changeset 39150:a56fcade3c7e

(Help, mainline code): Add new option -L FILE. (Copyright): Update year. (LANG, LANGUAGE, LC_ALL, LC_COLLATE, LC_CTYPE, LC_MESSAGES, LC_NUMERIC, LC_TIME): New shell vars, to make sure we live in the C locale. (mainline code): Handle nonstandard -u option differently, by transforming it to standard form. Check for "Working file: ", not "Working file:". Allow file names with spaces. (SOH, rlogfile): New shell vars. (rlogout): Remove. Its old functionality is mostly migrated to rlogfile. Append ';;' to the last arm of every case statement, for portability to ancient broken BSD shells. (logins): Fix bug; was not being computed at all, lowering performance. (pository): New var. This fixes some bugs where repositories are remote, or have trailing slashes. (authors): $llogout is never an empty shell var, so don't worry about that possibility. (printlogline, mainline code): Fix bug with SOH's being put into the output.

--- a/lib-src/rcs2log	Wed Sep 05 14:46:26 2001 +0000
+++ b/lib-src/rcs2log	Wed Sep 05 23:07:46 2001 +0000
@@ -18,6 +18,7 @@
   -h HOSTNAME  Use HOSTNAME in change log entries (default current host).
   -i INDENT  Indent change log lines by INDENT spaces (default 8).
   -l LENGTH  Try to limit log lines to LENGTH characters (default 79).
+  -L FILE  Use rlog-format FILE for source of logs.
   -R  If no FILEs are given and RCS is used, recurse through working directory.
   -r OPTION  Pass OPTION to subsidiary log command.
   -t TABWIDTH  Tab stops are every TABWIDTH characters (default 8).
@@ -28,9 +29,9 @@
 
 Report bugs to <[e-mail]>.'
 
-Id='$Id: rcs2log,v 1.46 2001/01/03 12:04:06 gerd Exp $'
+Id='$Id: rcs2log,v 1.47 2001/07/20 09:05:26 gerd Exp eggert $'
 
-# Copyright 1992, 93, 94, 95, 96, 97, 1998, 2001
+# Copyright 1992, 1993, 1994, 1995, 1996, 1997, 1998, 2001
 #  Free Software Foundation, Inc.
 
 # This program is free software; you can redistribute it and/or modify
@@ -48,13 +49,29 @@
 # Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 # Boston, MA 02111-1307, USA.
 
-Copyright='Copyright 1998 Free Software Foundation, Inc.
+Copyright='Copyright 2001 Free Software Foundation, Inc.
 This program comes with NO WARRANTY, to the extent permitted by law.
 You may redistribute copies of this program
 under the terms of the GNU General Public License.
 For more information about these matters, see the files named COPYING.
 Author: Paul Eggert <[e-mail]>'
 
+# Use the traditional C locale.
+LANG=C
+LANGUAGE=C
+LC_ALL=C
+LC_COLLATE=C
+LC_CTYPE=C
+LC_MESSAGES=C
+LC_NUMERIC=C
+LC_TIME=C
+export LANG LANGUAGE LC_ALL LC_COLLATE LC_CTYPE LC_MESSAGES LC_NUMERIC LC_TIME
+
+# These variables each contain a single ASCII character.
+# Unfortunately, there's no portable way of writing these characters
+# in older Unix implementations, other than putting them directly into
+# this text file.
+SOH='
' # SOH, octal code 001
 tab='	'
 nl='
 '
@@ -75,6 +92,7 @@
 recursive= # t if we want recursive rlog
 revision= # t if we want revision numbers
 rlog_options= # options to pass to rlog
+rlogfile= # log file to read from
 tabwidth=8 # width of horizontal tab
 
