Bug 88900 - Force cvs_ondisk_charset

git-svn-id: svn://svn.office.custis.ru/3rdparty/viewvc.org/trunk@1451 6955db30-a419-402b-8a0d-67ecbb4d7f56

bin/cvsdbadmin

@@ -47,12 +47,12 @@ import vclib.ccvs
 from stat import *
   
 
-def UpdateFile(db, repository, path, update, latest_checkin, quiet_level):
+def UpdateFile(db, repository, path, update, latest_checkin, quiet_level, encoding = None):
     try:
         if update:
             mtime = os.stat(repository.rcsfile(path, 1))[ST_MTIME]
             if mtime < latest_checkin:
-              return
+                return
             commit_list = cvsdb.GetUnrecordedCommitList(repository, path, db)
         else:
             commit_list = cvsdb.GetCommitListFromRCSFile(repository, path)
@@ -75,6 +75,8 @@ def UpdateFile(db, repository, path, update, latest_checkin, quiet_level):
 
     ## add the commits into the database
     for commit in commit_list:
+        if encoding:
+            commit.SetFile(commit.GetFile().decode(encoding).encode('utf-8'))
         db.AddCommit(commit)
         if printing:
             sys.stdout.write('.')
@@ -84,7 +86,7 @@ def UpdateFile(db, repository, path, update, latest_checkin, quiet_level):
 
 
 def RecurseUpdate(db, repository, directory, update, latest_checkin,
-                  quiet_level):
+                  quiet_level, encoding = None):
     for entry in repository.listdir(directory, None, {}):
         path = directory + [entry.name]
 
@@ -93,12 +95,13 @@ def RecurseUpdate(db, repository, directory, update, latest_checkin,
 
         if entry.kind is vclib.DIR:
             RecurseUpdate(db, repository, path, update, latest_checkin,
-                          quiet_level)
+                          quiet_level, encoding)
             continue
 
         if entry.kind is vclib.FILE:
             UpdateFile(db, repository, path, update, latest_checkin,
-                       quiet_level)
+                       quiet_level, encoding)
+
 
 def RootPath(path, quiet_level):
     """Break os path into cvs root path and other parts"""
@@ -200,9 +203,10 @@ if __name__ == '__main__':
             if latest_checkin is None:
                 command = 'rebuild'
             RecurseUpdate(db, repository, path_parts,
-                          command == 'update', latest_checkin, quiet_level)
+                          command == 'update', latest_checkin, quiet_level,
+                          cfg.options.cvs_ondisk_charset)
     except KeyboardInterrupt:
         print
         print '** break **'
-        
+
     sys.exit(0)

lib/cvsdb.py

@@ -1188,6 +1188,7 @@ def ConnectDatabase(cfg, request=None, readonly=0):
 def ConnectDatabaseReadOnly(cfg, request):
     return ConnectDatabase(cfg, request, 1)
 
+# Get all commits from rcsfile (CVS)
 def GetCommitListFromRCSFile(repository, path_parts, revision=None):
     commit_list = []
 
@@ -1227,6 +1228,7 @@ def GetCommitListFromRCSFile(repository, path_parts, revision=None):
 
     return commit_list
 
+# Get unrecorded commits from rcsfile (CVS)
 def GetUnrecordedCommitList(repository, path_parts, db):
     commit_list = GetCommitListFromRCSFile(repository, path_parts)
 

lib/viewvc.py

@@ -3811,11 +3811,11 @@ def build_commit(request, files, max_files, dir_strip, format):
       return None
     if my_repos.roottype == 'cvs':
       # we store UTF8 in the DB
-      try: where = where.decode('utf-8')
-      except: pass
       # FIXME maybe also store "real" non-UTF8 filesystem path in the DB instead of having such setting?
-      try: where = where.encode(cfg.options.cvs_ondisk_charset)
-      except: pass
+      try:
+        where = where.decode('utf-8').encode(cfg.options.cvs_ondisk_charset)
+      except:
+        raise Exception("Invalid string encoding: "+where)
     path_parts = _path_parts(where)
 
     # In CVS, we can actually look at deleted revisions; in Subversion

lib/viewvcmagic.py

@@ -35,10 +35,12 @@ class ContentMagic:
 
     # returns (utf8_content, charset)
     def guess_charset(self, content):
-        # Try to guess with chardet
-        charset = None
-        if have_chardet:
-            # Try chardet
+        # Try UTF-8
+        charset = 'utf-8'
+        try: content = content.decode('utf-8')
+        except: charset = None
+        if charset is None and have_chardet:
+            # Try to guess with chardet
             try:
                 # Only detect on first 256KB if content is longer
                 if len(content) > 256*1024:
@@ -56,11 +58,6 @@ class ContentMagic:
                 else:
                     content = content.decode(charset)
             except: charset = None
-        else:
-            # Try UTF-8
-            charset = 'utf-8'
-            try: content = content.decode('utf-8')
-            except: charset = None
         # Then try to guess primitively
         if charset is None:
             for charset in self.encodings: