cvs2svn_lib/symbol_database.py

   1 # (Be in -*- python -*- mode.)
   2 #
   3 # ====================================================================
   4 # Copyright (c) 2000-2006 CollabNet.  All rights reserved.
   5 #
   6 # This software is licensed as described in the file COPYING, which
   7 # you should have received as part of this distribution.  The terms
   8 # are also available at http://subversion.tigris.org/license-1.html.
   9 # If newer versions of this license are posted there, you may use a
  10 # newer version instead, at your option.
  11 #
  12 # This software consists of voluntary contributions made by many
  13 # individuals.  For exact contribution history, see the revision
  14 # history and logs, available at http://cvs2svn.tigris.org/.
  15 # ====================================================================
  16
  17 """This module contains database facilities used by cvs2svn."""
  18
  19
  20 from __future__ import generators
  21
  22 from boolean import *
  23 import config
  24 from artifact_manager import artifact_manager
  25
  26
  27 def match_regexp_list(regexp_list, s):
  28   """Test whether string S matches any of the compiled regexps in
  29   REGEXP_LIST."""
  30
  31   for regexp in regexp_list:
  32     if regexp.match(s):
  33       return True
  34   return False
  35
  36
  37 class SymbolDatabase:
  38   """This database records information on all symbols in the RCS
  39   files.  It is created in pass 1 and it is used in pass 2."""
  40
  41   def __init__(self):
  42     # A hash that maps tag names to commit counts
  43     self.tags = { }
  44     # A hash that maps branch names to lists of the format
  45     # [ create_count, commit_count, blockers ], where blockers
  46     # is a hash that lists the symbols that depend on the
  47     # the branch.  The blockers hash is used as a set, so the
  48     # values are not used.
  49     self.branches = { }
  50
  51   def register_tag_creation(self, name):
  52     """Register the creation of the tag NAME."""
  53
  54     self.tags[name] = self.tags.get(name, 0) + 1
  55
  56   def _branch(self, name):
  57     """Helper function to get a branch node that will create and
  58     initialize the node if it does not exist."""
  59
  60     if not self.branches.has_key(name):
  61       self.branches[name] = [ 0, 0, { } ]
  62     return self.branches[name]
  63
  64   def register_branch_creation(self, name):
  65     """Register the creation of the branch NAME."""
  66
  67     self._branch(name)[0] += 1
  68
  69   def register_branch_commit(self, name):
  70     """Register a commit on the branch NAME."""
  71
  72     self._branch(name)[1] += 1
  73
  74   def register_branch_blocker(self, name, blocker):
  75     """Register BLOCKER as a blocker on the branch NAME."""
  76
  77     self._branch(name)[2][blocker] = None
  78
  79   def branch_has_commit(self, name):
  80     """Return non-zero if NAME has commits.  Returns 0 if name
  81     is not a branch or if it has no commits."""
  82
  83     return self.branches.has_key(name) and self.branches[name][1]
  84
  85   def find_excluded_symbols(self, regexp_list):
  86     """Returns a hash of all symbols that match the regexps in
  87     REGEXP_LIST.  The hash is used as a set so the values are
  88     not used."""
  89
  90     excludes = { }
  91     for tag in self.tags:
  92       if match_regexp_list(regexp_list, tag):
  93         excludes[tag] = None
  94     for branch in self.branches:
  95       if match_regexp_list(regexp_list, branch):
  96         excludes[branch] = None
  97     return excludes
  98
  99   def find_branch_exclude_blockers(self, branch, excludes):
 100     """Find all blockers of BRANCH, excluding the ones in the hash
 101     EXCLUDES."""
 102
 103     blockers = { }
 104     if excludes.has_key(branch):
 105       for blocker in self.branches[branch][2]:
 106         if not excludes.has_key(blocker):
 107           blockers[blocker] = None
 108     return blockers
 109
 110   def find_blocked_excludes(self, excludes):
 111     """Find all branches not in EXCLUDES that have blocking symbols that
 112     are not themselves excluded.  Return a hash that maps branch names
 113     to a hash of blockers.  The hash of blockes is used as a set so the
 114     values are not used."""
 115
 116     blocked_branches = { }
 117     for branch in self.branches:
 118       blockers = self.find_branch_exclude_blockers(branch, excludes)
 119       if blockers:
 120         blocked_branches[branch] = blockers
 121     return blocked_branches
 122
 123   def find_mismatches(self, excludes=None):
 124     """Find all symbols that are defined as both tags and branches,
 125     excluding the ones in EXCLUDES.  Returns a list of 4-tuples with
 126     the symbol name, tag count, branch count and commit count."""
 127
 128     if excludes is None:
 129       excludes = { }
 130     mismatches = [ ]
 131     for branch in self.branches:
 132       if not excludes.has_key(branch) and self.tags.has_key(branch):
 133         mismatches.append((branch,                    # name
 134                            self.tags[branch],         # tag count
 135                            self.branches[branch][0],  # branch count
 136                            self.branches[branch][1])) # commit count
 137     return mismatches
 138
 139   def read(self):
 140     """Read the symbol database from files."""
 141
 142     f = open(artifact_manager.get_temp_file(config.TAGS_LIST))
 143     while 1:
 144       line = f.readline()
 145       if not line:
 146         break
 147       tag, count = line.split()
 148       self.tags[tag] = int(count)
 149
 150     f = open(artifact_manager.get_temp_file(config.BRANCHES_LIST))
 151     while 1:
 152       line = f.readline()
 153       if not line:
 154         break
 155       words = line.split()
 156       self.branches[words[0]] = [ int(words[1]), int(words[2]), { } ]
 157       for blocker in words[3:]:
 158         self.branches[words[0]][2][blocker] = None
 159
 160   def write(self):
 161     """Store the symbol database to files."""
 162
 163     f = open(artifact_manager.get_temp_file(config.TAGS_LIST), "w")
 164     for tag, count in self.tags.items():
 165       f.write("%s %d\n" % (tag, count))
 166
 167     f = open(artifact_manager.get_temp_file(config.BRANCHES_LIST), "w")
 168     for branch, info in self.branches.items():
 169       f.write("%s %d %d" % (branch, info[0], info[1]))
 170       if info[2]:
 171         f.write(" ")
 172         f.write(" ".join(info[2].keys()))
 173       f.write("\n")
 174
 175