tools/dev/normalize-dump.py

   1 #!/usr/bin/env python
   2
   3 import sys
   4 import re
   5 import string
   6
   7 header_re = re.compile(r'^([^:]*): ?(.*)$')
   8
   9 class NodePath:
  10     def __init__(self, path, headers):
  11         self.path = path
  12         self.headers = headers
  13
  14     def dump(self):
  15         print (' ' * 3) + self.path
  16         headers = self.headers.keys()
  17         headers.sort()
  18         for header in headers:
  19             print (' ' * 6) + header + ': ' + self.headers[header]
  20
  21
  22 def dump_revision(rev, nodepaths):
  23     sys.stderr.write('* Normalizing revision ' + rev + '...')
  24     print 'Revision ' + rev
  25     paths = nodepaths.keys()
  26     paths.sort()
  27     for path in paths:
  28         nodepath = nodepaths[path]
  29         nodepath.dump()
  30     sys.stderr.write('done\n')
  31
  32
  33
  34 def parse_header_block(fp):
  35     headers = {}
  36     while 1:
  37         line = fp.readline()
  38         if line == '':
  39             return headers, 1
  40         line = string.strip(line)
  41         if line == '':
  42             return headers, 0
  43         matches = header_re.match(line)
  44         if not matches:
  45             raise Exception('Malformed header block')
  46         headers[matches.group(1)] = matches.group(2)
  47
  48
  49 def parse_file(fp):
  50     nodepaths = {}
  51     current_rev = None
  52
  53     while 1:
  54         # Parse a block of headers
  55         headers, eof = parse_header_block(fp)
  56
  57         # This is a revision header block
  58         if headers.has_key('Revision-number'):
  59
  60             # If there was a previous revision, dump it
  61             if current_rev:
  62                 dump_revision(current_rev, nodepaths)
  63
  64             # Reset the data for this revision
  65             current_rev = headers['Revision-number']
  66             nodepaths = {}
  67
  68             # Skip the contents
  69             prop_len = headers.get('Prop-content-length', 0)
  70             fp.read(int(prop_len))
  71
  72         # This is a node header block
  73         elif headers.has_key('Node-path'):
  74
  75             # Make a new NodePath object, and add it to the
  76             # dictionary thereof
  77             path = headers['Node-path']
  78             node = NodePath(path, headers)
  79             nodepaths[path] = node
  80
  81             # Skip the content
  82             text_len = headers.get('Text-content-length', 0)
  83             prop_len = headers.get('Prop-content-length', 0)
  84             fp.read(int(text_len) + int(prop_len))
  85
  86         # Not a revision, not a node -- if we've already seen at least
  87         # one revision block, we are in an errorful state.
  88         elif current_rev and len(headers.keys()):
  89             raise Exception('Header block from outta nowhere')
  90
  91         if eof:
  92             if current_rev:
  93                 dump_revision(current_rev, nodepaths)
  94             break
  95
  96 def usage():
  97     print 'Usage: ' + sys.argv[0] + ' [DUMPFILE]'
  98     print ''
  99     print 'Reads a Subversion dumpfile from DUMPFILE (or, if not provided,'
 100     print 'from stdin) and normalizes the metadata contained therein,'
 101     print 'printing summarized and sorted information.  This is useful for'
 102     print 'generating data about dumpfiles in a diffable fashion.'
 103     sys.exit(0)
 104
 105 def main():
 106     if len(sys.argv) > 1:
 107         if sys.argv[1] == '--help':
 108             usage()
 109         fp = open(sys.argv[1], 'rb')
 110     else:
 111         fp = sys.stdin
 112     parse_file(fp)
 113
 114
 115 if __name__ == '__main__':
 116     main()
 117
 118
 119
 120