cvs2svn_lib/git_revision_recorder.py

   1 # (Be in -*- python -*- mode.)
   2 #
   3 # ====================================================================
   4 # Copyright (c) 2007-2009 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 """Write file contents to a stream of git-fast-import blobs."""
  18
  19 import itertools
  20
  21 from cvs2svn_lib.symbol import Trunk
  22 from cvs2svn_lib.cvs_item import CVSRevisionDelete
  23 from cvs2svn_lib.cvs_item import CVSSymbol
  24 from cvs2svn_lib.fulltext_revision_recorder import FulltextRevisionRecorder
  25 from cvs2svn_lib.key_generator import KeyGenerator
  26
  27
  28 class GitRevisionRecorder(FulltextRevisionRecorder):
  29   """Output file revisions to git-fast-import."""
  30
  31   def __init__(self, blob_filename):
  32     self.blob_filename = blob_filename
  33
  34   def start(self):
  35     self.dump_file = open(self.blob_filename, 'wb')
  36     self._mark_generator = KeyGenerator()
  37
  38   def start_file(self, cvs_file_items):
  39     self._cvs_file_items = cvs_file_items
  40
  41   def _get_original_source(self, cvs_rev):
  42     """Return the original source of the contents of CVS_REV.
  43
  44     Return the first non-delete CVSRevision with the same contents as
  45     CVS_REV.  'First' here refers to deltatext order; i.e., the very
  46     first revision is HEAD on trunk, then backwards to the root of a
  47     branch, then out to the tip of a branch.
  48
  49     The candidates are all revisions along the CVS delta-dependency
  50     chain until the next one that has a deltatext (inclusive).  Of the
  51     candidates, CVSRevisionDeletes are disqualified because, even
  52     though CVS records their contents, it is impossible to extract
  53     their fulltext using commands like 'cvs checkout -p'.
  54
  55     If there is no other CVSRevision that has the same content, return
  56     CVS_REV itself."""
  57
  58     # Keep track of the "best" source CVSRevision found so far:
  59     best_source_rev = None
  60
  61     for cvs_rev in itertools.chain(
  62           [cvs_rev], self._cvs_file_items.iter_deltatext_ancestors(cvs_rev)
  63           ):
  64       if not isinstance(cvs_rev, CVSRevisionDelete):
  65         best_source_rev = cvs_rev
  66
  67       if cvs_rev.deltatext_exists:
  68         break
  69
  70     return best_source_rev
  71
  72   def record_fulltext(self, cvs_rev, log, fulltext):
  73     """Write the fulltext to a blob if it is original and not a delete.
  74
  75     The reason we go to this trouble is to avoid writing the same file
  76     contents multiple times for a string of revisions that don't have
  77     deltatexts (as, for example, happens with dead revisions and
  78     imported revisions)."""
  79
  80     if isinstance(cvs_rev, CVSRevisionDelete):
  81       # There is no need to record a delete revision, and its token
  82       # will never be needed:
  83       return None
  84
  85     source = self._get_original_source(cvs_rev)
  86
  87     if source.id == cvs_rev.id:
  88       # Revision is its own source; write it out:
  89       mark = self._mark_generator.gen_id()
  90       self.dump_file.write('blob\n')
  91       self.dump_file.write('mark :%d\n' % (mark,))
  92       self.dump_file.write('data %d\n' % (len(fulltext),))
  93       self.dump_file.write(fulltext)
  94       self.dump_file.write('\n')
  95       return mark
  96     else:
  97       # Return as revision_recorder_token the CVSRevision.id of the
  98       # original source revision:
  99       return source.revision_recorder_token
 100
 101   def finish_file(self, cvs_file_items):
 102     # Determine the original source of each CVSSymbol, and store it as
 103     # the symbol's revision_recorder_token.
 104     for cvs_item in cvs_file_items.values():
 105       if isinstance(cvs_item, CVSSymbol):
 106         cvs_source = cvs_item.get_cvs_revision_source(cvs_file_items)
 107         cvs_item.revision_recorder_token = cvs_source.revision_recorder_token
 108
 109     del self._cvs_file_items
 110
 111   def finish(self):
 112     self.dump_file.close()
 113
 114