git2cl

   1 #!/usr/bin/perl
   2
   3 # Copyright (C) 2007-2024 Simon Josefsson <simon@josefsson.org>
   4 # Copyright (C) 2007 Luis Mondesi <lemsx1@gmail.com>
   5 #
   6 # The functions mywrap, last_line_len, wrap_log_entry are derived from
   7 # the GPLv2+ cvs2cl tool, see <http://www.red-bean.com/cvs2cl/>:
   8 # Copyright (C) 2001,2002,2003,2004 Martyn J. Pearce <fluffy@cpan.org>
   9 # Copyright (C) 1999 Karl Fogel <kfogel@red-bean.com>
  10 #
  11 # This program is free software: you can redistribute it and/or modify
  12 # it under the terms of the GNU General Public License as published by
  13 # the Free Software Foundation, either version 3 of the License, or
  14 # (at your option) any later version.
  15 #
  16 # This program is distributed in the hope that it will be useful,
  17 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  18 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 # GNU General Public License for more details.
  20 #
  21 # You should have received a copy of the GNU General Public License
  22 # along with this program.  If not, see <https://www.gnu.org/licenses/>.
  23
  24 =head1 NAME
  25
  26 git2cl        -  tool to convert git logs to GNU ChangeLog
  27
  28 =head1 SYNOPSIS
  29
  30 git2cl > ChangeLog
  31
  32 If you don't want git2cl to invoke git log internally, you can use it
  33 as a pipe.
  34 It needs a git log generated with --pretty --numstat and --summary.
  35 You can use it as follows:
  36
  37     git log --pretty --numstat --summary | git2cl > ChangeLog
  38
  39 =head1 DESCRIPTION
  40
  41 Convert git logs to GNU ChangeLog format.
  42
  43 The tool invokes `git log` internally unless you pipe a log to it.
  44 Thus, typically you would use it as follows:
  45
  46 =head1 SEE ALSO
  47
  48 Output format specification:
  49     <http://www.gnu.org/prep/standards/html_node/Change-Logs.html>
  50
  51 =head1 AUTHORS
  52
  53 git2cl is developed by Simon Josefsson <simon@josefsson.org>
  54                    and Luis Mondesi <lemsx1@gmail.com>
  55
  56 =cut
  57
  58 use strict;
  59 use POSIX qw(strftime);
  60 use Text::Wrap qw(wrap);
  61 use FileHandle;
  62
  63 use constant EMPTY_LOG_MESSAGE => '*** empty log message ***';
  64
  65 # this is a helper hash for stptime.
  66 # Assumes you are calling 'git log ...' with LC_ALL=C
  67 my %month = (
  68     'Jan'=>0,
  69     'Feb'=>1,
  70     'Mar'=>2,
  71     'Apr'=>3,
  72     'May'=>4,
  73     'Jun'=>5,
  74     'Jul'=>6,
  75     'Aug'=>7,
  76     'Sep'=>8,
  77     'Oct'=>9,
  78     'Nov'=>10,
  79     'Dec'=>11,
  80 );
  81
  82 my $fh = new FileHandle;
  83
  84 sub strptime {
  85     my $str = shift;
  86     return undef if not defined $str;
  87
  88     # we are parsing this format
  89     # Fri Oct 26 00:42:56 2007 -0400
  90     # to these fields
  91     # sec, min, hour, mday, mon, year, wday = -1, yday = -1, isdst = -1
  92     # Luis Mondesi <lemsx1@gmail.com>
  93     my @date;
  94     if ($str =~ /([[:alpha:]]{3})\s+([[:alpha:]]{3})\s+([[:digit:]]{1,2})\s+([[:digit:]]{1,2}):([[:digit:]]{1,2}):([[:digit:]]{1,2})\s+([[:digit:]]{4})/){
  95         push(@date,$6,$5,$4,$3,$month{$2},($7 - 1900),-1,-1,-1);
  96     } else {
  97         die ("Cannot parse date '$str'\n'");
  98     }
  99     return @date;
 100 }
 101
 102 sub mywrap {
 103     my ($indent1, $indent2, @text) = @_;
 104     # If incoming text looks preformatted, don't get clever
 105     my $text = Text::Wrap::wrap($indent1, $indent2, @text);
 106     if ( grep /^\s+/m, @text ) {
 107         return $text;
 108     }
 109     my @lines = split /\n/, $text;
 110     $indent2 =~ s!^((?: {8})+)!"\t" x (length($1)/8)!e;
 111     $lines[0] =~ s/^$indent1\s+/$indent1/;
 112     s/^$indent2\s+/$indent2/
 113         for @lines[1..$#lines];
 114     my $newtext = join "\n", @lines;
 115     $newtext .= "\n"
 116         if substr($text, -1) eq "\n";
 117     return $newtext;
 118 }
 119
 120 sub last_line_len {
 121     my $files_list = shift;
 122     my @lines = split (/\n/, $files_list);
 123     my $last_line = pop (@lines);
 124     return length ($last_line);
 125 }
 126
 127 # A custom wrap function, sensitive to some common constructs used in
 128 # log entries.
