cvs2cl.pl

java写的html的解析器parser

#!/bin/sh
exec perl -w -x $0 ${1+"$@"} # -*- mode: perl; perl-indent-level: 2; -*-
#!perl -w


##############################################################
###                                                        ###
### cvs2cl.pl: produce ChangeLog(s) from `cvs log` output. ###
###                                                        ###
##############################################################

## $Revision: 1.1 $
## $Date: 2003/09/23 03:41:34 $
## $Author: derrickoswald $
##
##   (C) 2001,2002,2003 Martyn J. Pearce <[e-mail]>, under the GNU GPL.
##   (C) 1999 Karl Fogel <[e-mail]>, under the GNU GPL.
##
##   (Extensively hacked on by Melissa O'Neill <[e-mail]>.)
##   (Gecos hacking by Robin Johnson <[e-mail]>.)
##
## cvs2cl.pl is free software; you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation; either version 2, or (at your option)
## any later version.
##
## cvs2cl.pl is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
## GNU General Public License for more details.
##
## You may have received a copy of the GNU General Public License
## along with cvs2cl.pl; see the file COPYING.  If not, write to the
## Free Software Foundation, Inc., 59 Temple Place - Suite 330,
## Boston, MA 02111-1307, USA.


use strict;
use Text::Wrap;
use Time::Local;
use File::Basename qw( fileparse );
use User::pwent;


# The Plan:
#
# Read in the logs for multiple files, spit out a nice ChangeLog that
# mirrors the information entered during `cvs commit'.
#
# The problem presents some challenges. In an ideal world, we could
# detect files with the same author, log message, and checkin time --
# each <filelist, author, time, logmessage> would be a changelog entry.
# We'd sort them; and spit them out.  Unfortunately, CVS is *not atomic*
# so checkins can span a range of times.  Also, the directory structure
# could be hierarchical.
#
# Another question is whether we really want to have the ChangeLog
# exactly reflect commits. An author could issue two related commits,
# with different log entries, reflecting a single logical change to the
# source. GNU style ChangeLogs group these under a single author/date.
# We try to do the same.
#
# So, we parse the output of `cvs log', storing log messages in a
# multilevel hash that stores the mapping:
#   directory => author => time => message => filelist
# As we go, we notice "nearby" commit times and store them together
# (i.e., under the same timestamp), so they appear in the same log
# entry.
#
# When we've read all the logs, we twist this mapping into
# a time => author => message => filelist mapping for each directory.
#
# If we're not using the `--distributed' flag, the directory is always
# considered to be `./', even as descend into subdirectories.


############### Globals ################

use constant MAILNAME => "/etc/mailname";

# What we run to generate it:
my $Log_Source_Command = "cvs log";

# In case we have to print it out:
my $VERSION = '$Revision: 1.1 $';
$VERSION =~ s/\S+\s+(\S+)\s+\S+/$1/;

## Vars set by options:

# Print debugging messages?
my $Debug = 0;

# Just show version and exit?
my $Print_Version = 0;

# Just print usage message and exit?
my $Print_Usage = 0;

# Single top-level ChangeLog, or one per subdirectory?
my $Distributed = 0;

# What file should we generate (defaults to "ChangeLog")?
my $Log_File_Name = "ChangeLog";

# Grab most recent entry date from existing ChangeLog file, just add
# to that ChangeLog.
my $Cumulative = 0;

# `cvs log -d`, this will repeat the last entry in the old log.  This is OK,
# as it guarantees at least one entry in the update changelog, which means
# that there will always be a date to extract for the next update.  The repeat
# entry can be removed in postprocessing, if necessary.
my $Update = 0;

# Expand usernames to email addresses based on a map file?
my $User_Map_File = "";
my $User_Passwd_File;
my $Mail_Domain;

