Changed default logging level to INFO.

[redakcja.git] / lib / vstorage / __init__.py

diff --git a/lib/vstorage/__init__.py b/lib/vstorage/__init__.py
index 1060bc2..9f7084b 100644 (file)
--- a/lib/vstorage/__init__.py
+++ b/lib/vstorage/__init__.py
@@ -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'


@@ -16,21 +24,21 @@ import mercurial.util
 
 
 def urlquote(url, safe='/'):
 
 
 def urlquote(url, safe='/'):

-    """Quotes URL 
-    
+    """Quotes URL
+

     >>> urlquote(u'Za\u017c\xf3\u0142\u0107 g\u0119\u015bl\u0105 ja\u017a\u0144')
     >>> 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'
+    'Za%C5%BC%C3%B3%C5%82%C4%87%20g%C4%99%C5%9Bl%C4%85%20ja%C5%BA%C5%84'

     """
     """

-    return urllib.quote(url.replace(' ', '_').encode('utf-8', 'ignore'), safe)
+    return urllib.quote(url.encode('utf-8', 'ignore'), 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')
     # >>> 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'
+    # u'Za\u017c\xf3\u0142\u0107_g\u0119\u015bl\u0105 ja\u017a\u0144'

     """
     """

-    return unicode(urllib.unquote(url), 'utf-8', 'ignore').replace('_', ' ')
+    return unicode(urllib.unquote(url), 'utf-8', 'ignore')

 
 
 def find_repo_path(path):
 
 
 def find_repo_path(path):


@@ -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'


@@ -109,42 +128,39 @@ 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,
                                             create=create)
 
     def reopen(self):
         """Close and reopen the repo, to make sure we are up to date."""
         self.repo_prefix = self.path[len(self.repo_path):].strip('/')
         self.repo = mercurial.hg.repository(self.ui, self.repo_path,
                                             create=create)
 
     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):
         return os.path.join(self.path, urlquote(title, safe=''))
 
         self.repo = mercurial.hg.repository(self.ui, self.repo_path)
 
     def _file_path(self, title):
         return os.path.join(self.path, urlquote(title, safe=''))
 

-    def _title_to_file(self, title):
-        return os.path.join(self.repo_prefix, urlquote(title, safe=''))
+    def _title_to_file(self, title, type=".xml"):
+        return os.path.join(self.repo_prefix, urlquote(title, safe='')) + type

 
     def _file_to_title(self, filename):
         assert filename.startswith(self.repo_prefix)
 
     def _file_to_title(self, filename):
         assert filename.startswith(self.repo_prefix)

-        name = filename[len(self.repo_prefix):].strip('/')
+        name = filename[len(self.repo_prefix):].strip('/').split('.', 1)[0]

         return urlunquote(name)
 
     def __contains__(self, title):
         return urlunquote(name)
 
     def __contains__(self, title):

-        return os.path.exists(self._file_path(title))
+        return urlquote(title) in self.repo['tip']

 
     def __iter__(self):
         return self.all_pages()
 
     def __iter__(self):
         return self.all_pages()


@@ -157,16 +173,19 @@ class VersionedStorage(object):
 
         self.repo.dirstate.setparents(parent_node)
         node = self._commit([repo_file], text, user)
 
         self.repo.dirstate.setparents(parent_node)
         node = self._commit([repo_file], text, user)

-        
+

         partial = lambda filename: repo_file == filename
         partial = lambda filename: repo_file == filename

-        
+

         # If p1 is equal to p2, there is no work to do. Even the dirstate is correct.
         p1, p2 = self.repo[None].parents()[0], self.repo[tip_node]
         if p1 == p2:
             return text
         # If p1 is equal to p2, there is no work to do. Even the dirstate is correct.
         p1, p2 = self.repo[None].parents()[0], self.repo[tip_node]
         if p1 == p2:
             return text

-        
-        # TODO: Check if merge was successful
-        mercurial.merge.update(self.repo, tip_node, True, False, partial)
+
+        try:
+            mercurial.merge.update(self.repo, tip_node, True, False, partial)
+            msg = 'merge of edit conflict'
+        except mercurial.util.Abort:
+            msg = 'failed merge of edit conflict'

 
         self.repo.dirstate.setparents(tip_node, node)
         # Mercurial 1.1 and later need updating the merge state
 
         self.repo.dirstate.setparents(tip_node, node)
         # Mercurial 1.1 and later need updating the merge state


@@ -174,52 +193,46 @@ class VersionedStorage(object):
             mercurial.merge.mergestate(self.repo).mark(repo_file, "r")
         except (AttributeError, KeyError):
             pass
             mercurial.merge.mergestate(self.repo).mark(repo_file, "r")
         except (AttributeError, KeyError):
             pass

-        return u'merge of edit conflict'
+        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:
             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()
         try:
             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()

-            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)


@@ -230,24 +243,17 @@ 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
+    @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'
     def delete_page(self, title, author=u'', comment=u''):
         user = author.encode('utf-8') or 'anon'
         text = comment.encode('utf-8') or 'deleted'


@@ -260,103 +266,137 @@ 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):
+    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:
-            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."""

