porting vclib/ccvs into Python 3

  * now ViewVC works for CVS on Python 3, if use_rcsparse is False
  * if use_rcsparse, it cannot work correctly for checkout non text file.
   (to fix it, make rcsparse to parse bytes stream. Currently it works
    for text stream only, for corner-cutting porting for Python 3)

lib/vclib/ccvs/__init__.py

@@ -60,8 +60,8 @@ def find_root_in_parent(parent_path, rootname):
 def CVSRepository(name, rootpath, authorizer, utilities, use_rcsparse):
   rootpath = canonicalize_rootpath(rootpath)
   if use_rcsparse:
-    import ccvs
+    from . import ccvs
     return ccvs.CCVSRepository(name, rootpath, authorizer, utilities)
   else:
-    import bincvs
+    from . import bincvs
     return bincvs.BinCVSRepository(name, rootpath, authorizer, utilities)

lib/vclib/ccvs/bincvs.py

@@ -21,11 +21,20 @@
 import re
 import time
 import calendar
+import subprocess
 
 # ViewVC libs
-import popen
 import vclib.ccvs
 
+if sys.version_info[0] >= 3:
+  PY3 = True
+  import functools
+  # Python 3: workaround for cmp()
+  def cmp(a, b):
+    return (a > b) - (a < b)
+else:
+  PY3 = False
+
 def _path_join(path_parts):
   return '/'.join(path_parts)
 
@@ -123,7 +132,7 @@ def listdir(self, path_parts, rev, options):
     return data
 
   def _getpath(self, path_parts):
-    return apply(os.path.join, (self.rootpath,) + tuple(path_parts))
+    return os.path.join(*((self.rootpath,) + tuple(path_parts)))
 
   def _atticpath(self, path_parts):
     return path_parts[:-1] + ['Attic'] + path_parts[-1:]
@@ -317,9 +326,15 @@ def itemlog(self, path_parts, rev, sortby, first, limit, options):
 
     options['cvs_tags'] = tags
     if sortby == vclib.SORTBY_DATE:
-      filtered_revs.sort(_logsort_date_cmp)
+      if PY3:
+        filtered_revs.sort(key=functools.cmp_to_key(_logsort_date_cmp))
+      else:
+        filtered_revs.sort(_logsort_date_cmp)
     elif sortby == vclib.SORTBY_REV:
-      filtered_revs.sort(_logsort_rev_cmp)
+      if PY3:
+        filtered_revs.sort(key=functools.cmp_to_key(_logsort_rev_cmp))
+      else:
+        filtered_revs.sort(_logsort_rev_cmp)
 
     if len(filtered_revs) < first:
       return []
@@ -328,14 +343,21 @@ def itemlog(self, path_parts, rev, sortby, first, limit, options):
     return filtered_revs
 
   def rcs_popen(self, rcs_cmd, rcs_args, mode, capture_err=1):
+    # as we use this function as "r" mode only, we don't care stdin
+    # to communicate child process.
     if self.utilities.cvsnt:
       cmd = self.utilities.cvsnt
       args = ['rcsfile', rcs_cmd]
       args.extend(list(rcs_args))
     else:
       cmd = os.path.join(self.utilities.rcs_dir, rcs_cmd)
       args = rcs_args
-    return popen.popen(cmd, args, mode, capture_err)
+    prc = subprocess.Popen([cmd] + list(args), bufsize = -1,
+                           stdout=subprocess.PIPE,
+                           stderr=(subprocess.STDOUT if capture_err else None),
+                           universal_newlines=('t' in mode),
+                           close_fds=(sys.platform != "win32"))
+    return prc.stdout
 
   def annotate(self, path_parts, rev=None, include_text=False):
     if self.itemtype(path_parts, rev) != vclib.FILE:  # does auth-check
@@ -366,7 +388,7 @@ def rawdiff(self, path_parts1, rev1, path_parts2, rev2, type, options={}):
 
     rcsfile = self.rcsfile(path_parts1, 1)
     if path_parts1 != path_parts2:
-      raise NotImplementedError, "cannot diff across paths in cvs"
+      raise NotImplementedError("cannot diff across paths in cvs")
     args.extend(['-r' + rev1, '-r' + rev2, rcsfile])
 
     fp = self.rcs_popen('rcsdiff', args, 'rt')
@@ -487,7 +509,7 @@ def _match_revs_tags(revlist, taglist):
 
   # loop through revisions, setting properties and storing state in "history"
   for rev in revlist:
-    depth = len(rev.number) / 2 - 1
+    depth = len(rev.number) // 2 - 1
 
     # set "prev" and "next" properties
     rev.prev = rev.next = None
@@ -574,7 +596,7 @@ def _revision_tuple(revision_string):
 def _tag_tuple(revision_string):
   """convert a revision number or branch number into a tuple of integers"""
   if revision_string:
-    t = map(int, revision_string.split('.'))
+    t = [int(x) for x in revision_string.split('.')]
     l = len(t)
     if l == 1:
       return ()
@@ -603,20 +625,29 @@ class COMissingRevision(vclib.Error):
   pass
 
