contrib/cvs/contrib/rcs2log.sh

   1 #! /bin/sh
   2
   3 # RCS to ChangeLog generator
   4
   5 # Generate a change log prefix from RCS files and the ChangeLog (if any).
   6 # Output the new prefix to standard output.
   7 # You can edit this prefix by hand, and then prepend it to ChangeLog.
   8
   9 # Ignore log entries that start with `#'.
  10 # Clump together log entries that start with `{topic} ',
  11 # where `topic' contains neither white space nor `}'.
  12
  13 # Author: Paul Eggert <eggert@twinsun.com>
  14
  15 # Copyright 1992, 1993, 1994, 1995 Free Software Foundation, Inc.
  16
  17 # This program is free software; you can redistribute it and/or modify
  18 # it under the terms of the GNU General Public License as published by
  19 # the Free Software Foundation; either version 2, or (at your option)
  20 # any later version.
  21 #
  22 # This program is distributed in the hope that it will be useful,
  23 # but WITHOUT ANY WARRANTY; without even the implied warranty of
  24 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  25 # GNU General Public License for more details.
  26 #
  27 # You should have received a copy of the GNU General Public License
  28 # along with this program; see the file COPYING.  If not, write to the
  29 # Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  30 # Boston, MA 02111-1307, USA.
  31
  32 tab='   '
  33 nl='
  34 '
  35
  36 # Parse options.
  37
  38 # defaults
  39 : ${AWK=awk}
  40 : ${TMPDIR=/tmp}
  41 hostname= # name of local host (if empty, will deduce it later)
  42 indent=8 # indent of log line
  43 length=79 # suggested max width of log line
  44 logins= # login names for people we know fullnames and mailaddrs of
  45 loginFullnameMailaddrs= # login<tab>fullname<tab>mailaddr triplets
  46 recursive= # t if we want recursive rlog
  47 rlog_options= # options to pass to rlog
  48 tabwidth=8 # width of horizontal tab
  49
  50 while :
  51 do
  52         case $1 in
  53         -i)     indent=${2?}; shift;;
  54         -h)     hostname=${2?}; shift;;
  55         -l)     length=${2?}; shift;;
  56         -[nu])  # -n is obsolescent; it is replaced by -u.
  57                 case $1 in
  58                 -n)     case ${2?}${3?}${4?} in
  59                         *"$tab"* | *"$nl"*)
  60                                 echo >&2 "$0: -n '$2' '$3' '$4': tabs, newlines not allowed"
  61                                 exit 1
  62                         esac
  63                         loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$2$tab$3$tab$4
  64                         shift; shift; shift;;
  65                 -u)
  66                         # If $2 is not tab-separated, use colon for separator.
  67                         case ${2?} in
  68                         *"$nl"*)
  69                                 echo >&2 "$0: -u '$2': newlines not allowed"
  70                                 exit 1;;
  71                         *"$tab"*)
  72                                 t=$tab;;
  73                         *)
  74                                 t=:
  75                         esac
  76                         case $2 in
  77                         *"$t"*"$t"*"$t"*)
  78                                 echo >&2 "$0: -u '$2': too many fields"
  79                                 exit 1;;
  80                         *"$t"*"$t"*)
  81                                 ;;
  82                         *)
  83                                 echo >&2 "$0: -u '$2': not enough fields"
  84                                 exit 1
  85                         esac
  86                         loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$2
  87                         shift
  88                 esac
  89                 logins=$logins$nl$login
  90                 ;;
  91         -r)     rlog_options=$rlog_options$nl${2?}; shift;;
  92         -R)     recursive=t;;
  93         -t)     tabwidth=${2?}; shift;;
  94         -*)     echo >&2 "$0: usage: $0 [options] [file ...]
  95 Options:
  96         [-h hostname] [-i indent] [-l length] [-R] [-r rlog_option]
  97         [-t tabwidth] [-u 'login<TAB>fullname<TAB>mailaddr']..."