+        fctx = self._find_filectx(title, revision)

 
 

-        filectx_tip = self._find_filectx(title)
-        if filectx_tip is None:
-            raise DocumentNotFound()
-            #return -1, None, u'', u''
-        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."""

+        tip = self._changectx()  # start with tip

 
 

-        repo_file = self._title_to_file(title)
-        changectx = self._changectx()
-        stack = [changectx]
-        while repo_file not in changectx:
-            if not stack:
-                return None
-            changectx = stack.pop()
-            for parent in changectx.parents():
-                if parent != changectx:
-                    stack.append(parent)
-        return changectx[repo_file]
+        def tree_search(tip, repo_file):
+            logging.info("Searching for %r", repo_file)
+            current = tip
+            visited = set()
+
+            stack = [current]
+            visited.add(current)
+
+            while repo_file not in current:
+                if not stack:
+                    raise LookupError
+
+                current = stack.pop()
+                for parent in current.parents():
+                    if parent not in visited:
+                        stack.append(parent)
+                        visited.add(parent)
+
+            fctx = current[repo_file]
+            if rev is not None:
+                fctx = fctx.filectx(rev)
+                fctx.filerev()
+            return fctx
+
+        try:
+            return tree_search(tip, self._title_to_file(title))
+        except (IndexError, LookupError) as e:
+            logging.info("XML file not found, trying plain")
+            try:
+                return tree_search(tip, self._title_to_file(title, type=''))
+            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
         maxrev = filectx_tip.filerev()
         minrev = 0

-        for rev in range(maxrev, minrev-1, -1):
+        for rev in range(maxrev, minrev - 1, -1):

             filectx = filectx_tip.filectx(rev)
             date = datetime.datetime.fromtimestamp(filectx.date()[0])
             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')

 
 

-    def revision_text(self, title, rev):
-        data = self.page_revision(title, rev)
-        text = unicode(data, self.charset, 'replace')
-        return text
+        message = u"Assigned tag {tag!r} to version {rev!r} of {ctitle!r}".format(**locals()).encode('utf-8')
+
+        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."""


@@ -364,12 +404,11 @@ class VersionedStorage(object):
         changectx = self._changectx()
         maxrev = changectx.rev()
         minrev = 0
         changectx = self._changectx()
         maxrev = changectx.rev()
         minrev = 0

-        for wiki_rev in range(maxrev, minrev-1, -1):
+        for wiki_rev in range(maxrev, minrev - 1, -1):

             change = self.repo.changectx(wiki_rev)
             date = datetime.datetime.fromtimestamp(change.date()[0])
             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)


@@ -379,13 +418,12 @@ class VersionedStorage(object):
                         rev = -1
                     yield title, rev, date, author, comment
 
                         rev = -1
                     yield title, rev, date, author, comment
 

-    def all_pages(self):
+    def all_pages(self, type=''):
+        tip = self.repo['tip']

         """Iterate over the titles of all pages in the wiki."""
         """Iterate over the titles of all pages in the wiki."""

-
-        for filename in os.listdir(self.path):
-            if (os.path.isfile(os.path.join(self.path, filename))
-                and not filename.startswith('.')):
-                yield urlunquote(filename)
+        return [self._file_to_title(filename) for filename in tip
+                  if not filename.startswith('.')
+                    and filename.endswith(type) ]

 
     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."""


@@ -399,6 +437,6 @@ class VersionedStorage(object):
         current = self.repo.lookup('tip')
         status = self.repo.status(current, last)
         modified, added, removed, deleted, unknown, ignored, clean = status
         current = self.repo.lookup('tip')
         status = self.repo.status(current, last)
         modified, added, removed, deleted, unknown, ignored, clean = status

-        for filename in modified+added+removed+deleted:
+        for filename in modified + added + removed + deleted:

             if filename.startswith(self.repo_prefix):
                 yield self._file_to_title(filename)
             if filename.startswith(self.repo_prefix):
                 yield self._file_to_title(filename)