Tools/scripts/logmerge.py

   1 #! /usr/bin/env python
   2
   3 """Consolidate a bunch of CVS or RCS logs read from stdin.
   4
   5 Input should be the output of a CVS or RCS logging command, e.g.
   6
   7     cvs log -rrelease14:
   8
   9 which dumps all log messages from release1.4 upwards (assuming that
  10 release 1.4 was tagged with tag 'release14').  Note the trailing
  11 colon!
  12
  13 This collects all the revision records and outputs them sorted by date
  14 rather than by file, collapsing duplicate revision record, i.e.,
  15 records with the same message for different files.
  16
  17 The -t option causes it to truncate (discard) the last revision log
  18 entry; this is useful when using something like the above cvs log
  19 command, which shows the revisions including the given tag, while you
  20 probably want everything *since* that tag.
  21
  22 XXX This code was created by reverse engineering CVS 1.9 and RCS 5.7
  23 from their output.
  24
  25 """
  26
  27 import os, sys, getopt, string, re
  28
  29 sep1 = '='*77 + '\n'                    # file separator
  30 sep2 = '-'*28 + '\n'                    # revision separator
  31
  32 def main():
  33     """Main program"""
  34     truncate_last = 0
  35     reverse = 0
  36     opts, args = getopt.getopt(sys.argv[1:], "tr")
  37     for o, a in opts:
  38         if o == '-t':
  39             truncate_last = 1
  40         elif o == '-r':
  41             reverse = 1
  42     database = []
  43     while 1:
  44         chunk = read_chunk(sys.stdin)
  45         if not chunk:
  46             break
  47         records = digest_chunk(chunk)
  48         if truncate_last:
  49             del records[-1]
  50         database[len(database):] = records
  51     database.sort()
  52     if not reverse:
  53         database.reverse()
  54     format_output(database)
  55
  56 def read_chunk(fp):
  57     """Read a chunk -- data for one file, ending with sep1.
  58
  59     Split the chunk in parts separated by sep2.
  60
  61     """
  62     chunk = []
  63     lines = []
  64     while 1:
  65         line = fp.readline()
  66         if not line:
  67             break
  68         if line == sep1:
  69             if lines:
  70                 chunk.append(lines)
  71             break
  72         if line == sep2:
  73             if lines:
  74                 chunk.append(lines)
  75                 lines = []
  76         else:
  77             lines.append(line)
  78     return chunk
  79
  80 def digest_chunk(chunk):
  81     """Digest a chunk -- extrach working file name and revisions"""
  82     lines = chunk[0]
  83     key = 'Working file:'
  84     keylen = len(key)
  85     for line in lines:
  86         if line[:keylen] == key:
  87             working_file = string.strip(line[keylen:])
  88             break
  89     else:
  90         working_file = None
  91     records = []
  92     for lines in chunk[1:]:
  93         revline = lines[0]
  94         dateline = lines[1]
  95         text = lines[2:]
  96         words = string.split(dateline)
  97         author = None
  98         if len(words) >= 3 and words[0] == 'date:':
  99             dateword = words[1]
 100             timeword = words[2]
 101             if timeword[-1:] == ';':
 102                 timeword = timeword[:-1]
 103             date = dateword + ' ' + timeword
 104             if len(words) >= 5 and words[3] == 'author:':
 105                 author = words[4]
 106                 if author[-1:] == ';':
 107                     author = author[:-1]
 108         else:
 109             date = None
 110             text.insert(0, revline)
 111         words = string.split(revline)
 112         if len(words) >= 2 and words[0] == 'revision':
 113             rev = words[1]
 114         else:
 115             rev = None
 116             text.insert(0, revline)
 117         records.append((date, working_file, rev, author, text))
 118     return records
 119
 120 def format_output(database):
 121     prevtext = None
 122     prev = []
 123     database.append((None, None, None, None, None)) # Sentinel
 124     for (date, working_file, rev, author, text) in database:
 125         if text != prevtext:
 126             if prev:
 127                 print sep2,
 128                 for (p_date, p_working_file, p_rev, p_author) in prev:
 129                     print p_date, p_author, p_working_file, p_rev
 130                 sys.stdout.writelines(prevtext)
 131             prev = []
 132         prev.append((date, working_file, rev, author))
 133         prevtext = text
 134
 135 main()