  98                 exit 1;;
  99         *)      break
 100         esac
 101         shift
 102 done
 103
 104 month_data='
 105         m[0]="Jan"; m[1]="Feb"; m[2]="Mar"
 106         m[3]="Apr"; m[4]="May"; m[5]="Jun"
 107         m[6]="Jul"; m[7]="Aug"; m[8]="Sep"
 108         m[9]="Oct"; m[10]="Nov"; m[11]="Dec"
 109
 110         # days in non-leap year thus far, indexed by month (0-12)
 111         mo[0]=0; mo[1]=31; mo[2]=59; mo[3]=90
 112         mo[4]=120; mo[5]=151; mo[6]=181; mo[7]=212
 113         mo[8]=243; mo[9]=273; mo[10]=304; mo[11]=334
 114         mo[12]=365
 115 '
 116
 117
 118 # Put rlog output into $rlogout.
 119
 120 # If no rlog options are given,
 121 # log the revisions checked in since the first ChangeLog entry.
 122 case $rlog_options in
 123 '')
 124         date=1970-01-01
 125         if test -s ChangeLog
 126         then
 127                 # Add 1 to seconds to avoid duplicating most recent log.
 128                 e='
 129                         /^... ... [ 0-9][0-9] [ 0-9][0-9]:[0-9][0-9]:[0-9][0-9] [0-9]+ /{
 130                                 '"$month_data"'
 131                                 year = $5
 132                                 for (i=0; i<=11; i++) if (m[i] == $2) break
 133                                 dd = $3
 134                                 hh = substr($0,12,2)
 135                                 mm = substr($0,15,2)
 136                                 ss = substr($0,18,2)
 137                                 ss++
 138                                 if (ss == 60) {
 139                                         ss = 0
 140                                         mm++
 141                                         if (mm == 60) {
 142                                                 mm = 0
 143                                                 hh++
 144                                                 if (hh == 24) {
 145                                                         hh = 0
 146                                                         dd++
 147                                                         monthdays = mo[i+1] - mo[i]
 148                                                         if (i == 1 && year%4 == 0 && (year%100 != 0 || year%400 == 0)) monthdays++
 149                                                         if (dd == monthdays + 1) {
 150                                                                 dd = 1
 151                                                                 i++
 152                                                                 if (i == 12) {
 153                                                                         i = 0
 154                                                                         year++
 155                                                                 }
 156                                                         }
 157                                                 }
 158                                         }
 159                                 }
 160                                 printf "%02d/%02d/%d %02d:%02d:%02d\n", i+1,dd,year,hh,mm,ss
 161                                 exit
 162                         }
 163                 '
 164                 d=`$AWK "$e" <ChangeLog` || exit
 165                 case $d in
 166                 ?*) date=$d
 167                 esac
 168         fi
 169         datearg="-d>$date"
 170 esac
 171
 172 # If CVS is in use, examine its repository, not the normal RCS files.
 173 if test ! -f CVS/Repository
 174 then
 175         rlog=rlog
 176         repository=
 177 else
 178         rlog='cvs log'
 179         repository=`sed 1q <CVS/Repository` || exit
 180         test ! -f CVS/Root || CVSROOT=`cat <CVS/Root` || exit
 181         case $CVSROOT in
 182         *:/*)
 183                 # remote repository
 184                 ;;
 185         *)
 186                 # local repository
 187                 case $repository in
 188                 /*) ;;
 189                 *) repository=${CVSROOT?}/$repository
 190                 esac
 191                 if test ! -d "$repository"
 192                 then
 193                         echo >&2 "$0: $repository: bad repository (see CVS/Repository)"
 194                         exit 1
 195                 fi
 196         esac
 197 fi
 198
 199 # With no arguments, examine all files under the RCS directory.
