1 # -*- coding: utf-8 -*-
10 logger = logging.getLogger('fnp.hazlenut.vstorage')
12 # Note: we have to set these before importing Mercurial
13 os.environ['HGENCODING'] = 'utf-8'
14 os.environ['HGMERGE'] = "internal:merge"
18 import mercurial.revlog
22 def urlquote(url, safe = '/'):
25 >>> urlquote(u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144')
26 'Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84'
28 return urllib.quote(url.replace(' ', '_').encode('utf-8', 'ignore'), safe)
34 # >>> urlunquote('Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84')
35 # u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144'
37 return unicode(urllib.unquote(url), 'utf-8', 'ignore').replace('_', ' ')
40 def find_repo_path(path):
41 """Go up the directory tree looking for a Mercurial repository (a directory containing a .hg subdirectory)."""
42 while not os.path.isdir(os.path.join(path, ".hg")):
43 old_path, path = path, os.path.dirname(path)
49 def with_working_copy_locked(func):
50 """A decorator for locking the repository when calling a method."""
52 @functools.wraps(func)
53 def wrapped(self, *args, **kwargs):
54 """Wrap the original function in locks."""
55 wlock = self.repo.wlock()
57 return func(self, *args, **kwargs)
62 def with_storage_locked(func):
63 """A decorator for locking the repository when calling a method."""
65 @functools.wraps(func)
66 def wrapped(self, *args, **kwargs):
67 """Wrap the original function in locks."""
68 lock = self.repo.lock()
70 return func(self, *args, **kwargs)
75 def guess_mime(file_name):
77 Guess file's mime type based on extension.
78 Default of text/x-wiki for files without an extension.
80 >>> guess_mime('something.txt')
82 >>> guess_mime('SomePage')
84 >>> guess_mime(u'ąęśUnicodePage')
86 >>> guess_mime('image.png')
88 >>> guess_mime('style.css')
90 >>> guess_mime('archive.tar.gz')
94 mime, encoding = mimetypes.guess_type(file_name, strict = False)
96 mime = 'archive/%s' % encoding
102 class DocumentNotFound(Exception):
106 class VersionedStorage(object):
108 Provides means of storing text pages and keeping track of their
109 change history, using Mercurial repository as the storage method.
112 def __init__(self, path, charset = None):
114 Takes the path to the directory where the pages are to be kept.
115 If the directory doen't exist, it will be created. If it's inside
116 a Mercurial repository, that repository will be used, otherwise
117 a new repository will be created in it.
120 self.charset = charset or 'utf-8'
122 if not os.path.exists(self.path):
123 os.makedirs(self.path)
124 self.repo_path = find_repo_path(self.path)
126 self.ui = mercurial.ui.ui()
128 self.ui._report_untrusted = False
129 self.ui.setconfig('ui', 'interactive', False)
131 if self.repo_path is None:
132 self.repo_path = self.path
137 self.repo_prefix = self.path[len(self.repo_path):].strip('/')
138 self.repo = mercurial.hg.repository(self.ui, self.repo_path,
142 """Close and reopen the repo, to make sure we are up to date."""
143 self.repo = mercurial.hg.repository(self.ui, self.repo_path)
145 def _file_path(self, title):
146 return os.path.join(self.path, urlquote(title, safe = ''))
148 def _title_to_file(self, title):
149 return os.path.join(self.repo_prefix, urlquote(title, safe = ''))
151 def _file_to_title(self, filename):
152 assert filename.startswith(self.repo_prefix)
153 name = filename[len(self.repo_prefix):].strip('/')
154 return urlunquote(name)
156 def __contains__(self, title):
157 return urlquote(title) in self.repo['tip']
160 return self.all_pages()
162 def merge_changes(self, changectx, repo_file, text, user, parent):
163 """Commits and merges conflicting changes in the repository."""
164 tip_node = changectx.node()
165 filectx = changectx[repo_file].filectx(parent)
166 parent_node = filectx.changectx().node()
168 self.repo.dirstate.setparents(parent_node)
169 node = self._commit([repo_file], text, user)
171 partial = lambda filename: repo_file == filename
173 # If p1 is equal to p2, there is no work to do. Even the dirstate is correct.
174 p1, p2 = self.repo[None].parents()[0], self.repo[tip_node]
179 mercurial.merge.update(self.repo, tip_node, True, False, partial)
180 msg = 'merge of edit conflict'
181 except mercurial.util.Abort:
182 msg = 'failed merge of edit conflict'
184 self.repo.dirstate.setparents(tip_node, node)
185 # Mercurial 1.1 and later need updating the merge state
187 mercurial.merge.mergestate(self.repo).mark(repo_file, "r")
188 except (AttributeError, KeyError):
192 @with_working_copy_locked
194 def save_file(self, title, file_name, author = u'', comment = u'', parent = None):
195 """Save an existing file as specified page."""
196 user = author.encode('utf-8') or u'anonymous'.encode('utf-8')
197 text = comment.encode('utf-8') or u'comment'.encode('utf-8')
199 repo_file = self._title_to_file(title)
200 file_path = self._file_path(title)
201 mercurial.util.rename(file_name, file_path)
202 changectx = self._changectx()
205 filectx_tip = changectx[repo_file]
206 current_page_rev = filectx_tip.filerev()
207 except mercurial.revlog.LookupError:
208 self.repo.add([repo_file])
209 current_page_rev = -1
211 if parent is not None and current_page_rev != parent:
212 msg = self.merge_changes(changectx, repo_file, text, user, parent)
214 text = msg.encode('utf-8')
216 self._commit([repo_file], text, user)
219 def save_data(self, title, data, **kwargs):
220 """Save data as specified page."""
