source: tracdarcs/tracdarcs/changesparser.py @ 176

# -*- coding: utf-8 -*-
#
# Copyright (C) 2005 Edgewall Software
# Copyright (C) 2006 K.S.Sreeram <sreeram@tachyontech.net>
# Copyright (C) 2007,2008,2009 Lele Gaifax <lele@metapensiero.it>
#
# This software is licensed as described in the file COPYING, which
# you should have received as part of this distribution. The terms
# are also available at http://trac.edgewall.com/license.html.
#
# This software consists of voluntary contributions made by many
# individuals. For the exact contribution history, see the revision
# history and logs, available at http://projects.edgewall.com/trac/.
#
# Author: K.S.Sreeram <sreeram@tachyontech.net>

from datetime import datetime
from re import compile
from time import mktime, strptime, timezone
from xml.sax import parseString
from xml.sax.handler import ContentHandler
from xml.sax._exceptions import SAXParseException

from trac.util.datefmt import utc

# Strip away darcs 2.1 junk: hopefully this regex is strict enough to
# not obliterate useful info...
IGNORE_THIS = compile('^Ignore-this: [a-f\\d]+\\n?')

# Filter out bad chars that can cause the XML parser to give up in despair.
# (Thanks to lelit of the tailor project and ndurner and warner for this hack.)
BADCHARS = "".join([chr(i) for i in range(0x0a) +
                    [0x0b, 0x0c] + range(0x0e, 0x20) + range(0x7f,0x100)])

class Patch(object):
    '''
    Represents a single darcs changeset.

    'time' is an int, the number of seconds since the epoch.
    It is in the local timezone.
    'entries' is a list of changes performed in the current changeset.
    The elements of 'entries' can be of the following form:
    ('add_file', 'some/file/path')
    ('remove_file', 'some/file/path')
    ('modify_file', 'some/file/path')
    ('add_directory', 'some/dir/path')
    ('remove_directory', 'some/dir/path')
    ('move', 'from/path', 'to/path')
    'move' is used for both directories and files.
    '''
    def __init__(self, author, time, hash, name, comment, entries):
        self.author = author
        self.time = time
        self.hash = hash
        self.name = name
        self.comment = comment
        self.entries = entries

    def __repr__(self):
        return "author: %s, time: %s, hash: %s, name: %s, comment: %s, entries: %s" % (self.author, self.time, self.hash, self.name, self.comment, self.entries)

def parse_changes(changes, log):
    '''
    Parses the output of 'darcs changes --xml-output --summary', and
    returns a list of 'Patch' objects.
    '''

    class SAXHandler(ContentHandler):
        RESET_CONTENT_ELTS = set('name comment add_file modify_file remove_file '
                                 'add_directory remove_directory'.split())
        ACTION_ELTS = set('add_file modify_file remove_file '
                          'add_directory remove_directory'.split())

        def __init__(self):
            self.patches = []
            self.content = []

        def startElement(self, name, attr):
            if name == 'patch':
                self.author = attr['author']
                date = attr['date']
                try:
                    # 20040619130027
                    timestamp = datetime(*strptime(date, '%Y%m%d%H%M%S')[:6])
                except ValueError:
                    # Old darcs patches use the form Sun Oct 20 20:01:05 EDT 2002
                    timestamp = datetime(*strptime(date[:19] + date[-5:], '%a %b %d %H:%M:%S %Y')[:6])
                self.time = timestamp.replace(tzinfo=utc)
                # Strip useless trailing .gz
                phash = attr['hash']
                if phash.endswith('.gz'):
                    phash = phash[:-3]
                self.hash = phash
                # darcs-1 rollbacks: darcs-2 uses a different way, rollbacks are
                # normal patches, and "inverted" is always False
                self.inverted = attr['inverted'] == 'True'
                self.name = ''
                self.comment = ''
                self.entries = []
            elif name == 'move':
                self.entries.append(('move',attr['from'],attr['to']))
            elif name in self.RESET_CONTENT_ELTS:
                self.content = []

        def characters(self, data):
            self.content.append(data)

        def endElement(self, name):
            if name == 'name':
                self.name = (self.inverted and 'UNDO: ' or '') + ''.join(self.content)
            elif name == 'comment':
                content = ''.join(self.content)
                self.comment = IGNORE_THIS.sub('', content)
            elif name in self.ACTION_ELTS:
                path = ''.join(self.content).strip()
                self.entries.append((name,path))
            elif name == 'patch':
                patch = Patch(self.author, self.time, self.hash,
                              self.name, self.comment, self.entries)
                self.patches.append(patch)

    from string import maketrans
    tt = maketrans(BADCHARS, "-"*len(BADCHARS))
    changes = changes.translate(tt)

    handler = SAXHandler()
    try:
        parseString(changes, handler)
    except SAXParseException, e:
        lines = changes.splitlines()
        line = lines[e._linenum-1]
        log.critical("Cannot complete the parse of the XML output due to a SAX Error, "
                     "on the changeset %d: %s", len(handler.patches), e)
        log.warning("%s", line)
        if e._colnum:
            log.warning('-'*(e._colnum-1)+'^'*5)
    return handler.patches