ping server every 10 minutes
[redakcja.git] / lib / vstorage.py
index b23234a..b4acaf3 100644 (file)
@@ -1,9 +1,17 @@
 # -*- coding: utf-8 -*-
 # -*- coding: utf-8 -*-
+#
+# This file is part of FNP-Redakcja, licensed under GNU Affero GPLv3 or later.
+# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
+#
 import os
 import tempfile
 import datetime
 import mimetypes
 import urllib
 import os
 import tempfile
 import datetime
 import mimetypes
 import urllib
+import functools
+
+import logging
+logger = logging.getLogger('fnp.hazlenut.vstorage')
 
 # Note: we have to set these before importing Mercurial
 os.environ['HGENCODING'] = 'utf-8'
 
 # Note: we have to set these before importing Mercurial
 os.environ['HGENCODING'] = 'utf-8'
@@ -15,9 +23,9 @@ import mercurial.revlog
 import mercurial.util
 
 
 import mercurial.util
 
 
-def urlquote(url, safe = '/'):
-    """Quotes URL 
-    
+def urlquote(url, safe='/'):
+    """Quotes URL
+
     >>> urlquote(u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144')
     'Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84'
     """
     >>> urlquote(u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144')
     'Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84'
     """
@@ -25,8 +33,8 @@ def urlquote(url, safe = '/'):
 
 
 def urlunquote(url):
 
 
 def urlunquote(url):
-    """Unqotes URL 
-    
+    """Unqotes URL
+
     # >>> urlunquote('Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84')
     # u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144'
     """
     # >>> urlunquote('Za%C5%BC%C3%B3%C5%82%C4%87_g%C4%99%C5%9Bl%C4%85_ja%C5%BA%C5%84')
     # u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144'
     """
@@ -42,27 +50,38 @@ def find_repo_path(path):
     return path
 
 
     return path
 
 
-def locked_repo(func):
+def with_working_copy_locked(func):
     """A decorator for locking the repository when calling a method."""
 
     """A decorator for locking the repository when calling a method."""
 
-    def new_func(self, *args, **kwargs):
+    @functools.wraps(func)
+    def wrapped(self, *args, **kwargs):
         """Wrap the original function in locks."""
         """Wrap the original function in locks."""
-
         wlock = self.repo.wlock()
         wlock = self.repo.wlock()
-        lock = self.repo.lock()
         try:
         try:
-            func(self, *args, **kwargs)
+            return func(self, *args, **kwargs)
         finally:
         finally:
-            lock.release()
             wlock.release()
             wlock.release()
+    return wrapped
 
 
-    return new_func
+
+def with_storage_locked(func):
+    """A decorator for locking the repository when calling a method."""
+
+    @functools.wraps(func)
+    def wrapped(self, *args, **kwargs):
+        """Wrap the original function in locks."""
+        lock = self.repo.lock()
+        try:
+            return func(self, *args, **kwargs)
+        finally:
+            lock.release()
+    return wrapped
 
 
 def guess_mime(file_name):
     """
     Guess file's mime type based on extension.
 
 
 def guess_mime(file_name):
     """
     Guess file's mime type based on extension.
-    Default ot text/x-wiki for files without an extension.
+    Default of text/x-wiki for files without an extension.
 
     >>> guess_mime('something.txt')
     'text/plain'
 
     >>> guess_mime('something.txt')
     'text/plain'
@@ -78,7 +97,7 @@ def guess_mime(file_name):
     'archive/gzip'
     """
 
     'archive/gzip'
     """
 
-    mime, encoding = mimetypes.guess_type(file_name, strict = False)
+    mime, encoding = mimetypes.guess_type(file_name, strict=False)
     if encoding:
         mime = 'archive/%s' % encoding
     if mime is None:
     if encoding:
         mime = 'archive/%s' % encoding
     if mime is None:
@@ -96,7 +115,7 @@ class VersionedStorage(object):
     change history, using Mercurial repository as the storage method.
     """
 
     change history, using Mercurial repository as the storage method.
     """
 