 200 case $# in
 201 0)
 202         case $repository in
 203         '')
 204                 oldIFS=$IFS
 205                 IFS=$nl
 206                 case $recursive in
 207                 t)
 208                         RCSdirs=`find . -name RCS -type d -print`
 209                         filesFromRCSfiles='s|,v$||; s|/RCS/|/|; s|^\./||'
 210                         files=`
 211                                 {
 212                                         case $RCSdirs in
 213                                         ?*) find $RCSdirs -type f -print
 214                                         esac
 215                                         find . -name '*,v' -print
 216                                 } |
 217                                 sort -u |
 218                                 sed "$filesFromRCSfiles"
 219                         `;;
 220                 *)
 221                         files=
 222                         for file in RCS/.* RCS/* .*,v *,v
 223                         do
 224                                 case $file in
 225                                 RCS/. | RCS/..) continue;;
 226                                 RCS/.\* | RCS/\* | .\*,v | \*,v) test -f "$file" || continue
 227                                 esac
 228                                 files=$files$nl$file
 229                         done
 230                         case $files in
 231                         '') exit 0
 232                         esac
 233                 esac
 234                 set x $files
 235                 shift
 236                 IFS=$oldIFS
 237         esac
 238 esac
 239
 240 llogout=$TMPDIR/rcs2log$$l
 241 rlogout=$TMPDIR/rcs2log$$r
 242 trap exit 1 2 13 15
 243 trap "rm -f $llogout $rlogout; exit 1" 0
 244
 245 case $rlog_options in
 246 ?*) $rlog $rlog_options ${1+"$@"} >$rlogout;;
 247 '') $rlog "$datearg" ${1+"$@"} >$rlogout
 248 esac || exit
 249
 250
 251 # Get the full name of each author the logs mention, and set initialize_fullname
 252 # to awk code that initializes the `fullname' awk associative array.
 253 # Warning: foreign authors (i.e. not known in the passwd file) are mishandled;
 254 # you have to fix the resulting output by hand.
 255
 256 initialize_fullname=
 257 initialize_mailaddr=
 258
 259 case $loginFullnameMailaddrs in
 260 ?*)
 261         case $loginFullnameMailaddrs in
 262         *\"* | *\\*)
 263                 sed 's/["\\]/\\&/g' >$llogout <<EOF || exit
 264 $loginFullnameMailaddrs
 265 EOF
 266                 loginFullnameMailaddrs=`cat $llogout`
 267         esac
 268
 269         oldIFS=$IFS
 270         IFS=$nl
 271         for loginFullnameMailaddr in $loginFullnameMailaddrs
 272         do
 273                 case $loginFullnameMailaddr in
 274                 *"$tab"*) IFS=$tab;;
 275                 *) IFS=:
 276                 esac
 277                 set x $loginFullnameMailaddr
 278                 login=$2
 279                 fullname=$3
 280                 mailaddr=$4
 281                 initialize_fullname="$initialize_fullname
 282                         fullname[\"$login\"] = \"$fullname\""
 283                 initialize_mailaddr="$initialize_mailaddr
 284                         mailaddr[\"$login\"] = \"$mailaddr\""
 285         done
 286         IFS=$oldIFS
 287 esac
 288
 289 case $llogout in
 290 ?*) sort -u -o $llogout <<EOF || exit
 291 $logins
 292 EOF
 293 esac
 294 output_authors='/^date: / {
 295         if ($2 ~ /^[0-9]*[-\/][0-9][0-9][-\/][0-9][0-9]$/ && $3 ~ /^[0-9][0-9]:[0-9][0-9]:[0-9][0-9][-+0-9:]*;$/ && $4 == "author:" && $5 ~ /^[^;]*;$/) {
 296                 print substr($5, 1, length($5)-1)
 297         }
 298 }'
 299 authors=`
 300         $AWK "$output_authors" <$rlogout |
 301         case $llogout in
 302         '') sort -u;;
 303         ?*) sort -u | comm -23 - $llogout
 304         esac
 305 `
 306 case $authors in
 307 ?*)
 308         cat >$llogout <<EOF || exit
 309 $authors
 310 EOF
 311         initialize_author_script='s/["\\]/\\&/g; s/.*/author[\"&\"] = 1/'
 312         initialize_author=`sed -e "$initialize_author_script" <$llogout`
 313         awkscript='
 314                 BEGIN {
 315                         alphabet = "abcdefghijklmnopqrstuvwxyz"
 316                         ALPHABET = "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 317                         '"$initialize_author"'
 318                 }
 319                 {
 320                         if (author[$1]) {
 321                                 fullname = $5
 322                                 if (fullname ~ /[0-9]+-[^(]*\([0-9]+\)$/) {
 323                                         # Remove the junk from fullnames like "0000-Admin(0000)".