 ### suck up other warnings in _re_co_warning?
-_re_co_filename = re.compile(r'^(.*),v\s+-->\s+(?:(?:standard output)|(?:stdout))\s*\n?$')
-_re_co_warning = re.compile(r'^.*co: .*,v: warning: Unknown phrases like .*\n$')
-_re_co_missing_rev = re.compile(r'^.*co: .*,v: revision.*absent\n$')
-_re_co_side_branches = re.compile(r'^.*co: .*,v: no side branches present for [\d\.]+\n$')
-_re_co_revision = re.compile(r'^revision\s+([\d\.]+)\s*\n$')
+if PY3:
+  _re_co_filename = re.compile(br'^(.*),v\s+-->\s+(?:(?:standard output)|(?:stdout))\s*\n?$')
+  _re_co_warning = re.compile(br'^.*co: .*,v: warning: Unknown phrases like .*\n$')
+  _re_co_missing_rev = re.compile(br'^.*co: .*,v: revision.*absent\n$')
+  _re_co_side_branches = re.compile(br'^.*co: .*,v: no side branches present for [\d\.]+\n$')
+  _re_co_revision = re.compile(br'^revision\s+([\d\.]+)\s*\n$')
+else:
+  _re_co_filename = re.compile(r'^(.*),v\s+-->\s+(?:(?:standard output)|(?:stdout))\s*\n?$')
+  _re_co_warning = re.compile(r'^.*co: .*,v: warning: Unknown phrases like .*\n$')
+  _re_co_missing_rev = re.compile(r'^.*co: .*,v: revision.*absent\n$')
+  _re_co_side_branches = re.compile(r'^.*co: .*,v: no side branches present for [\d\.]+\n$')
+  _re_co_revision = re.compile(r'^revision\s+([\d\.]+)\s*\n$')
 
-def _parse_co_header(fp):
+def _parse_co_header(fp, encoding='utf-8'):
   """Parse RCS co header.
 
   fp is a file (pipe) opened for reading the co standard error stream.
 
   Returns: (filename, revision) or (None, None) if output is empty
   """
 
+  # Python 3: in this context, fp is raw mode.
+
   # header from co:
   #
   #/home/cvsroot/mod_dav/dav_shared_stub.c,v  -->  standard output
@@ -634,7 +665,7 @@ def _parse_co_header(fp):
     return None, None
   match = _re_co_filename.match(line)
   if not match:
-    raise COMalformedOutput, "Unable to find filename in co output stream"
+    raise COMalformedOutput("Unable to find filename in co output stream")
   filename = match.group(1)
 
   # look through subsequent lines for a revision.  we might encounter
@@ -646,15 +677,18 @@ def _parse_co_header(fp):
     # look for a revision.
     match = _re_co_revision.match(line)
     if match:
-      return filename, match.group(1)
+      if PY3:
+        return filename.decode(encoding, 'surrogateescape'), match.group(1).decode(encoding, 'surrogateescape')
+      else:
+        return filename, match.group(1)
     elif _re_co_missing_rev.match(line) or _re_co_side_branches.match(line):
-      raise COMissingRevision, "Got missing revision error from co output stream"
+      raise COMissingRevision("Got missing revision error from co output stream")
     elif _re_co_warning.match(line):
       pass
     else:
       break
 
-  raise COMalformedOutput, "Unable to find revision in co output stream"
+  raise COMalformedOutput("Unable to find revision in co output stream")
 
 # if your rlog doesn't use 77 '=' characters, then this must change
 LOG_END_MARKER = '=' * 77 + '\n'
@@ -719,15 +753,15 @@ def _parse_log_header(fp):
 
     if state == 1:
       if line[0] == '\t':
-        [ tag, rev ] = map(lambda x: x.strip(), line.split(':'))
+        [ tag, rev ] = [x.strip() for x in line.split(':')]
         taginfo[tag] = rev
       else:
         # oops. this line isn't tag info. stop parsing tags.
         state = 0
 
     if state == 2:
       if line[0] == '\t':
-        [ locker, rev ] = map(lambda x: x.strip(), line.split(':'))
+        [ locker, rev ] = [x.strip() for x in line.split(':')]
         lockinfo[rev] = locker
       else:
         # oops. this line isn't lock info. stop parsing tags.
@@ -846,7 +880,7 @@ def _parse_log_entry(fp):
     if (tm[0] - 1900) < 70:
       tm[0] = tm[0] + 100
     if tm[0] < EPOCH:
-      raise ValueError, 'invalid year'
+      raise ValueError('invalid year')
   date = calendar.timegm(tm)
 
   return Revision(rev, date,
@@ -890,7 +924,7 @@ def _file_log(revs, taginfo, lockinfo, cur_branch, filter):
   # Create tag objects
   for name, num in taginfo.items():
     taginfo[name] = Tag(name, num)
-  tags = taginfo.values()
+  tags = list(taginfo.values())
 