-    def __init__(self, path, charset = None):
+    def __init__(self, path, charset=None):
         """
         Takes the path to the directory where the pages are to be kept.
         If the directory doen't exist, it will be created. If it's inside
         """
         Takes the path to the directory where the pages are to be kept.
         If the directory doen't exist, it will be created. If it's inside
@@ -109,34 +128,31 @@ class VersionedStorage(object):
         if not os.path.exists(self.path):
             os.makedirs(self.path)
         self.repo_path = find_repo_path(self.path)
         if not os.path.exists(self.path):
             os.makedirs(self.path)
         self.repo_path = find_repo_path(self.path)
-        try:
-            self.ui = mercurial.ui.ui(report_untrusted = False,
-                                      interactive = False, quiet = True)
-        except TypeError:
-            # Mercurial 1.3 changed the way we setup the ui object.
-            self.ui = mercurial.ui.ui()
-            self.ui.quiet = True
-            self.ui._report_untrusted = False
-            self.ui.setconfig('ui', 'interactive', False)
+
+        self.ui = mercurial.ui.ui()
+        self.ui.quiet = True
+        self.ui._report_untrusted = False
+        self.ui.setconfig('ui', 'interactive', False)
+
         if self.repo_path is None:
             self.repo_path = self.path
             create = True
         else:
             create = False
         if self.repo_path is None:
             self.repo_path = self.path
             create = True
         else:
             create = False
+
         self.repo_prefix = self.path[len(self.repo_path):].strip('/')
         self.repo = mercurial.hg.repository(self.ui, self.repo_path,
         self.repo_prefix = self.path[len(self.repo_path):].strip('/')
         self.repo = mercurial.hg.repository(self.ui, self.repo_path,
-                                            create = create)
+                                            create=create)
 
     def reopen(self):
         """Close and reopen the repo, to make sure we are up to date."""
 
     def reopen(self):
         """Close and reopen the repo, to make sure we are up to date."""
-
         self.repo = mercurial.hg.repository(self.ui, self.repo_path)
 
     def _file_path(self, title):
         self.repo = mercurial.hg.repository(self.ui, self.repo_path)
 
     def _file_path(self, title):
-        return os.path.join(self.path, urlquote(title, safe = ''))
+        return os.path.join(self.path, urlquote(title, safe=''))
 
     def _title_to_file(self, title):
 
     def _title_to_file(self, title):
-        return os.path.join(self.repo_prefix, urlquote(title, safe = ''))
+        return os.path.join(self.repo_prefix, urlquote(title, safe=''))
 
     def _file_to_title(self, filename):
         assert filename.startswith(self.repo_prefix)
 
     def _file_to_title(self, filename):
         assert filename.startswith(self.repo_prefix)
@@ -144,8 +160,7 @@ class VersionedStorage(object):
         return urlunquote(name)
 
     def __contains__(self, title):
         return urlunquote(name)
 
     def __contains__(self, title):
-        print "Checking ", title
-        return urlquote(title) in self.repo.dirstate
+        return urlquote(title) in self.repo['tip']
 
     def __iter__(self):
         return self.all_pages()
 
     def __iter__(self):
         return self.all_pages()
@@ -180,50 +195,44 @@ class VersionedStorage(object):
             pass
         return msg
 
             pass
         return msg
 
-    @locked_repo
-    def save_file(self, title, file_name, author = u'', comment = u'', parent = None):
+    @with_working_copy_locked
+    @with_storage_locked
+    def save_file(self, title, file_name, **kwargs):
         """Save an existing file as specified page."""
 
         """Save an existing file as specified page."""
 
-        user = author.encode('utf-8') or u'anon'.encode('utf-8')
-        text = comment.encode('utf-8') or u'comment'.encode('utf-8')
+        author = kwargs.get('author', u'anonymous').encode('utf-8')
+        comment = kwargs.get('comment', u'Empty comment.').encode('utf-8')
+        parent = kwargs.get('parent', None)
+
         repo_file = self._title_to_file(title)
         file_path = self._file_path(title)
         mercurial.util.rename(file_name, file_path)
         changectx = self._changectx()
         repo_file = self._title_to_file(title)
         file_path = self._file_path(title)
         mercurial.util.rename(file_name, file_path)
         changectx = self._changectx()
+
         try:
             filectx_tip = changectx[repo_file]
             current_page_rev = filectx_tip.filerev()
         except mercurial.revlog.LookupError:
             self.repo.add([repo_file])
             current_page_rev = -1
         try:
             filectx_tip = changectx[repo_file]
             current_page_rev = filectx_tip.filerev()
         except mercurial.revlog.LookupError:
             self.repo.add([repo_file])
             current_page_rev = -1
-        if parent is not None and current_page_rev != parent:
-            msg = self.merge_changes(changectx, repo_file, text, user, parent)
-            user = '<wiki>'
-            text = msg.encode('utf-8')
-        self._commit([repo_file], text, user)
 