 324                                         fullname = substr(fullname, index(fullname, "-") + 1)
 325                                         fullname = substr(fullname, 1, index(fullname, "(") - 1)
 326                                 }
 327                                 if (fullname ~ /,[^ ]/) {
 328                                         # Some sites put comma-separated junk after the fullname.
 329                                         # Remove it, but leave "Bill Gates, Jr" alone.
 330                                         fullname = substr(fullname, 1, index(fullname, ",") - 1)
 331                                 }
 332                                 abbr = index(fullname, "&")
 333                                 if (abbr) {
 334                                         a = substr($1, 1, 1)
 335                                         A = a
 336                                         i = index(alphabet, a)
 337                                         if (i) A = substr(ALPHABET, i, 1)
 338                                         fullname = substr(fullname, 1, abbr-1) A substr($1, 2) substr(fullname, abbr+1)
 339                                 }
 340
 341                                 # Quote quotes and backslashes properly in full names.
 342                                 # Do not use gsub; traditional awk lacks it.
 343                                 quoted = ""
 344                                 rest = fullname
 345                                 for (;;) {
 346                                         p = index(rest, "\\")
 347                                         q = index(rest, "\"")
 348                                         if (p) {
 349                                                 if (q && q<p) p = q
 350                                         } else {
 351                                                 if (!q) break
 352                                                 p = q
 353                                         }
 354                                         quoted = quoted substr(rest, 1, p-1) "\\" substr(rest, p, 1)
 355                                         rest = substr(rest, p+1)
 356                                 }
 357
 358                                 printf "fullname[\"%s\"] = \"%s%s\"\n", $1, quoted, rest
 359                                 author[$1] = 0
 360                         }
 361                 }
 362         '
 363
 364         initialize_fullname=`
 365                 (
 366                         cat /etc/passwd
 367                         for author in $authors
 368                         do nismatch $author passwd.org_dir
 369                         done
 370                         ypmatch $authors passwd
 371                 ) 2>/dev/null |
 372                 $AWK -F: "$awkscript"
 373         `$initialize_fullname
 374 esac
 375
 376
 377 # Function to print a single log line.
 378 # We don't use awk functions, to stay compatible with old awk versions.
 379 # `Log' is the log message (with \n replaced by \r).
 380 # `files' contains the affected files.
 381 printlogline='{
 382
 383         # Following the GNU coding standards, rewrite
 384         #       * file: (function): comment
 385         # to
 386         #       * file (function): comment
 387         if (Log ~ /^\([^)]*\): /) {
 388                 i = index(Log, ")")
 389                 files = files " " substr(Log, 1, i)
 390                 Log = substr(Log, i+3)
 391         }
 392
 393         # If "label: comment" is too long, break the line after the ":".
 394         sep = " "
 395         if ('"$length"' <= '"$indent"' + 1 + length(files) + index(Log, CR)) sep = "\n" indent_string
 396
 397         # Print the label.
 398         printf "%s*%s:", indent_string, files
 399
 400         # Print each line of the log, transliterating \r to \n.