 129 sub wrap_log_entry {
 130     my $text = shift;                  # The text to wrap.
 131     my $left_pad_str = shift;          # String to pad with on the left.
 132
 133     # These do NOT take left_pad_str into account:
 134     my $length_remaining = shift;      # Amount left on current line.
 135     my $max_line_length  = shift;      # Amount left for a blank line.
 136
 137     my $wrapped_text = '';             # The accumulating wrapped entry.
 138     my $user_indent = '';              # Inherited user_indent from prev line.
 139
 140     my $first_time = 1;                # First iteration of the loop?
 141     my $suppress_line_start_match = 0; # Set to disable line start checks.
 142
 143     my @lines = split (/\n/, $text);
 144     while (@lines)   # Don't use `foreach' here, it won't work.
 145     {
 146         my $this_line = shift (@lines);
 147         chomp $this_line;
 148
 149         if ($this_line =~ /^(\s+)/) {
 150             $user_indent = $1;
 151         }
 152         else {
 153             $user_indent = '';
 154         }
 155
 156         # If it matches any of the line-start regexps, print a newline now...
 157         if ($suppress_line_start_match)
 158         {
 159             $suppress_line_start_match = 0;
 160         }
 161         elsif (($this_line =~ /^(\s*)\*\s+[a-zA-Z0-9]/)
 162                || ($this_line =~ /^(\s*)\* [a-zA-Z0-9_\.\/\+-]+/)
 163                || ($this_line =~ /^(\s*)\([a-zA-Z0-9_\.\/\+-]+(\)|,\s*)/)
 164                || ($this_line =~ /^(\s+)(\S+)/)
 165                || ($this_line =~ /^(\s*)- +/)
 166                || ($this_line =~ /^()\s*$/)
 167                || ($this_line =~ /^(\s*)\*\) +/)
 168                || ($this_line =~ /^(\s*)[a-zA-Z0-9](\)|\.|\:) +/))
 169         {
 170             $length_remaining = $max_line_length - (length ($user_indent));
 171         }
 172
 173         # Now that any user_indent has been preserved, strip off leading
 174         # whitespace, so up-folding has no ugly side-effects.
 175         $this_line =~ s/^\s*//;
 176
 177         # Accumulate the line, and adjust parameters for next line.
 178         my $this_len = length ($this_line);
 179         if ($this_len == 0)
 180         {
 181             # Blank lines should cancel any user_indent level.
 182             $user_indent = '';
 183             $length_remaining = $max_line_length;
 184         }
 185         elsif ($this_len >= $length_remaining) # Line too long, try breaking it.
 186         {
 187             # Walk backwards from the end.  At first acceptable spot, break
 188             # a new line.
 189             my $idx = $length_remaining - 1;
 190             if ($idx < 0) { $idx = 0 };
 191             while ($idx > 0)
 192             {
 193                 if (substr ($this_line, $idx, 1) =~ /\s/)
 194                 {
 195                     my $line_now = substr ($this_line, 0, $idx);
 196                     my $next_line = substr ($this_line, $idx);
 197                     $this_line = $line_now;
 198
 199                     # Clean whitespace off the end.
 200                     chomp $this_line;
 201
 202                     # The current line is ready to be printed.
 203                     $this_line .= "\n${left_pad_str}";
 204
 205                     # Make sure the next line is allowed full room.
 206                     $length_remaining = $max_line_length - (length ($user_indent));
 207
 208                     # Strip next_line, but then preserve any user_indent.
 209                     $next_line =~ s/^\s*//;
 210
 211                     # Sneak a peek at the user_indent of the upcoming line, so
 212                     # $next_line (which will now precede it) can inherit that
 213                     # indent level.  Otherwise, use whatever user_indent level
 214                     # we currently have, which might be none.
 215                     my $next_next_line = shift (@lines);
 216                     if ((defined ($next_next_line)) && ($next_next_line =~ /^(\s+)/)) {
 217                         $next_line = $1 . $next_line if (defined ($1));
 218                         # $length_remaining = $max_line_length - (length ($1));
 219                         $next_next_line =~ s/^\s*//;
 220                     }
 221                     else {
 222                         $next_line = $user_indent . $next_line;
 223                     }
 224                     if (defined ($next_next_line)) {
 225                         unshift (@lines, $next_next_line);
 226                     }
 227                     unshift (@lines, $next_line);
 228
 229                     # Our new next line might, coincidentally, begin with one of
 230                     # the line-start regexps, so we temporarily turn off
 231                     # sensitivity to that until we're past the line.
 232                     $suppress_line_start_match = 1;
 233
 234                     last;
 235                 }
 236                 else
 237                 {
 238                     $idx--;
 239                 }
 240             }
 241
 242             if ($idx == 0)
 243             {
 244                 # We bottomed out because the line is longer than the
 245                 # available space.  But that could be because the space is
 246                 # small, or because the line is longer than even the maximum
 247                 # possible space.  Handle both cases below.