 
+        if parent is not None and current_page_rev != parent:
+            msg = self.merge_changes(changectx, repo_file, comment, author, parent)
+            author = '<wiki>'
+            comment = msg.encode('utf-8')
 
 
-    def _commit(self, files, text, user):
-        try:
-            return self.repo.commit(files = files, text = text, user = user,
-                                    force = True, empty_ok = True)
-        except TypeError:
-            # Mercurial 1.3 doesn't accept empty_ok or files parameter
-            match = mercurial.match.exact(self.repo_path, '', list(files))
-            return self.repo.commit(match = match, text = text, user = user,
-                                    force = True)
-
+        self._commit([repo_file], comment, author)
 
 
-    def save_data(self, title, data, author = u'', comment = u'', parent = None):
+    def save_data(self, title, data, **kwargs):
         """Save data as specified page."""
         """Save data as specified page."""
-
         try:
         try:
-            temp_path = tempfile.mkdtemp(dir = self.path)
+            temp_path = tempfile.mkdtemp(dir=self.path)
             file_path = os.path.join(temp_path, 'saved')
             f = open(file_path, "wb")
             f.write(data)
             f.close()
             file_path = os.path.join(temp_path, 'saved')
             f = open(file_path, "wb")
             f.write(data)
             f.close()
-            self.save_file(title, file_path, author, comment, parent)
+
+            return self.save_file(title=title, file_name=file_path, **kwargs)
         finally:
             try:
                 os.unlink(file_path)
         finally:
             try:
                 os.unlink(file_path)
@@ -234,25 +243,18 @@ class VersionedStorage(object):
             except OSError:
                 pass
 
             except OSError:
                 pass
 
-    def save_text(self, title, text, author = u'', comment = u'', parent = None):
+    def save_text(self, **kwargs):
         """Save text as specified page, encoded to charset."""
         """Save text as specified page, encoded to charset."""
+        text = kwargs.pop('text')
+        return self.save_data(data=text.encode(self.charset), **kwargs)
 
 
-        data = text.encode(self.charset)
-        self.save_data(title, data, author, comment, parent)
-
-    def page_text(self, title):
-        """Read unicode text of a page."""
-
-        data = self.open_page(title).read()
-        text = unicode(data, self.charset, 'replace')
-        return text
-
-    def page_lines(self, page):
-        for data in page:
-            yield unicode(data, self.charset, 'replace')
+    def _commit(self, files, comment, user):
+        match = mercurial.match.exact(self.repo_path, '', list(files))
+        return self.repo.commit(match=match, text=comment, user=user, force=True)
 
 
-    @locked_repo
-    def delete_page(self, title, author = u'', comment = u''):
+    @with_working_copy_locked
+    @with_storage_locked
+    def delete_page(self, title, author=u'', comment=u''):
         user = author.encode('utf-8') or 'anon'
         text = comment.encode('utf-8') or 'deleted'
         repo_file = self._title_to_file(title)
         user = author.encode('utf-8') or 'anon'
         text = comment.encode('utf-8') or 'deleted'
         repo_file = self._title_to_file(title)
@@ -264,109 +266,130 @@ class VersionedStorage(object):
         self.repo.remove([repo_file])
         self._commit([repo_file], text, user)
 
         self.repo.remove([repo_file])
         self._commit([repo_file], text, user)
 
-    def open_page(self, title):
-        if title not in self:
-            raise DocumentNotFound()
+    def page_text(self, title, revision=None):
+        """Read unicode text of a page."""
+        ctx = self._find_filectx(title, revision)
+
+        if ctx is None:
+            raise DocumentNotFound(title)
+
+        return ctx.data().decode(self.charset, 'replace'), ctx.filerev()
+
+    def page_text_by_tag(self, title, tag):
+        """Read unicode text of a taged page."""
+        fname = self._title_to_file(title)
+        tag = u"{fname}#{tag}".format(**locals()).encode('utf-8')
 
         try:
 
         try:
-            return open(self._file_path(title), "rb")
-        except IOError:
-            import traceback
-            print traceback.print_exc()
-            raise DocumentNotFound()
+            ctx = self.repo[tag][fname]
+            return ctx.data().decode(self.charset, 'replace'), ctx.filerev()
+        except IndexError:
+            raise DocumentNotFound(fname)
 