 401         while ((i = index(Log, CR)) != 0) {
 402                 logline = substr(Log, 1, i-1)
 403                 if (logline ~ /[^'"$tab"' ]/) {
 404                         printf "%s%s\n", sep, logline
 405                 } else {
 406                         print ""
 407                 }
 408                 sep = indent_string
 409                 Log = substr(Log, i+1)
 410         }
 411 }'
 412
 413 case $hostname in
 414 '')
 415         hostname=`(
 416                 hostname || uname -n || uuname -l || cat /etc/whoami
 417         ) 2>/dev/null` || {
 418                 echo >&2 "$0: cannot deduce hostname"
 419                 exit 1
 420         }
 421
 422         case $hostname in
 423         *.*) ;;
 424         *)
 425                 domainname=`(domainname) 2>/dev/null` &&
 426                 case $domainname in
 427                 *.*) hostname=$hostname.$domainname
 428                 esac
 429         esac
 430 esac
 431
 432
 433 # Process the rlog output, generating ChangeLog style entries.
 434
 435 # First, reformat the rlog output so that each line contains one log entry.
 436 # Transliterate \n to \r so that multiline entries fit on a single line.
 437 # Discard irrelevant rlog output.
 438 $AWK <$rlogout '
 439         BEGIN { repository = "'"$repository"'" }
 440         /^RCS file:/ {
 441                 if (repository != "") {
 442                         filename = $3
 443                         if (substr(filename, 1, length(repository) + 1) == repository "/") {
 444                                 filename = substr(filename, length(repository) + 2)
 445                         }
 446                         if (filename ~ /,v$/) {
 447                                 filename = substr(filename, 1, length(filename) - 2)
 448                         }
 449                 }
 450         }
 451         /^Working file:/ { if (repository == "") filename = $3 }
 452         /^date: /, /^(-----------*|===========*)$/ {
 453                 if ($0 ~ /^branches: /) { next }
 454                 if ($0 ~ /^date: [0-9][- +\/0-9:]*;/) {
 455                         date = $2
 456                         if (date ~ /-/) {
 457                                 # An ISO format date.  Replace all "-"s with "/"s.
 458                                 newdate = ""
 459                                 while ((i = index(date, "-")) != 0) {
 460                                         newdate = newdate substr(date, 1, i-1) "/"
 461                                         date = substr(date, i+1)
 462                                 }
 463                                 date = newdate date
 464                         }
 465                         # Ignore any time zone; ChangeLog has no room for it.
 466                         time = substr($3, 1, 8)
 467                         author = substr($5, 1, length($5)-1)
 468                         printf "%s %s %s %s %c", filename, date, time, author, 13
 469                         next
 470                 }
 471                 if ($0 ~ /^(-----------*|===========*)$/) { print ""; next }
 472                 printf "%s%c", $0, 13
 473         }
 474 ' |
 475
 476 # Now each line is of the form
 477 # FILENAME YYYY/MM/DD HH:MM:SS AUTHOR \rLOG
 478 #       where \r stands for a carriage return,
 479 #       and each line of the log is terminated by \r instead of \n.
 480 # Sort the log entries, first by date+time (in reverse order),
 481 # then by author, then by log entry, and finally by file name (just in case).
 482 sort +1 -3r +3 +0 |
 483
 484 # Finally, reformat the sorted log entries.
 485 $AWK '
 486         BEGIN {
 487                 # Some awk variants do not understand "\r" or "\013", so we have to
 488                 # put a carriage return directly in the file.
 489                 CR="\r" # <-- There is a single CR between the " chars here.
 490
 491                 # Initialize the fullname and mailaddr associative arrays.
 492                 '"$initialize_fullname"'
 493                 '"$initialize_mailaddr"'
 494
 495                 # Initialize indent string.
 496                 indent_string = ""
 497                 i = '"$indent"'
 498                 if (0 < '"$tabwidth"')
 499                         for (;  '"$tabwidth"' <= i;  i -= '"$tabwidth"')
 500                                 indent_string = indent_string "\t"
 501                 while (1 <= i--)
 502                         indent_string = indent_string " "
 503
 504                 # Set up date conversion tables.
 505                 # RCS uses a nice, clean, sortable format,
 506                 # but ChangeLog wants the traditional, ugly ctime format.
 507
 508                 # January 1, 0 AD (Gregorian) was Saturday = 6
 509                 EPOCH_WEEKDAY = 6
 510                 # Of course, there was no 0 AD, but the algorithm works anyway.
