| 1 | # -*- coding: iso-8859-1 -*- |
|---|
| 2 | # |
|---|
| 3 | # Copyright (C) 2005 Edgewall Software |
|---|
| 4 | # Copyright (C) 2006 K.S.Sreeram <sreeram@tachyontech.net> |
|---|
| 5 | # Copyright (C) 2007,2008 Lele Gaifax <lele@metapensiero.it> |
|---|
| 6 | # |
|---|
| 7 | # This software is licensed as described in the file COPYING, which |
|---|
| 8 | # you should have received as part of this distribution. The terms |
|---|
| 9 | # are also available at http://trac.edgewall.com/license.html. |
|---|
| 10 | # |
|---|
| 11 | # This software consists of voluntary contributions made by many |
|---|
| 12 | # individuals. For the exact contribution history, see the revision |
|---|
| 13 | # history and logs, available at http://projects.edgewall.com/trac/. |
|---|
| 14 | # |
|---|
| 15 | # Author: K.S.Sreeram <sreeram@tachyontech.net> |
|---|
| 16 | |
|---|
| 17 | from datetime import datetime |
|---|
| 18 | from time import mktime, strptime, timezone |
|---|
| 19 | from xml.sax import parseString |
|---|
| 20 | from xml.sax.handler import ContentHandler |
|---|
| 21 | |
|---|
| 22 | from trac.util.datefmt import utc |
|---|
| 23 | |
|---|
| 24 | class Patch(object): |
|---|
| 25 | ''' |
|---|
| 26 | Represents a single darcs changeset. |
|---|
| 27 | |
|---|
| 28 | 'time' is an int, the number of seconds since the epoch. |
|---|
| 29 | It is in the local timezone. |
|---|
| 30 | 'entries' is a list of changes performed in the current changeset. |
|---|
| 31 | The elements of 'entries' can be of the following form: |
|---|
| 32 | ('add_file', 'some/file/path') |
|---|
| 33 | ('remove_file', 'some/file/path') |
|---|
| 34 | ('modify_file', 'some/file/path') |
|---|
| 35 | ('add_directory', 'some/dir/path') |
|---|
| 36 | ('remove_directory', 'some/dir/path') |
|---|
| 37 | ('move', 'from/path', 'to/path') |
|---|
| 38 | 'move' is used for both directories and files. |
|---|
| 39 | ''' |
|---|
| 40 | def __init__(self, author, time, hash, name, comment, entries): |
|---|
| 41 | self.author = author |
|---|
| 42 | self.time = time |
|---|
| 43 | self.hash = hash |
|---|
| 44 | self.name = name |
|---|
| 45 | self.comment = comment |
|---|
| 46 | self.entries = entries |
|---|
| 47 | |
|---|
| 48 | def __repr__(self): |
|---|
| 49 | 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) |
|---|
| 50 | |
|---|
| 51 | def parse_changes(changes): |
|---|
| 52 | ''' |
|---|
| 53 | Parses the output of 'darcs changes --xml-output --summary', and |
|---|
| 54 | returns a list of 'Patch' objects. |
|---|
| 55 | ''' |
|---|
| 56 | |
|---|
| 57 | class SAXHandler(ContentHandler): |
|---|
| 58 | RESET_CONTENT_ELTS = set('name comment add_file modify_file remove_file ' |
|---|
| 59 | 'add_directory remove_directory'.split()) |
|---|
| 60 | ACTION_ELTS = set('add_file modify_file remove_file ' |
|---|
| 61 | 'add_directory remove_directory'.split()) |
|---|
| 62 | |
|---|
| 63 | def __init__(self): |
|---|
| 64 | self.patches = [] |
|---|
| 65 | self.content = [] |
|---|
| 66 | |
|---|
| 67 | def startElement(self, name, attr): |
|---|
| 68 | if name == 'patch': |
|---|
| 69 | self.author = attr['author'] |
|---|
| 70 | date = attr['date'] |
|---|
| 71 | try: |
|---|
| 72 | # 20040619130027 |
|---|
| 73 | timestamp = datetime(*strptime(date, '%Y%m%d%H%M%S')[:6]) |
|---|
| 74 | except ValueError: |
|---|
| 75 | # Old darcs patches use the form Sun Oct 20 20:01:05 EDT 2002 |
|---|
| 76 | timestamp = datetime(*strptime(date[:19] + date[-5:], '%a %b %d %H:%M:%S %Y')[:6]) |
|---|
| 77 | self.time = timestamp.replace(tzinfo=utc) |
|---|
| 78 | self.hash = attr['hash'] |
|---|
| 79 | # darcs-1 rollbacks: darcs-2 uses a different way, rollbacks are |
|---|
| 80 | # normal patches, and "inverted" is always False |
|---|
| 81 | self.inverted = attr['inverted'] == 'True' |
|---|
| 82 | self.name = '' |
|---|
| 83 | self.comment = '' |
|---|
| 84 | self.entries = [] |
|---|
| 85 | elif name == 'move': |
|---|
| 86 | self.entries.append(('move',attr['from'],attr['to'])) |
|---|
| 87 | elif name in self.RESET_CONTENT_ELTS: |
|---|
| 88 | self.content = [] |
|---|
| 89 | |
|---|
| 90 | def characters(self, data): |
|---|
| 91 | self.content.append(data) |
|---|
| 92 | |
|---|
| 93 | def endElement(self, name): |
|---|
| 94 | if name == 'name': |
|---|
| 95 | self.name = (self.inverted and 'UNDO: ' or '') + ''.join(self.content) |
|---|
| 96 | elif name == 'comment': |
|---|
| 97 | self.comment = ''.join(self.content) |
|---|
| 98 | elif name in self.ACTION_ELTS: |
|---|
| 99 | path = ''.join(self.content).strip() |
|---|
| 100 | self.entries.append((name,path)) |
|---|
| 101 | elif name == 'patch': |
|---|
| 102 | patch = Patch(self.author, self.time, self.hash, |
|---|
| 103 | self.name, self.comment, self.entries) |
|---|
| 104 | self.patches.append(patch) |
|---|
| 105 | |
|---|
| 106 | handler = SAXHandler() |
|---|
| 107 | |
|---|
| 108 | def fixup_badchars(s, map): |
|---|
| 109 | if not map: |
|---|
| 110 | return s |
|---|
| 111 | |
|---|
| 112 | ret = [map.get(c, c) for c in s] |
|---|
| 113 | return "".join(ret) |
|---|
| 114 | |
|---|
| 115 | replace_badchars = { |
|---|
| 116 | '\xc1': 'Á', '\xc9': 'É', '\xcd': 'Í', |
|---|
| 117 | '\xd3': 'Ó', '\xd6': 'Ö', '\xd5': 'Ő', |
|---|
| 118 | '\xda': 'Ú', '\xdc': 'Ü', '\xdb': 'Ű', |
|---|
| 119 | '\xe1': 'á', '\xe9': 'é', '\xed': 'í', |
|---|
| 120 | '\xf3': 'ó', '\xf6': 'ö', '\xf5': 'ő', |
|---|
| 121 | '\xfa': 'ú', '\xfc': 'ü', '\xfb': 'ű', |
|---|
| 122 | '\xf1': 'ñ', '\xdf': 'ß', '\xe5': 'å', |
|---|
| 123 | } |
|---|
| 124 | |
|---|
| 125 | cleaned = fixup_badchars(changes, replace_badchars) |
|---|
| 126 | parseString(cleaned, handler) |
|---|
| 127 | return handler.patches |
|---|
