kind

            old_item = next(old_it)

        elif (not old_item) or (new_item and (new_name < old_name)):

            mutter("     extra entry in new-tree sequence")

            if old_tree.has_id(new_id):

                yield 'R', new_id, old_tree.id2path(new_id), new_name, new_kind

            else:

                yield 'A', new_id, None, new_name, new_kind

            new_item = next(new_it)

        elif old_id != new_id:

            assert old_name == new_name

            # both trees have a file of this name, but it is not the

            # same file.  in other words, the old filename has been

            # overwritten by either a newly-added or a renamed file.

            # (should we return something about the overwritten file?)

            if old_tree.has_id(new_id):

                # renaming, overlying a deleted file

                yield 'R', new_id, old_tree.id2path(new_id), new_name, new_kind

            else:

                yield 'A', new_id, None, new_name, new_kind

            new_item = next(new_it)

            old_item = next(old_it)

        else:

            assert old_id == new_id

            assert old_id != None

            assert old_name == new_name

            assert old_kind == new_kind

            if old_kind == 'directory':

                yield '.', new_id, old_name, new_name, new_kind

            elif old_tree.get_file_size(old_id) != new_tree.get_file_size(old_id):

                mutter("    file size has changed, must be different")

                yield 'M', new_id, old_name, new_name, new_kind

            elif old_tree.get_file_sha1(old_id) == new_tree.get_file_sha1(old_id):

                mutter("      SHA1 indicates they're identical")

                ## assert compare_files(old_tree.get_file(i), new_tree.get_file(i))

                yield '.', new_id, old_name, new_name, new_kind

            else:

                mutter("      quick compare shows different")

                yield 'M', new_id, old_name, new_name, new_kind

            new_item = next(new_it)

            old_item = next(old_it)

def show_diff(b, revision, file_list):

    import difflib, sys, types

    if revision == None:

        old_tree = b.basis_tree()

    else:

        old_tree = b.revision_tree(b.lookup_revision(revision))

    new_tree = b.working_tree()

    # TODO: Options to control putting on a prefix or suffix, perhaps as a format string

    old_label = ''

    new_label = ''

    DEVNULL = '/dev/null'

    # Windows users, don't panic about this filename -- it is a

    # special signal to GNU patch that the file should be created or

    # deleted respectively.

    # TODO: Generation of pseudo-diffs for added/deleted files could

    # be usefully made into a much faster special case.

    # TODO: Better to return them in sorted order I think.

    if file_list:

        file_list = [b.relpath(f) for f in file_list]

    # FIXME: If given a file list, compare only those files rather

    # than comparing everything and then throwing stuff away.

    for file_state, fid, old_name, new_name, kind in diff_trees(old_tree, new_tree):

        if file_list and (new_name not in file_list):

            continue

        # Don't show this by default; maybe do it if an option is passed

        # idlabel = '      {%s}' % fid

        idlabel = ''

        # FIXME: Something about the diff format makes patch unhappy

        # with newly-added files.

        def diffit(oldlines, newlines, **kw):

            # FIXME: difflib is wrong if there is no trailing newline.

            # The syntax used by patch seems to be "\ No newline at

            # end of file" following the last diff line from that

            # file.  This is not trivial to insert into the

            # unified_diff output and it might be better to just fix

            # or replace that function.

            # In the meantime we at least make sure the patch isn't

            # mangled.

            # Special workaround for Python2.3, where difflib fails if

            # both sequences are empty.

            if not oldlines and not newlines:

                return

            nonl = False

            if oldlines and (oldlines[-1][-1] != '\n'):

                oldlines[-1] += '\n'

                nonl = True

            if newlines and (newlines[-1][-1] != '\n'):

                newlines[-1] += '\n'

                nonl = True

            ud = difflib.unified_diff(oldlines, newlines, **kw)

            sys.stdout.writelines(ud)

            if nonl:

                print "\\ No newline at end of file"

            sys.stdout.write('\n')

        if file_state in ['.', '?', 'I']:

            continue

        elif file_state == 'A':

            print '*** added %s %r' % (kind, new_name)

            if kind == 'file':

                diffit([],

                       new_tree.get_file(fid).readlines(),

                       fromfile=DEVNULL,

                       tofile=new_label + new_name + idlabel)

        elif file_state == 'D':

            assert isinstance(old_name, types.StringTypes)

            print '*** deleted %s %r' % (kind, old_name)

            if kind == 'file':

                diffit(old_tree.get_file(fid).readlines(), [],

                       fromfile=old_label + old_name + idlabel,

                       tofile=DEVNULL)

        elif file_state in ['M', 'R']:

            if file_state == 'M':

                assert kind == 'file'

                assert old_name == new_name

                print '*** modified %s %r' % (kind, new_name)

            elif file_state == 'R':

                print '*** renamed %s %r => %r' % (kind, old_name, new_name)

            if kind == 'file':

                diffit(old_tree.get_file(fid).readlines(),

                       new_tree.get_file(fid).readlines(),

                       fromfile=old_label + old_name + idlabel,

                       tofile=new_label + new_name)

        else:

            raise BzrError("can't represent state %s {%s}" % (file_state, fid))

class TreeDelta:

    """Describes changes from one tree to another.

    Contains four lists:

    added

        (path, id)

    removed

        (path, id)

    renamed

        (oldpath, newpath, id)

    modified

        (path, id)

    A path may occur in more than one list if it was e.g. deleted

    under an old id and renamed into place in a new id.

    Files are listed in either modified or renamed, not both.  In

    other words, renamed files may also be modified.

    """

    def __init__(self):

        self.added = []

        self.removed = []

        self.renamed = []

        self.modified = []

def compare_inventories(old_inv, new_inv):

    """Return a TreeDelta object describing changes between inventories.

    This only describes changes in the shape of the tree, not the

    actual texts.

    This is an alternative to diff_trees() and should probably

    eventually replace it.

    """

    old_ids = old_inv.id_set()

    new_ids = new_inv.id_set()

    delta = TreeDelta()

    delta.removed = [(old_inv.id2path(fid), fid) for fid in (old_ids - new_ids)]

    delta.removed.sort()

    delta.added = [(new_inv.id2path(fid), fid) for fid in (new_ids - old_ids)]

    delta.added.sort()

    for fid in old_ids & new_ids:

        old_ie = old_inv[fid]

        new_ie = new_inv[fid]

        old_path = old_inv.id2path(fid)

        new_path = new_inv.id2path(fid)

        if old_path != new_path:

            delta.renamed.append((old_path, new_path, fid))

        elif old_ie.text_sha1 != new_ie.text_sha1:

            delta.modified.append((new_path, fid))

    delta.modified.sort()

    delta.renamed.sort()    

    return delta