 while :
@@ -84,17 +102,16 @@
 	-i)	indent=${2?}; shift;;
 	-h)	hostname=${2?}; shift;;
 	-l)	length=${2?}; shift;;
+	-L)	rlogfile=${2?}; shift;;
 	-[nu])	# -n is obsolescent; it is replaced by -u.
 		case $1 in
 		-n)	case ${2?}${3?}${4?} in
 			*"$tab"* | *"$nl"*)
 				echo >&2 "$0: -n '$2' '$3' '$4': tabs, newlines not allowed"
-				exit 1
+				exit 1;;
 			esac
-			case $loginFullnameMailaddrs in
-			'') loginFullnameMailaddrs=$2$tab$3$tab$4;;
-			?*) loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$2$tab$3$tab$4
-			esac
+			login=$2
+			lfm=$2$tab$3$tab$4
 			shift; shift; shift;;
 		-u)
 			# If $2 is not tab-separated, use colon for separator.
@@ -105,33 +122,39 @@
 			*"$tab"*)
 				t=$tab;;
 			*)
-				t=:
+				t=':';;
 			esac
 			case $2 in
 			*"$t"*"$t"*"$t"*)
 				echo >&2 "$0: -u '$2': too many fields"
 				exit 1;;
 			*"$t"*"$t"*)
-				;;
+				uf="[^$t]*$t" # An unselected field, followed by a separator.
+				sf="\\([^$t]*\\)" # The selected field.
+				login=`expr "X$2" : "X$sf"`
+				lfm="$login$tab"`
+					expr "X$2" : "$uf$sf"
+				  `"$tab"`
+					expr "X$2" : "$uf$uf$sf"
+				`;;
 			*)
 				echo >&2 "$0: -u '$2': not enough fields"
-				exit 1
+				exit 1;;
 			esac
-			case $loginFullnameMailaddrs in
-			'') loginFullnameMailaddrs=$2;;
-			?*) loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$2
-			esac
-			shift
+			shift;;
 		esac
 		case $logins in
 		'') logins=$login;;
-		?*) logins=$logins$nl$login
+		?*) logins=$logins$nl$login;;
 		esac
-		;;
+		case $loginFullnameMailaddrs in
+		'') loginFullnameMailaddrs=$lfm;;
+		?*) loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$lfm;;
+		esac;;
 	-r)
 		case $rlog_options in
 		'') rlog_options=${2?};;
-		?*) rlog_options=$rlog_options$nl${2?}
+		?*) rlog_options=$rlog_options$nl${2?};;
 		esac
 		shift;;
 	-R)	recursive=t;;
@@ -145,9 +168,9 @@
 	-*)	echo >&2 "Usage: $0 [OPTION]... [FILE ...]$nl$Help"
 		case $1 in
 		--help) exit 0;;
-		*) exit 1
+		*) exit 1;;
 		esac;;
-	*)	break
+	*)	break;;
 	esac
 	shift
 done
@@ -159,159 +182,176 @@
 	m[9]="Oct"; m[10]="Nov"; m[11]="Dec"
 '
 