   # Set view_tag to a Tag object in order to filter results. We can filter by
   # revision number or branch number
@@ -991,7 +1025,7 @@ def _get_logs(repos, dir_path_parts, entries, view_tag, get_dirs):
       #       we'll search the output for the appropriate revision
       # fetch the latest revision on the default branch
       args.append('-r')
-    args.extend(map(lambda x: x.path, chunk))
+    args.extend([x.path for x in chunk])
     rlog = repos.rcs_popen('rlog', args, 'rt')
 
     # consume each file found in the resulting log
@@ -1041,7 +1075,7 @@ def _get_logs(repos, dir_path_parts, entries, view_tag, get_dirs):
       tag = None
       if view_tag == 'MAIN' or view_tag == 'HEAD':
         tag = Tag(None, default_branch)
-      elif taginfo.has_key(view_tag):
+      elif view_tag in taginfo:
         tag = Tag(None, taginfo[view_tag])
       elif view_tag and (eof != _EOF_FILE):
         # the tag wasn't found, so skip this file (unless we already
@@ -1150,7 +1184,7 @@ def _check_path(path):
   def _check_path(pathname):
     try:
       info = os.stat(pathname)
-    except os.error, e:
+    except os.error as e:
       return None, ["stat error: %s" % e]
 
     kind = None

lib/vclib/ccvs/blame.py

@@ -29,7 +29,7 @@
 import re
 import time
 import math
-import rcsparse
+from . import rcsparse
 import vclib
 
 class CVSParser(rcsparse.Sink):
@@ -151,7 +151,7 @@ def extract_revision(self, revision):
           add_lines_remaining = count
           lines_added_now = lines_added_now + count
         else:
-          raise RuntimeError, 'Error parsing diff commands'
+          raise RuntimeError('Error parsing diff commands')
 
       self.lines_added[revision]   = self.lines_added[revision]   + lines_added_now
       self.lines_removed[revision] = self.lines_removed[revision] + lines_removed_now
@@ -262,10 +262,10 @@ def parse_cvs_file(self, rcs_pathname, opt_rev = None, opt_m_timestamp = None):
 
     # CheckHidden(rcs_pathname)
     try:
-      rcsfile = open(rcs_pathname, 'rb')
+      rcsfile = open(rcs_pathname, 'r')
     except:
-      raise RuntimeError, ('error: %s appeared to be under CVS control, ' +
-              'but the RCS file is inaccessible.') % rcs_pathname
+      raise RuntimeError(('error: %s appeared to be under CVS control, ' +
+              'but the RCS file is inaccessible.') % rcs_pathname)
 
     rcsparse.parse(rcsfile, self)
     rcsfile.close()
@@ -277,7 +277,7 @@ def parse_cvs_file(self, rcs_pathname, opt_rev = None, opt_m_timestamp = None):
       # Symbolic tag or specific revision number specified.
       revision = self.map_tag_to_revision(opt_rev)
       if revision == '':
-        raise RuntimeError, 'error: -r: No such revision: ' + opt_rev
+        raise RuntimeError('error: -r: No such revision: ' + opt_rev)
 
     # The primordial revision is not always 1.1!  Go find it.
     primordial = revision
@@ -320,7 +320,7 @@ def parse_cvs_file(self, rcs_pathname, opt_rev = None, opt_m_timestamp = None):
           skip       = count
           line_count = line_count + count
         else:
-          raise RuntimeError, 'error: illegal RCS file'
+          raise RuntimeError('error: illegal RCS file')
 
       rev = self.prev_revision.get(rev)
 
@@ -372,7 +372,7 @@ def parse_cvs_file(self, rcs_pathname, opt_rev = None, opt_m_timestamp = None):
               del self.revision_map[start_line:start_line + count]
               skip = count
             else:
-              raise RuntimeError, 'Error parsing diff commands'
+              raise RuntimeError('Error parsing diff commands')
 
       else:
         # Revisions on a branch are arranged backwards from those on
@@ -405,7 +405,7 @@ def parse_cvs_file(self, rcs_pathname, opt_rev = None, opt_m_timestamp = None):
                       temp + self.revision_map[start_line + adjust:])
               adjust = adjust + skip
             else:
-              raise RuntimeError, 'Error parsing diff commands'
+              raise RuntimeError('Error parsing diff commands')
 
       last_revision = revision
 
@@ -420,7 +420,7 @@ def __init__(self, rcs_file, opt_rev=None, include_text=False):
     count = len(parser.revision_map)
     lines = parser.extract_revision(revision)
     if len(lines) != count:
-      raise RuntimeError, 'Internal consistency error'
+      raise RuntimeError('Internal consistency error')
 
     # set up some state variables
     self.revision = revision