1 # -*- coding: utf-8 -*-
8 # Note: we have to set these before importing Mercurial
9 os.environ['HGENCODING'] = 'utf-8'
10 os.environ['HGMERGE'] = "internal:merge"
14 import mercurial.revlog
18 def urlquote(url, safe='/'):
21 >>> urlquote(u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144')
22 'Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84'
24 return urllib.quote(url.replace(' ', '_').encode('utf-8', 'ignore'), safe)
29 # >>> urlunquote('Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84')
30 # u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144'
32 return unicode(urllib.unquote(url), 'utf-8', 'ignore').replace('_', ' ')
34 def find_repo_path(path):
35 """Go up the directory tree looking for a Mercurial repository (a directory containing a .hg subdirectory)."""
36 while not os.path.isdir(os.path.join(path, ".hg")):
37 old_path, path = path, os.path.dirname(path)
42 def locked_repo(func):
43 """A decorator for locking the repository when calling a method."""
45 def new_func(self, *args, **kwargs):
46 """Wrap the original function in locks."""
48 wlock = self.repo.wlock()
49 lock = self.repo.lock()
51 func(self, *args, **kwargs)
59 class DocumentNotFound(Exception):
63 class VersionedStorage(object):
65 Provides means of storing text pages and keeping track of their
66 change history, using Mercurial repository as the storage method.
69 def __init__(self, path, charset=None):
71 Takes the path to the directory where the pages are to be kept.
72 If the directory doen't exist, it will be created. If it's inside
73 a Mercurial repository, that repository will be used, otherwise
74 a new repository will be created in it.
77 self.charset = charset or 'utf-8'
79 if not os.path.exists(self.path):
80 os.makedirs(self.path)
81 self.repo_path = find_repo_path(self.path)
83 self.ui = mercurial.ui.ui(report_untrusted=False,
84 interactive=False, quiet=True)
86 # Mercurial 1.3 changed the way we setup the ui object.
87 self.ui = mercurial.ui.ui()
89 self.ui._report_untrusted = False
90 self.ui.setconfig('ui', 'interactive', False)
91 if self.repo_path is None:
92 self.repo_path = self.path
96 self.repo_prefix = self.path[len(self.repo_path):].strip('/')
97 self.repo = mercurial.hg.repository(self.ui, self.repo_path,
101 """Close and reopen the repo, to make sure we are up to date."""
103 self.repo = mercurial.hg.repository(self.ui, self.repo_path)
105 def _file_path(self, title):
106 return os.path.join(self.path, urlquote(title, safe=''))
108 def _title_to_file(self, title):
109 return os.path.join(self.repo_prefix, urlquote(title, safe=''))
111 def _file_to_title(self, filename):
112 assert filename.startswith(self.repo_prefix)
113 name = filename[len(self.repo_prefix):].strip('/')
114 return urlunquote(name)
116 def __contains__(self, title):
117 return os.path.exists(self._file_path(title))
120 return self.all_pages()
122 def merge_changes(self, changectx, repo_file, text, user, parent):
123 """Commits and merges conflicting changes in the repository."""
124 tip_node = changectx.node()
125 filectx = changectx[repo_file].filectx(parent)
126 parent_node = filectx.changectx().node()
128 self.repo.dirstate.setparents(parent_node)
129 node = self._commit([repo_file], text, user)
131 partial = lambda filename: repo_file == filename
133 # If p1 is equal to p2, there is no work to do. Even the dirstate is correct.
134 p1, p2 = self.repo[None].parents()[0], self.repo[tip_node]
138 # TODO: Check if merge was successful
139 mercurial.merge.update(self.repo, tip_node, True, False, partial)
141 self.repo.dirstate.setparents(tip_node, node)
142 # Mercurial 1.1 and later need updating the merge state
144 mercurial.merge.mergestate(self.repo).mark(repo_file, "r")
145 except (AttributeError, KeyError):
147 return u'merge of edit conflict'
150 def save_file(self, title, file_name, author=u'', comment=u'', parent=None):
151 """Save an existing file as specified page."""
153 user = author.encode('utf-8') or u'anon'.encode('utf-8')
154 text = comment.encode('utf-8') or u'comment'.encode('utf-8')
155 repo_file = self._title_to_file(title)
156 file_path = self._file_path(title)
157 mercurial.util.rename(file_name, file_path)
158 changectx = self._changectx()
160 filectx_tip = changectx[repo_file]
161 current_page_rev = filectx_tip.filerev()
162 except mercurial.revlog.LookupError:
163 self.repo.add([repo_file])
164 current_page_rev = -1
165 if parent is not None and current_page_rev != parent:
166 msg = self.merge_changes(changectx, repo_file, text, user, parent)
168 text = msg.encode('utf-8')
169 self._commit([repo_file], text, user)
172 def _commit(self, files, text, user):
174 return self.repo.commit(files=files, text=text, user=user,
175 force=True, empty_ok=True)
177 # Mercurial 1.3 doesn't accept empty_ok or files parameter
178 match = mercurial.match.exact(self.repo_path, '', list(files))
179 return self.repo.commit(match=match, text=text, user=user,
183 def save_data(self, title, data, author=u'', comment=u'', parent=None):
184 """Save data as specified page."""
187 temp_path = tempfile.mkdtemp(dir=self.path)
188 file_path = os.path.join(temp_path, 'saved')
189 f = open(file_path, "wb")
192 self.save_file(title, file_path, author, comment, parent)
203 def save_text(self, title, text, author=u'', comment=u'', parent=None):
204 """Save text as specified page, encoded to charset."""