-
-# Put rlog output into $rlogout.
-
-# If no rlog options are given,
-# log the revisions checked in since the first ChangeLog entry.
-# Since ChangeLog is only by date, some of these revisions may be duplicates of
-# what's already in ChangeLog; it's the user's responsibility to remove them.
-case $rlog_options in
-'')
-	if test -s "$changelog"
-	then
-		e='
-			/^[0-9]+-[0-9][0-9]-[0-9][0-9]/{
-				# ISO 8601 date
-				print $1
-				exit
-			}
-			/^... ... [ 0-9][0-9] [ 0-9][0-9]:[0-9][0-9]:[0-9][0-9] [0-9]+ /{
-				# old-fashioned date and time (Emacs 19.31 and earlier)
-				'"$month_data"'
-				year = $5
-				for (i=0; i<=11; i++) if (m[i] == $2) break
-				dd = $3
-				printf "%d-%02d-%02d\n", year, i+1, dd
-				exit
-			}
-		'
-		d=`$AWK "$e" <"$changelog"` || exit
-		case $d in
-		?*) datearg="-d>$d"
-		esac
-	fi
-esac
-
-# Use TZ specified by ChangeLog local variable, if any.
-if test -s "$changelog"
-then
-	extractTZ='
-		/^.*change-log-time-zone-rule['"$tab"' ]*:['"$tab"' ]*"\([^"]*\)".*/{
-			s//\1/; p; q
-		}
-		/^.*change-log-time-zone-rule['"$tab"' ]*:['"$tab"' ]*t.*/{
-			s//UTC0/; p; q
-		}
-	'
-	logTZ=`tail "$changelog" | sed -n "$extractTZ"`
-	case $logTZ in
-	?*) TZ=$logTZ; export TZ
-	esac
-fi
-
-# If CVS is in use, examine its repository, not the normal RCS files.
-if test ! -f CVS/Repository
-then
-	rlog=rlog
-	repository=
-else
-	rlog='cvs -q log'
-	repository=`sed 1q <CVS/Repository` || exit
-	test ! -f CVS/Root || CVSROOT=`cat <CVS/Root` || exit
-	case $CVSROOT in
-	*:/*)
-		# remote repository
-		;;
-	*)
-		# local repository
-		case $repository in
-		/*) ;;
-		*) repository=${CVSROOT?}/$repository
-		esac
-		if test ! -d "$repository"
-		then
-			echo >&2 "$0: $repository: bad repository (see CVS/Repository)"
-			exit 1
-		fi
-	esac
-fi
-
-# Use $rlog's -zLT option, if $rlog supports it.
-case `$rlog -zLT 2>&1` in
-*' option'*) ;;
-*)
-	case $rlog_options in
-	'') rlog_options=-zLT;;
-	?*) rlog_options=-zLT$nl$rlog_options
-	esac
-esac
-
-# With no arguments, examine all files under the RCS directory.
-case $# in
-0)
-	case $repository in
-	'')
-		oldIFS=$IFS
-		IFS=$nl
-		case $recursive in
-		t)
-			RCSdirs=`find . -name RCS -type d -print`
-			filesFromRCSfiles='s|,v$||; s|/RCS/|/|; s|^\./||'
-			files=`
-				{
-					case $RCSdirs in
-					?*) find $RCSdirs \
-							-type f \
-							! -name '*_' \
-							! -name ',*,' \
-							! -name '.*_' \
-							! -name .rcsfreeze.log \
-							! -name .rcsfreeze.ver \
-							-print
-					esac
-					find . -name '*,v' -print
-				} |
-				sort -u |
-				sed "$filesFromRCSfiles"
-			`;;
-		*)
-			files=
-			for file in RCS/.* RCS/* .*,v *,v
-			do
-				case $file in
-				RCS/. | RCS/.. | RCS/,*, | RCS/*_) continue;;
-				RCS/.rcsfreeze.log | RCS/.rcsfreeze.ver) continue;;
-				RCS/.\* | RCS/\* | .\*,v | \*,v) test -f "$file" || continue;;
-				RCS/*,v | RCS/.*,v) ;;
-				RCS/* | RCS/.*) test -f "$file" || continue
-				esac
-				case $files in
-				'') files=$file;;
-				?*) files=$files$nl$file
-				esac
-			done
-			case $files in
-			'') exit 0
-			esac
-		esac
-		set x $files
-		shift
-		IFS=$oldIFS
-	esac
-esac
-
 logdir=$TMPDIR/rcs2log$$
 llogout=$logdir/l
-rlogout=$logdir/r
 trap exit 1 2 13 15
 trap "rm -fr $logdir 2>/dev/null" 0
 (umask 077 && exec mkdir $logdir) || exit
 