 
+    @with_working_copy_locked
     def page_file_meta(self, title):
         """Get page's inode number, size and last modification time."""
     def page_file_meta(self, title):
         """Get page's inode number, size and last modification time."""
-
         try:
         try:
-            (st_mode, st_ino, st_dev, st_nlink, st_uid, st_gid, st_size,
-             st_atime, st_mtime, st_ctime) = os.stat(self._file_path(title))
+            (_st_mode, st_ino, _st_dev, _st_nlink, _st_uid, _st_gid, st_size,
+             _st_atime, st_mtime, _st_ctime) = os.stat(self._file_path(title))
         except OSError:
             return 0, 0, 0
         return st_ino, st_size, st_mtime
 
         except OSError:
             return 0, 0, 0
         return st_ino, st_size, st_mtime
 
-    def page_meta(self, title):
+    @with_working_copy_locked
+    def page_meta(self, title, revision=None):
         """Get page's revision, date, last editor and his edit comment."""
         """Get page's revision, date, last editor and his edit comment."""
-        if not title in self:
-            raise DocumentNotFound()
+        fctx = self._find_filectx(title, revision)
 
 
-        filectx_tip = self._find_filectx(title)
-        if filectx_tip is None:
-            raise DocumentNotFound()
-        rev = filectx_tip.filerev()
-        filectx = filectx_tip.filectx(rev)
-        date = datetime.datetime.fromtimestamp(filectx.date()[0])
-        author = unicode(filectx.user(), "utf-8",
-                         'replace').split('<')[0].strip()
-        comment = unicode(filectx.description(), "utf-8", 'replace')
-        return rev, date, author, comment
+        if fctx is None:
+            raise DocumentNotFound(title)
+
+        return {
+            "revision": fctx.filerev(),
+            "date": datetime.datetime.fromtimestamp(fctx.date()[0]),
+            "author": fctx.user().decode("utf-8", 'replace'),
+            "comment": fctx.description().decode("utf-8", 'replace'),
+        }
 
     def repo_revision(self):
 
     def repo_revision(self):
-        return self._changectx().rev()
+        return self.repo['tip'].rev()
+
+    def _changectx(self):
+        return self.repo['tip']
 
     def page_mime(self, title):
         """
         Guess page's mime type based on corresponding file name.
         Default ot text/x-wiki for files without an extension.
         """
 
     def page_mime(self, title):
         """
         Guess page's mime type based on corresponding file name.
         Default ot text/x-wiki for files without an extension.
         """
-        return guess_type(self._file_path(title))
-
-    def _changectx(self):
-        """Get the changectx of the tip."""
-        try:
-            # This is for Mercurial 1.0
-            return self.repo.changectx()
-        except TypeError:
-            # Mercurial 1.3 (and possibly earlier) needs an argument
-            return self.repo.changectx('tip')
+        return guess_mime(self._file_path(title))
 
 
-    def _find_filectx(self, title):
+    def _find_filectx(self, title, rev=None):
         """Find the last revision in which the file existed."""
         """Find the last revision in which the file existed."""
-
         repo_file = self._title_to_file(title)
         repo_file = self._title_to_file(title)
-        changectx = self._changectx()
+        changectx = self._changectx()  # start with tip
+        visited = set()
+
         stack = [changectx]
         stack = [changectx]
+        visited.add(changectx)
+
         while repo_file not in changectx:
             if not stack:
         while repo_file not in changectx:
             if not stack:
-                return None
+                raise DocumentNotFound(title)
+
             changectx = stack.pop()
             for parent in changectx.parents():
             changectx = stack.pop()
             for parent in changectx.parents():
-                if parent != changectx:
+                if parent not in visited:
                     stack.append(parent)
                     stack.append(parent)
-        return changectx[repo_file]
+                    visited.add(parent)
+
+        try:
+            fctx = changectx[repo_file]
+
+            if rev is not None:
+                fctx = fctx.filectx(rev)
+                fctx.filerev()
+
+            return fctx
+        except (IndexError, LookupError) as e:
+            raise DocumentNotFound(title)
 
     def page_history(self, title):
         """Iterate over the page's history."""
 