206 data = text.encode(self.charset)
207 self.save_data(title, data, author, comment, parent)
209 def page_text(self, title):
210 """Read unicode text of a page."""
212 data = self.open_page(title).read()
213 text = unicode(data, self.charset, 'replace')
216 def page_lines(self, page):
218 yield unicode(data, self.charset, 'replace')
221 def delete_page(self, title, author=u'', comment=u''):
222 user = author.encode('utf-8') or 'anon'
223 text = comment.encode('utf-8') or 'deleted'
224 repo_file = self._title_to_file(title)
225 file_path = self._file_path(title)
230 self.repo.remove([repo_file])
231 self._commit([repo_file], text, user)
233 def open_page(self, title):
235 return open(self._file_path(title), "rb")
237 raise DocumentNotFound()
239 def page_file_meta(self, title):
240 """Get page's inode number, size and last modification time."""
243 (st_mode, st_ino, st_dev, st_nlink, st_uid, st_gid, st_size,
244 st_atime, st_mtime, st_ctime) = os.stat(self._file_path(title))
247 return st_ino, st_size, st_mtime
249 def page_meta(self, title):
250 """Get page's revision, date, last editor and his edit comment."""
252 filectx_tip = self._find_filectx(title)
253 if filectx_tip is None:
254 raise DocumentNotFound()
255 #return -1, None, u'', u''
256 rev = filectx_tip.filerev()
257 filectx = filectx_tip.filectx(rev)
258 date = datetime.datetime.fromtimestamp(filectx.date()[0])
259 author = unicode(filectx.user(), "utf-8",
260 'replace').split('<')[0].strip()
261 comment = unicode(filectx.description(), "utf-8", 'replace')
262 return rev, date, author, comment
264 def repo_revision(self):
265 return self._changectx().rev()
267 def page_mime(self, title):
269 Guess page's mime type ased on corresponding file name.
270 Default ot text/x-wiki for files without an extension.
272 # >>> page_mime('something.txt')
274 # >>> page_mime('SomePage')
276 # >>> page_mime(u'ąęśUnicodePage')
278 # >>> page_mime('image.png')
280 # >>> page_mime('style.css')
282 # >>> page_mime('archive.tar.gz')
286 addr = self._file_path(title)
287 mime, encoding = mimetypes.guess_type(addr, strict=False)
289 mime = 'archive/%s' % encoding
294 def _changectx(self):
295 """Get the changectx of the tip."""
297 # This is for Mercurial 1.0
298 return self.repo.changectx()
300 # Mercurial 1.3 (and possibly earlier) needs an argument
301 return self.repo.changectx('tip')
303 def _find_filectx(self, title):
304 """Find the last revision in which the file existed."""
306 repo_file = self._title_to_file(title)
307 changectx = self._changectx()
309 while repo_file not in changectx:
312 changectx = stack.pop()
313 for parent in changectx.parents():
314 if parent != changectx:
316 return changectx[repo_file]
318 def page_history(self, title):
319 """Iterate over the page's history."""
321 filectx_tip = self._find_filectx(title)
322 if filectx_tip is None:
324 maxrev = filectx_tip.filerev()
326 for rev in range(maxrev, minrev-1, -1):
327 filectx = filectx_tip.filectx(rev)
328 date = datetime.datetime.fromtimestamp(filectx.date()[0])
329 author = unicode(filectx.user(), "utf-8",
330 'replace').split('<')[0].strip()
331 comment = unicode(filectx.description(), "utf-8", 'replace')
332 yield rev, date, author, comment
334 def page_revision(self, title, rev):
335 """Get unicode contents of specified revision of the page."""
337 filectx_tip = self._find_filectx(title)
338 if filectx_tip is None:
339 raise DocumentNotFound()
341 data = filectx_tip.filectx(rev).data()
343 raise DocumentNotFound()
346 def revision_text(self, title, rev):
347 data = self.page_revision(title, rev)
348 text = unicode(data, self.charset, 'replace')
352 """Iterate over the history of entire wiki."""
354 changectx = self._changectx()
355 maxrev = changectx.rev()
357 for wiki_rev in range(maxrev, minrev-1, -1):
358 change = self.repo.changectx(wiki_rev)
359 date = datetime.datetime.fromtimestamp(change.date()[0])
360 author = unicode(change.user(), "utf-8",
361 'replace').split('<')[0].strip()
362 comment = unicode(change.description(), "utf-8", 'replace')
363 for repo_file in change.files():
364 if repo_file.startswith(self.repo_prefix):
365 title = self._file_to_title(repo_file)
367 rev = change[repo_file].filerev()
368 except mercurial.revlog.LookupError:
370 yield title, rev, date, author, comment
373 """Iterate over the titles of all pages in the wiki."""
375 for filename in os.listdir(self.path):
376 if (os.path.isfile(os.path.join(self.path, filename))
377 and not filename.startswith('.')):
378 yield urlunquote(filename)
380 def changed_since(self, rev):
381 """Return all pages that changed since specified repository revision."""
384 last = self.repo.lookup(int(rev))
386 for page in self.all_pages():
389 current = self.repo.lookup('tip')
390 status = self.repo.status(current, last)
391 modified, added, removed, deleted, unknown, ignored, clean = status
392 for filename in modified+added+removed+deleted:
393 if filename.startswith(self.repo_prefix):
394 yield self._file_to_title(filename)