-case $datearg in
-?*) $rlog $rlog_options "$datearg" ${1+"$@"} >$rlogout;;
-'') $rlog $rlog_options ${1+"$@"} >$rlogout
-esac || exit
+# If no rlog-format log file is given, generate one into $rlogfile.
+case $rlogfile in
+'')
+	rlogfile=$logdir/r
+
+	# If no rlog options are given,
+	# log the revisions checked in since the first ChangeLog entry.
+	# Since ChangeLog is only by date, some of these revisions may be duplicates of
+	# what's already in ChangeLog; it's the user's responsibility to remove them.
+	case $rlog_options in
+	'')
+		if test -s "$changelog"
+		then
+			e='
+				/^[0-9]+-[0-9][0-9]-[0-9][0-9]/{
+					# ISO 8601 date
+					print $1
+					exit
+				}
+				/^... ... [ 0-9][0-9] [ 0-9][0-9]:[0-9][0-9]:[0-9][0-9] [0-9]+ /{
+					# old-fashioned date and time (Emacs 19.31 and earlier)
+					'"$month_data"'
+					year = $5
+					for (i=0; i<=11; i++) if (m[i] == $2) break
+					dd = $3
+					printf "%d-%02d-%02d\n", year, i+1, dd
+					exit
+				}
+			'
+			d=`$AWK "$e" <"$changelog"` || exit
+			case $d in
+			?*) datearg="-d>$d";;
+			esac
+		fi;;
+	esac
+
+	# Use TZ specified by ChangeLog local variable, if any.
+	if test -s "$changelog"
+	then
+		extractTZ='
+			/^.*change-log-time-zone-rule['"$tab"' ]*:['"$tab"' ]*"\([^"]*\)".*/{
+				s//\1/; p; q
+			}
+			/^.*change-log-time-zone-rule['"$tab"' ]*:['"$tab"' ]*t.*/{
+				s//UTC0/; p; q
+			}
+		'
+		logTZ=`tail "$changelog" | sed -n "$extractTZ"`
+		case $logTZ in
+		?*) TZ=$logTZ; export TZ;;
+		esac
+	fi
+
+	# If CVS is in use, examine its repository, not the normal RCS files.
+	if test ! -f CVS/Repository
+	then
+		rlog=rlog
+		repository=
+	else
+		rlog='cvs -q log'
+		repository=`sed 1q <CVS/Repository` || exit
+		test ! -f CVS/Root || CVSROOT=`cat <CVS/Root` || exit
+		case $CVSROOT in
+		*:/*:/*)
+			echo >&2 "$0: $CVSROOT: CVSROOT has multiple ':/'s"
+			exit 1;;
+		*:/*)
+			# remote repository
+			pository=`expr "X$repository" : '.*:\(/.*\)'`;;
+		*)
+			# local repository
+			case $repository in
+			/*) ;;
+			*) repository=${CVSROOT?}/$repository;;
+			esac
+			if test ! -d "$repository"
+			then
+				echo >&2 "$0: $repository: bad repository (see CVS/Repository)"
+				exit 1
+			fi
+			pository=$repository;;
+		esac
+
+		# Ensure that $pository ends in exactly one slash.
+		while :
+		do
+			case $pository in
+			*//) pository=`expr "X$pository" : 'X\(.*\)/'`;;
+			*/) break;;
+			*) pository=$pository/; break;;
+			esac
+		done
+
+	fi
+
+	# Use $rlog's -zLT option, if $rlog supports it.
+	case `$rlog -zLT 2>&1` in
+	*' option'*) ;;
+	*)
+		case $rlog_options in
+		'') rlog_options=-zLT;;
+		?*) rlog_options=-zLT$nl$rlog_options;;
+		esac;;
+	esac
+
+	# With no arguments, examine all files under the RCS directory.
+	case $# in
+	0)
+		case $repository in
+		'')
+			oldIFS=$IFS
+			IFS=$nl
+			case $recursive in
+			t)
+				RCSdirs=`find . -name RCS -type d -print`
+				filesFromRCSfiles='s|,v$||; s|/RCS/|/|; s|^\./||'
+				files=`
+					{
+						case $RCSdirs in
+						?*) find $RCSdirs \
+								-type f \
+								! -name '*_' \
+								! -name ',*,' \
+								! -name '.*_' \
+								! -name .rcsfreeze.log \
+								! -name .rcsfreeze.ver \
+								-print;;
+						esac
+						find . -name '*,v' -print
+					} |
+					sort -u |
+					sed "$filesFromRCSfiles"
+				`;;
+			*)
+				files=
+				for file in RCS/.* RCS/* .*,v *,v
+				do
+					case $file in
+					RCS/. | RCS/.. | RCS/,*, | RCS/*_) continue;;
+					RCS/.rcsfreeze.log | RCS/.rcsfreeze.ver) continue;;
+					RCS/.\* | RCS/\* | .\*,v | \*,v) test -f "$file" || continue;;
+					RCS/*,v | RCS/.*,v) ;;
+					RCS/* | RCS/.*) test -f "$file" || continue;;
+					esac
+					case $files in
+					'') files=$file;;
+					?*) files=$files$nl$file;;
+					esac
+				done
+				case $files in
+				'') exit 0;;
+				esac;;
+			esac
+			set x $files
+			shift
+			IFS=$oldIFS;;
+		esac;;
+	esac
+
+	case $datearg in
+	?*) $rlog $rlog_options "$datearg" ${1+"$@"} >$rlogfile;;
+	'') $rlog $rlog_options ${1+"$@"} >$rlogfile;;
+	esac || exit;;
+esac
 