         filectx_tip = self._find_filectx(title)
 
     def page_history(self, title):
         """Iterate over the page's history."""
 
         filectx_tip = self._find_filectx(title)
-        if filectx_tip is None:
-            return
+
         maxrev = filectx_tip.filerev()
         minrev = 0
         for rev in range(maxrev, minrev - 1, -1):
             filectx = filectx_tip.filectx(rev)
             date = datetime.datetime.fromtimestamp(filectx.date()[0])
         maxrev = filectx_tip.filerev()
         minrev = 0
         for rev in range(maxrev, minrev - 1, -1):
             filectx = filectx_tip.filectx(rev)
             date = datetime.datetime.fromtimestamp(filectx.date()[0])
-            author = unicode(filectx.user(), "utf-8",
-                             'replace').split('<')[0].strip()
-            comment = unicode(filectx.description(), "utf-8", 'replace')
-            yield rev, date, author, comment
+            author = filectx.user().decode('utf-8', 'replace')
+            comment = filectx.description().decode("utf-8", 'replace')
+            tags = [t.rsplit('#', 1)[-1] for t in filectx.changectx().tags() if '#' in t]
 
 
-    def page_revision(self, title, rev):
-        """Get unicode contents of specified revision of the page."""
+            yield {
+                "version": rev,
+                "date": date,
+                "author": author,
+                "description": comment,
+                "tag": tags,
+            }
 
 
-        filectx_tip = self._find_filectx(title)
-        if filectx_tip is None:
-            raise DocumentNotFound()
-        try:
-            data = filectx_tip.filectx(rev).data()
-        except IndexError:
-            raise DocumentNotFound()
-        return data
+    @with_working_copy_locked
+    def add_page_tag(self, title, rev, tag, user, doctag=True):
+        ctitle = self._title_to_file(title)
+
+        if doctag:
+            tag = u"{ctitle}#{tag}".format(**locals()).encode('utf-8')
+
+        message = u"Assigned tag {tag!r} to version {rev!r} of {ctitle!r}".format(**locals()).encode('utf-8')
 
 
-    def revision_text(self, title, rev):
-        data = self.page_revision(title, rev)
-        text = unicode(data, self.charset, 'replace')
-        return text
+        fctx = self._find_filectx(title, rev)
+        self.repo.tag(
+            names=tag, node=fctx.node(), local=False,
+            user=user, message=message, date=None,
+        )
 
     def history(self):
         """Iterate over the history of entire wiki."""
 
     def history(self):
         """Iterate over the history of entire wiki."""
@@ -377,9 +400,8 @@ class VersionedStorage(object):
         for wiki_rev in range(maxrev, minrev - 1, -1):
             change = self.repo.changectx(wiki_rev)
             date = datetime.datetime.fromtimestamp(change.date()[0])
         for wiki_rev in range(maxrev, minrev - 1, -1):
             change = self.repo.changectx(wiki_rev)
             date = datetime.datetime.fromtimestamp(change.date()[0])
-            author = unicode(change.user(), "utf-8",
-                             'replace').split('<')[0].strip()
-            comment = unicode(change.description(), "utf-8", 'replace')
+            author = change.user().decode('utf-8', 'replace')
+            comment = change.description().decode("utf-8", 'replace')
             for repo_file in change.files():
                 if repo_file.startswith(self.repo_prefix):
                     title = self._file_to_title(repo_file)
             for repo_file in change.files():
                 if repo_file.startswith(self.repo_prefix):
                     title = self._file_to_title(repo_file)
@@ -390,12 +412,9 @@ class VersionedStorage(object):
                     yield title, rev, date, author, comment
 
     def all_pages(self):
                     yield title, rev, date, author, comment
 
     def all_pages(self):
+        tip = self.repo['tip']
         """Iterate over the titles of all pages in the wiki."""
         """Iterate over the titles of all pages in the wiki."""
-        return [ urlunquote(filename) for filename in self.repo.dirstate]
-        #status = self.repo.status(self.repo[None], None, None, True, True, True)
-        #clean_files = status[6]
-        #for filename in clean_files:
-        #    yield
+        return [urlunquote(filename) for filename in tip]
 
     def changed_since(self, rev):
         """Return all pages that changed since specified repository revision."""
 
     def changed_since(self, rev):
         """Return all pages that changed since specified repository revision."""