tools/pcre/CleanTxt
author John Schoenick <john@pointysoftware.net>
Wed Jun 01 20:36:16 2011 -0400 (2011-06-01)
changeset 3292 f428ffc8a3fe
parent 2762 d2d467bf1b5a
permissions -rwxr-xr-x
Fixed regression in FindEntityByClassname from bug 4899 (bug 4941, r=predcrab).
dvander@2762
     1
#! /usr/bin/perl -w
dvander@2762
     2
dvander@2762
     3
# Script to take the output of nroff -man and remove all the backspacing and
dvander@2762
     4
# the page footers and the screen commands etc so that it is more usefully
dvander@2762
     5
# readable online. In fact, in the latest nroff, intermediate footers don't
dvander@2762
     6
# seem to be generated any more.
dvander@2762
     7
dvander@2762
     8
$blankcount = 0;
dvander@2762
     9
$lastwascut = 0;
dvander@2762
    10
$firstheader = 1;
dvander@2762
    11
dvander@2762
    12
# Input on STDIN; output to STDOUT.
dvander@2762
    13
dvander@2762
    14
while (<STDIN>)
dvander@2762
    15
  {
dvander@2762
    16
  s/\x1b\[\d+m//g;   # Remove screen controls "ESC [ number m"
dvander@2762
    17
  s/.\x8//g;         # Remove "char, backspace"
dvander@2762
    18
dvander@2762
    19
  # Handle header lines. Retain only the first one we encounter, but remove
dvander@2762
    20
  # the blank line that follows. Any others (e.g. at end of document) and the
dvander@2762
    21
  # following blank line are dropped.
dvander@2762
    22
dvander@2762
    23
  if (/^PCRE(\w*)\(([13])\)\s+PCRE\1\(\2\)$/)
dvander@2762
    24
    {
dvander@2762
    25
    if ($firstheader)
dvander@2762
    26
      {
dvander@2762
    27
      $firstheader = 0;
dvander@2762
    28
      print;
dvander@2762
    29
      $lastprinted = $_;
dvander@2762
    30
      $lastwascut = 0;
dvander@2762
    31
      }
dvander@2762
    32
    $_=<STDIN>;       # Remove a blank that follows
dvander@2762
    33
    next;
dvander@2762
    34
    }
dvander@2762
    35
dvander@2762
    36
  # Count runs of empty lines
dvander@2762
    37
dvander@2762
    38
  if (/^\s*$/)
dvander@2762
    39
    {
dvander@2762
    40
    $blankcount++;
dvander@2762
    41
    $lastwascut = 0;
dvander@2762
    42
    next;
dvander@2762
    43
    }
dvander@2762
    44
dvander@2762
    45
  # If a chunk of lines has been cut out (page footer) and the next line
dvander@2762
    46
  # has a different indentation, put back one blank line.
dvander@2762
    47
dvander@2762
    48
  if ($lastwascut && $blankcount < 1 && defined($lastprinted))
dvander@2762
    49
    {
dvander@2762
    50
    ($a) = $lastprinted =~ /^(\s*)/;
dvander@2762
    51
    ($b) = $_ =~ /^(\s*)/;
dvander@2762
    52
    $blankcount++ if ($a ne $b);
dvander@2762
    53
    }
dvander@2762
    54
dvander@2762
    55
  # We get here only when we have a non-blank line in hand. If it was preceded
dvander@2762
    56
  # by 3 or more blank lines, read the next 3 lines and see if they are blank.
dvander@2762
    57
  # If so, remove all 7 lines, and remember that we have just done a cut.
dvander@2762
    58
dvander@2762
    59
  if ($blankcount >= 3)
dvander@2762
    60
    {
dvander@2762
    61
    for ($i = 0; $i < 3; $i++)
dvander@2762
    62
      {
dvander@2762
    63
      $next[$i] = <STDIN>;
dvander@2762
    64
      $next[$i] = "" if !defined $next[$i];
dvander@2762
    65
      $next[$i] =~ s/\x1b\[\d+m//g;   # Remove screen controls "ESC [ number m"
dvander@2762
    66
      $next[$i] =~ s/.\x8//g;         # Remove "char, backspace"
dvander@2762
    67
      }
dvander@2762
    68
dvander@2762
    69
    # Cut out chunks of the form <3 blanks><non-blank><3 blanks>
dvander@2762
    70
dvander@2762
    71
    if ($next[0] =~ /^\s*$/ &&
dvander@2762
    72
        $next[1] =~ /^\s*$/ &&
dvander@2762
    73
        $next[2] =~ /^\s*$/)
dvander@2762
    74
      {
dvander@2762
    75
      $blankcount -= 3;
dvander@2762
    76
      $lastwascut = 1;
dvander@2762
    77
      }
dvander@2762
    78
dvander@2762
    79
    # Otherwise output the saved blanks, the current, and the next three
dvander@2762
    80
    # lines. Remember the last printed line.
dvander@2762
    81
dvander@2762
    82
    else
dvander@2762
    83
      {
dvander@2762
    84
      for ($i = 0; $i < $blankcount; $i++) { print "\n"; }
dvander@2762
    85
      print;
dvander@2762
    86
      for ($i = 0; $i < 3; $i++)
dvander@2762
    87
        {
dvander@2762
    88
        $next[$i] =~ s/.\x8//g;
dvander@2762
    89
        print $next[$i];
dvander@2762
    90
        $lastprinted = $_;
dvander@2762
    91
        }
dvander@2762
    92
      $lastwascut = 0;
dvander@2762
    93
      $blankcount = 0;
dvander@2762
    94
      }
dvander@2762
    95
    }
dvander@2762
    96
dvander@2762
    97
  # This non-blank line is not preceded by 3 or more blank lines. Output
dvander@2762
    98
  # any blanks there are, and the line. Remember it. Force two blank lines
dvander@2762
    99
  # before headings.
dvander@2762
   100
dvander@2762
   101
  else
dvander@2762
   102
    {
dvander@2762
   103
    $blankcount = 2 if /^\S/ && !/^Last updated/ && !/^Copyright/ &&
dvander@2762
   104
      defined($lastprinted);
dvander@2762
   105
    for ($i = 0; $i < $blankcount; $i++) { print "\n"; }
dvander@2762
   106
    print;
dvander@2762
   107
    $lastprinted = $_;
dvander@2762
   108
    $lastwascut = 0;
dvander@2762
   109
    $blankcount = 0;
dvander@2762
   110
    }
dvander@2762
   111
  }
dvander@2762
   112
dvander@2762
   113
# End