 248
 249                 if ($length_remaining == ($max_line_length - (length ($user_indent))))
 250                 {
 251                     # The line is simply too long -- there is no hope of ever
 252                     # breaking it nicely, so just insert it verbatim, with
 253                     # appropriate padding.
 254                     $this_line = "\n${left_pad_str}${this_line}";
 255                 }
 256                 else
 257                 {
 258                     # Can't break it here, but may be able to on the next round...
 259                     unshift (@lines, $this_line);
 260                     $length_remaining = $max_line_length - (length ($user_indent));
 261                     $this_line = "\n${left_pad_str}";
 262                 }
 263             }
 264         }
 265         else  # $this_len < $length_remaining, so tack on what we can.
 266         {
 267             # Leave a note for the next iteration.
 268             $length_remaining = $length_remaining - $this_len;
 269
 270             if ($this_line =~ /\.$/)
 271             {
 272                 $this_line .= "  ";
 273                 $length_remaining -= 2;
 274             }
 275             else  # not a sentence end
 276             {
 277                 $this_line .= " ";
 278                 $length_remaining -= 1;
 279             }
 280         }
 281
 282         # Unconditionally indicate that loop has run at least once.
 283         $first_time = 0;
 284
 285         $wrapped_text .= "${user_indent}${this_line}";
 286     }
 287
 288     # One last bit of padding.
 289     $wrapped_text .= "\n";
 290
 291     return $wrapped_text;
 292 }
 293
 294 # main
 295
 296 my @date;
 297 my $author;
 298 my @files;
 299 my $comment;
 300
 301 my $state; # 0-header 1-comment 2-files
 302 my $done = 0;
 303
 304 $state = 0;
 305
 306 # if reading from STDIN, we assume that we are
 307 # getting git log as input
 308 if (-f STDIN or -l STDIN or -p STDIN)
 309 {
 310     #my $dummyfh; # don't care about writing
 311     #($fh,$dummyfh) = FileHandle::pipe;
 312     $fh->fdopen(*STDIN, 'r');
 313 }
 314 else
 315 {
 316     $fh->open("LC_ALL=C git log --pretty --numstat --summary|")
 317         or die("Cannot execute git log...$!\n");
 318 }
 319
 320 while (my $_l = <$fh>) {
 321     #print STDERR "debug ($state, " . (@date ? (strftime "%Y-%m-%d", @date) : "") . "): `$_'\n";
 322     if ($state == 0) {
 323         if ($_l =~ m,^Author: (.*),) {
 324             $author = $1;
 325         }
 326         if ($_l =~ m,^Date: (.*),) {
 327             @date = strptime($1);
 328         }
 329         $state = 1 if ($_l =~ m,^$, and $author and (@date+0>0));
 330     } elsif ($state == 1) {
 331         # * modifying our input text is a bad choice
 332         #   let's make a copy of it first, then we remove spaces
 333         # * if we meet a "merge branch" statement, we need to start
 334         #   over and find a real entry
 335         # Luis Mondesi <lemsx1@gmail.com>
 336         my $_s = $_l;
 337         $_s =~ s/^    //g;
 338         if ($_s =~ m/^Merge branch/)
 339         {
 340             $state=0;
 341             next;
 342         }
 343         $comment = $comment . $_s;
 344         $state = 2 if ($_l =~ m,^$,);
 345     } elsif ($state == 2) {
 346         if ($_l =~ m,^([0-9]+)\t([0-9]+)\t(.*)$,) {
 347             push @files, $3;
 348         }
 349         $done = 1 if ($_l =~ m,^$,);
 350     }
 351
 352     if ($done) {
 353         print (strftime "%Y-%m-%d  $author\n\n", @date);
 354
 355         my $files = join (", ", @files);
 356         $files = mywrap ("\t", "\t", "* $files"), ": ";
 357
 358         if (index($comment, EMPTY_LOG_MESSAGE) > -1 ) {
 359             $comment = "[no log message]\n";
 360         }
 361
 362         my $files_last_line_len = 0;
 363         $files_last_line_len = last_line_len($files) + 1;
 364         my $msg = wrap_log_entry($comment, "\t", 69-$files_last_line_len, 69);
 365
 366         $msg =~ s/[ \t]+\n/\n/g;
 367
 368         print "$files: $msg\n";
 369
 370         @date = ();
 371         $author = "";
 372         @files = ();
 373         $comment = "";
 374
 375         $state = 0;
 376         $done = 0;
 377     }
 378 }
 379
 380 if (@date + 0)
 381 {
 382     print (strftime "%Y-%m-%d  $author\n\n", @date);
 383     my $msg = wrap_log_entry($comment, "\t", 69, 69);
 384     $msg =~ s/[ \t]+\n/\n/g;
 385     print "\t* $msg\n";
 386 }