# Output log in chronological order? [default is reverse chronological order]
my $Chronological_Order = 0;

# Grab user details via gecos
my $Gecos = 0;

# User domain for gecos email addresses
my $Domain = "";

# Output to a file or to stdout?
my $Output_To_Stdout = 0;

# Eliminate empty log messages?
my $Prune_Empty_Msgs = 0;

# Tags of which not to output
my %ignore_tags;

# Show only revisions with Tags
my %show_tags;

# Don't call Text::Wrap on the body of the message
my $No_Wrap = 0;

# Don't do any pretty print processing
my $Summary = 0;

# Separates header from log message.  Code assumes it is either " " or
# "\n\n", so if there's ever an option to set it to something else,
# make sure to go through all conditionals that use this var.
my $After_Header = " ";

# XML Encoding
my $XML_Encoding = '';

# Format more for programs than for humans.
my $XML_Output = 0;

# Do some special tweaks for log data that was written in FSF
# ChangeLog style.
my $FSF_Style = 0;

# Show times in UTC instead of local time
my $UTC_Times = 0;

# Show times in output?
my $Show_Times = 1;

# Show day of week in output?
my $Show_Day_Of_Week = 0;

# Show revision numbers in output?
my $Show_Revisions = 0;

# Show dead files in output?
my $Show_Dead = 0;

# Show tags (symbolic names) in output?
my $Show_Tags = 0;

# Show tags separately in output?
my $Show_Tag_Dates = 0;

# Show branches by symbolic name in output?
my $Show_Branches = 0;

# Show only revisions on these branches or their ancestors.
my @Follow_Branches;

# Don't bother with files matching this regexp.
my @Ignore_Files;

# How exactly we match entries.  We definitely want "o",
# and user might add "i" by using --case-insensitive option.
my $Case_Insensitive = 0;

# Maybe only show log messages matching a certain regular expression.
my $Regexp_Gate = "";

# Pass this global option string along to cvs, to the left of `log':
my $Global_Opts = "";

# Pass this option string along to the cvs log subcommand:
my $Command_Opts = "";

# Read log output from stdin instead of invoking cvs log?
my $Input_From_Stdin = 0;

# Don't show filenames in output.
my $Hide_Filenames = 0;

# Don't shorten directory names from filenames.
my $Common_Dir = 1;

# Max checkin duration. CVS checkin is not atomic, so we may have checkin
# times that span a range of time. We assume that checkins will last no
# longer than $Max_Checkin_Duration seconds, and that similarly, no
# checkins will happen from the same users with the same message less
# than $Max_Checkin_Duration seconds apart.
my $Max_Checkin_Duration = 180;

# What to put at the front of [each] ChangeLog.
my $ChangeLog_Header = "";

# Whether to enable 'delta' mode, and for what start/end tags.
my $Delta_Mode = 0;
my $Delta_From = "";
my $Delta_To = "";

my $TestCode;

# Whether to parse filenames from the RCS filename, and if so what
# prefix to strip.
my $RCS_Mode = 0;
my $RCS_Root = "";

## end vars set by options.

# latest observed times for the start/end tags in delta mode
my $Delta_StartTime = 0;
my $Delta_EndTime = 0;

# In 'cvs log' output, one long unbroken line of equal signs separates
# files:
my $file_separator = "======================================="
                   . "======================================";

# In 'cvs log' output, a shorter line of dashes separates log messages
# within a file:
my $logmsg_separator = "----------------------------";

my $No_Ancestors = 0;

############### End globals ############



&parse_options ();
if ( defined $TestCode ) {
  eval $TestCode;
  die "Eval failed: '$@'\n"
    if $@;
} else {
  &derive_change_log ();
}


### Everything below is subroutine definitions. ###

sub run_ext {
  my ($cmd) = @_;
  $cmd = [$cmd]
    unless ref $cmd;
  local $" = ' ';
  my $out = qx"@$cmd 2>&1";
  my $rv  = $?;
  my ($sig, $core, $exit) = ($? & 127, $? & 128, $? >> 8);
  return $out, $exit, $sig, $core;
}

# If accumulating, grab the boundary date from pre-existing ChangeLog.
sub maybe_grab_accumulation_date ()
{
  if (! $Cumulative || $Update) {
    return "";
  }