 511
 512                 w[0]="Sun"; w[1]="Mon"; w[2]="Tue"; w[3]="Wed"
 513                 w[4]="Thu"; w[5]="Fri"; w[6]="Sat"
 514
 515                 '"$month_data"'
 516         }
 517
 518         {
 519                 newlog = substr($0, 1 + index($0, CR))
 520
 521                 # Ignore log entries prefixed by "#".
 522                 if (newlog ~ /^#/) { next }
 523
 524                 if (Log != newlog || date != $2 || author != $4) {
 525
 526                         # The previous log and this log differ.
 527
 528                         # Print the old log.
 529                         if (date != "") '"$printlogline"'
 530
 531                         # Logs that begin with "{clumpname} " should be grouped together,
 532                         # and the clumpname should be removed.
 533                         # Extract the new clumpname from the log header,
 534                         # and use it to decide whether to output a blank line.
 535                         newclumpname = ""
 536                         sep = "\n"
 537                         if (date == "") sep = ""
 538                         if (newlog ~ /^\{[^'"$tab"' }]*}['"$tab"' ]/) {
 539                                 i = index(newlog, "}")
 540                                 newclumpname = substr(newlog, 1, i)
 541                                 while (substr(newlog, i+1) ~ /^['"$tab"' ]/) i++
 542                                 newlog = substr(newlog, i+1)
 543                                 if (clumpname == newclumpname) sep = ""
 544                         }
 545                         printf sep
 546                         clumpname = newclumpname
 547
 548                         # Get ready for the next log.
 549                         Log = newlog
 550                         if (files != "")
 551                                 for (i in filesknown)
 552                                         filesknown[i] = 0
 553                         files = ""
 554                 }
 555                 if (date != $2  ||  author != $4) {
 556                         # The previous date+author and this date+author differ.
 557                         # Print the new one.
 558                         date = $2
 559                         author = $4
 560
 561                         # Convert nice RCS date like "1992/01/03 00:03:44"
 562                         # into ugly ctime date like "Fri Jan  3 00:03:44 1992".
 563                         # Calculate day of week from Gregorian calendar.
 564                         i = index($2, "/")
 565                         year = substr($2, 1, i-1) + 0
 566                         monthday = substr($2, i+1)
 567                         i = index(monthday, "/")
 568                         month = substr(monthday, 1, i-1) + 0
 569                         day = substr(monthday, i+1) + 0
 570                         leap = 0
 571                         if (2 < month && year%4 == 0 && (year%100 != 0 || year%400 == 0)) leap = 1
 572                         days_since_Sunday_before_epoch = EPOCH_WEEKDAY + year * 365 + int((year + 3) / 4) - int((year + 99) / 100) + int((year + 399) / 400) + mo[month-1] + leap + day - 1
 573
 574                         # Print "date  fullname  (email address)".
 575                         # Get fullname and email address from associative arrays;
 576                         # default to author and author@hostname if not in arrays.
 577                         if (fullname[author])
 578                                 auth = fullname[author]
 579                         else
 580                                 auth = author
 581                         printf "%s %s %2d %s %d  %s  ", w[days_since_Sunday_before_epoch%7], m[month-1], day, $3, year, auth
 582                         if (mailaddr[author])
 583                                 printf "<%s>\n\n", mailaddr[author]
 584                         else
 585                                 printf "<%s@%s>\n\n", author, "'"$hostname"'"
 586                 }
 587                 if (! filesknown[$1]) {
 588                         filesknown[$1] = 1
 589                         if (files == "") files = " " $1
 590                         else files = files ", " $1
 591                 }
 592         }
 593         END {
 594                 # Print the last log.
 595                 if (date != "") {
 596                         '"$printlogline"'
 597                         printf "\n"
 598                 }
 599         }
 600 ' &&
 601
 602
 603 # Exit successfully.
 604
 605 exec rm -f $llogout $rlogout