222 temp_path = tempfile.mkdtemp(dir = self.path)
223 file_path = os.path.join(temp_path, 'saved')
224 f = open(file_path, "wb")
227 self.save_file(title = title, file_name = file_path, **kwargs)
238 def save_text(self, text, **kwargs):
239 """Save text as specified page, encoded to charset."""
240 self.save_data(data = text.encode(self.charset), **kwargs)
243 def _commit(self, files, text, user):
244 match = mercurial.match.exact(self.repo_path, '', list(files))
245 return self.repo.commit(match = match, text = text, user = user, force = True)
247 def page_text(self, title):
248 """Read unicode text of a page."""
249 data = self.open_page(title).read()
250 text = unicode(data, self.charset, 'replace')
253 def page_lines(self, page):
255 yield unicode(data, self.charset, 'replace')
257 @with_working_copy_locked
259 def delete_page(self, title, author = u'', comment = u''):
260 user = author.encode('utf-8') or 'anon'
261 text = comment.encode('utf-8') or 'deleted'
262 repo_file = self._title_to_file(title)
263 file_path = self._file_path(title)
268 self.repo.remove([repo_file])
269 self._commit([repo_file], text, user)
271 @with_working_copy_locked
272 def open_page(self, title):
273 if title not in self:
274 raise DocumentNotFound()
276 path = self._title_to_file(title)
277 logger.debug("Opening page %s", path)
279 return self.repo.wfile(path, 'rb')
281 logger.exception("Failed to open page %s", title)
282 raise DocumentNotFound()
284 @with_working_copy_locked
285 def page_file_meta(self, title):
286 """Get page's inode number, size and last modification time."""
288 (st_mode, st_ino, st_dev, st_nlink, st_uid, st_gid, st_size,
289 st_atime, st_mtime, st_ctime) = os.stat(self._file_path(title))
292 return st_ino, st_size, st_mtime
294 @with_working_copy_locked
295 def page_meta(self, title):
296 """Get page's revision, date, last editor and his edit comment."""
297 if not title in self:
298 raise DocumentNotFound()
300 filectx_tip = self._find_filectx(title)
301 if filectx_tip is None:
302 raise DocumentNotFound()
303 rev = filectx_tip.filerev()
304 filectx = filectx_tip.filectx(rev)
305 date = datetime.datetime.fromtimestamp(filectx.date()[0])
306 author = unicode(filectx.user(), "utf-8",
307 'replace').split('<')[0].strip()
308 comment = unicode(filectx.description(), "utf-8", 'replace')
309 return rev, date, author, comment
311 def repo_revision(self):
312 return self.repo['tip'].rev()
314 def _changectx(self):
315 return self.repo['tip']
317 def page_mime(self, title):
319 Guess page's mime type based on corresponding file name.
320 Default ot text/x-wiki for files without an extension.
322 return guess_mime(self._file_path(title))
324 def _find_filectx(self, title):
325 """Find the last revision in which the file existed."""
327 repo_file = self._title_to_file(title)
328 changectx = self._changectx()
330 while repo_file not in changectx:
333 changectx = stack.pop()
334 for parent in changectx.parents():
335 if parent != changectx:
337 return changectx[repo_file]
339 def page_history(self, title):
340 """Iterate over the page's history."""
342 filectx_tip = self._find_filectx(title)
343 if filectx_tip is None:
345 maxrev = filectx_tip.filerev()
347 for rev in range(maxrev, minrev - 1, -1):
348 filectx = filectx_tip.filectx(rev)
349 date = datetime.datetime.fromtimestamp(filectx.date()[0])
350 author = unicode(filectx.user(), "utf-8",
351 'replace').split('<')[0].strip()
352 comment = unicode(filectx.description(), "utf-8", 'replace')
353 yield {"version": rev, "date": date, "author": author, "description": comment}
355 def page_revision(self, title, rev):
356 """Get unicode contents of specified revision of the page."""
358 filectx_tip = self._find_filectx(title)
359 if filectx_tip is None:
360 raise DocumentNotFound()
362 data = filectx_tip.filectx(rev).data()
364 raise DocumentNotFound()
367 def revision_text(self, title, rev):
368 data = self.page_revision(title, rev)
369 text = unicode(data, self.charset, 'replace')
373 """Iterate over the history of entire wiki."""
375 changectx = self._changectx()
376 maxrev = changectx.rev()
378 for wiki_rev in range(maxrev, minrev - 1, -1):
379 change = self.repo.changectx(wiki_rev)
380 date = datetime.datetime.fromtimestamp(change.date()[0])
381 author = unicode(change.user(), "utf-8",
382 'replace').split('<')[0].strip()
383 comment = unicode(change.description(), "utf-8", 'replace')
384 for repo_file in change.files():
385 if repo_file.startswith(self.repo_prefix):
386 title = self._file_to_title(repo_file)
388 rev = change[repo_file].filerev()
389 except mercurial.revlog.LookupError:
391 yield title, rev, date, author, comment
394 tip = self.repo['tip']
395 """Iterate over the titles of all pages in the wiki."""
396 return [ urlunquote(filename) for filename in tip ]
398 def changed_since(self, rev):
399 """Return all pages that changed since specified repository revision."""
402 last = self.repo.lookup(int(rev))
404 for page in self.all_pages():
407 current = self.repo.lookup('tip')
408 status = self.repo.status(current, last)
409 modified, added, removed, deleted, unknown, ignored, clean = status
410 for filename in modified + added + removed + deleted:
411 if filename.startswith(self.repo_prefix):
412 yield self._file_to_title(filename)