  # else

  open (LOG, "$Log_File_Name")
      or die ("trouble opening $Log_File_Name for reading ($!)");

  my $boundary_date;
  while (<LOG>)
  {
    if (/^(\d\d\d\d-\d\d-\d\d\s+\d\d:\d\d)/)
    {
      $boundary_date = "$1";
      last;
    }
  }

  close (LOG);
  return $boundary_date;
}

# Fills up a ChangeLog structure in the current directory.
sub derive_change_log ()
{
  # See "The Plan" above for a full explanation.

  my %grand_poobah;

  my $file_full_path;
  my $time;
  my $revision;
  my $author;
  my $state;
  my $lines;
  my $cvsstate;
  my $msg_txt;
  my $detected_file_separator;

  my %tag_date_printed;

  # Might be adding to an existing ChangeLog
  my $accumulation_date = &maybe_grab_accumulation_date ();
  if ($accumulation_date) {
    # Insert -d immediately after 'cvs log'
    my $Log_Date_Command = "-d\'>${accumulation_date}\'";
    $Log_Source_Command =~ s/(^.*log\S*)/$1 $Log_Date_Command/;
    &debug ("(adding log msg starting from $accumulation_date)\n");
  }

  # We might be expanding usernames
  my %usermap;

  # In general, it's probably not very maintainable to use state
  # variables like this to tell the loop what it's doing at any given
  # moment, but this is only the first one, and if we never have more
  # than a few of these, it's okay.
  my $collecting_symbolic_names = 0;
  my %symbolic_names;    # Where tag names get stored.
  my %branch_names;      # We'll grab branch names while we're at it.
  my %branch_numbers;    # Save some revisions for @Follow_Branches
  my @branch_roots;      # For showing which files are branch ancestors.

  # Bleargh.  Compensate for a deficiency of custom wrapping.
  if (($After_Header ne " ") and $FSF_Style)
  {
    $After_Header .= "\t";
  }

  if (! $Input_From_Stdin) {
    &debug ("(run \"${Log_Source_Command}\")\n");
    open (LOG_SOURCE, "$Log_Source_Command |")
        or die "unable to run \"${Log_Source_Command}\"";
  }
  else {
    open (LOG_SOURCE, "-") or die "unable to open stdin for reading";
  }

  binmode LOG_SOURCE;

  %usermap = &maybe_read_user_map_file ();

  while (<LOG_SOURCE>)
  {
    # Canonicalize line endings
    s/\r$//;
    my $new_full_path;

    # If on a new file and don't see filename, skip until we find it, and
    # when we find it, grab it.
    if (! (defined $file_full_path))
    {
      if (/^Working file: (.*)/) {
        $new_full_path = $1;
      } elsif ($RCS_Mode && m|^RCS file: $RCS_Root/(.*),v$|) {
        $new_full_path = $1;
      }
    }

    if (defined $new_full_path)
    {
      $file_full_path = $new_full_path;
      if (@Ignore_Files)
      {
        my $base;
        ($base, undef, undef) = fileparse ($file_full_path);
        # Ouch, I wish trailing operators in regexps could be
        # evaluated on the fly!
        if ($Case_Insensitive) {
          if (grep ($file_full_path =~ m|$_|i, @Ignore_Files)) {
            undef $file_full_path;
          }
        }
        elsif (grep ($file_full_path =~ m|$_|, @Ignore_Files)) {
          undef $file_full_path;
        }
      }
      next;
    }

    # Just spin wheels if no file defined yet.
    next if (! $file_full_path);

    # Collect tag names in case we're asked to print them in the output.
    if (/^symbolic names:$/) {
      $collecting_symbolic_names = 1;
      next;  # There's no more info on this line, so skip to next
    }
    if ($collecting_symbolic_names)
    {
      # All tag names are listed with whitespace in front in cvs log
      # output; so if see non-whitespace, then we're done collecting.
      if (/^\S/) {
        $collecting_symbolic_names = 0;
      }
      else    # we're looking at a tag name, so parse & store it
      {
        # According to the Cederqvist manual, in node "Tags", tag
        # names must start with an uppercase or lowercase letter and
        # can contain uppercase and lowercase letters, digits, `-',
        # and `_'.  However, it's not our place to enforce that, so
        # we'll allow anything CVS hands us to be a tag:
        /^\s+([^:]+): ([\d.]+)$/;