 
 # Get the full name of each author the logs mention, and set initialize_fullname
@@ -329,17 +369,14 @@
 		sed 's/["\\]/\\&/g' >$llogout <<EOF || exit
 $loginFullnameMailaddrs
 EOF
-		loginFullnameMailaddrs=`cat $llogout`
+		loginFullnameMailaddrs=`cat $llogout`;;
 	esac
 
 	oldIFS=$IFS
 	IFS=$nl
 	for loginFullnameMailaddr in $loginFullnameMailaddrs
 	do
-		case $loginFullnameMailaddr in
-		*"$tab"*) IFS=$tab;;
-		*) IFS=:
-		esac
+		IFS=$tab
 		set x $loginFullnameMailaddr
 		login=$2
 		fullname=$3
@@ -349,25 +386,26 @@
 		initialize_mailaddr="$initialize_mailaddr
 			mailaddr[\"$login\"] = \"$mailaddr\""
 	done
-	IFS=$oldIFS
+	IFS=$oldIFS;;
 esac
 
-case $llogout in
-?*) sort -u -o $llogout <<EOF || exit
+case $logins in
+?*)
+	sort -u -o $llogout <<EOF
 $logins
 EOF
-esac
+	;;
+'')
+	: ;;
+esac >$llogout || exit
+
 output_authors='/^date: / {
 	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:" && $5 ~ /^[^;]*;$/) {
 		print substr($5, 1, length($5)-1)
 	}
 }'
 authors=`
-	$AWK "$output_authors" <$rlogout |
-	case $llogout in
-	'') sort -u;;
-	?*) sort -u | comm -23 - $llogout
-	esac
+	$AWK "$output_authors" <"$rlogfile" | sort -u | comm -23 - $llogout
 `
 case $authors in
 ?*)
@@ -439,13 +477,13 @@
 			)
 		} 2>/dev/null |
 		$AWK -F: "$awkscript"
-	`$initialize_fullname
+	`$initialize_fullname;;
 esac
 
 
 # Function to print a single log line.
 # We don't use awk functions, to stay compatible with old awk versions.
-# `Log' is the log message (with \n replaced by \001).
+# `Log' is the log message.
 # `files' contains the affected files.
 printlogline='{
 
@@ -455,19 +493,25 @@
 	#	* file (function): comment
 	if (Log ~ /^\([^)]*\): /) {
 		i = index(Log, ")")
-		files = files " " substr(Log, 1, i)
+		filefunc = substr(Log, 1, i)
+		while ((j = index(filefunc, "\n"))) {
+			files = files " " substr(filefunc, 1, j-1)
+			filefunc = substr(filefunc, j+1)
+		}
+		files = files " " filefunc
 		Log = substr(Log, i+3)
 	}
 
 	# If "label: comment" is too long, break the line after the ":".
 	sep = " "
-	if ('"$length"' <= '"$indent"' + 1 + length(files) + index(Log, SOH)) sep = "\n" indent_string
+	i = index(Log, "\n")
+	if ('"$length"' <= '"$indent"' + 1 + length(files) + i) sep = "\n" indent_string
 
 	# Print the label.
 	printf "%s*%s:", indent_string, files
 
-	# Print each line of the log, transliterating \001 to \n.
-	while ((i = index(Log, SOH)) != 0) {
+	# Print each line of the log.
+	while (i) {
 		logline = substr(Log, 1, i-1)
 		if (logline ~ /[^'"$tab"' ]/) {
 			printf "%s%s\n", sep, logline
@@ -476,6 +520,7 @@
 		}
 		sep = indent_string
 		Log = substr(Log, i+1)
+		i = index(Log, "\n")
 	}
 }'
 
@@ -496,24 +541,27 @@
 	*)
 		domainname=`(domainname) 2>/dev/null` &&
 		case $domainname in
-		*.*) hostname=$hostname.$domainname
-		esac
-	esac
+		*.*) hostname=$hostname.$domainname;;
+		esac;;
+	esac;;
 esac
 
 
 # Process the rlog output, generating ChangeLog style entries.
 
 # First, reformat the rlog output so that each line contains one log entry.
-# Transliterate \n to \001 so that multiline entries fit on a single line.
+# Transliterate \n to SOH so that multiline entries fit on a single line.
 # Discard irrelevant rlog output.
-$AWK <$rlogout '
-	BEGIN { repository = "'"$repository"'" }
-	/^RCS file:/ {
-		if (repository != "") {
-			filename = $3
-			if (substr(filename, 1, length(repository) + 1) == repository "/") {
-				filename = substr(filename, length(repository) + 2)
+$AWK '
+	BEGIN {
+		pository = "'"$pository"'"
+		SOH="'"$SOH"'"
+	}
+	/^RCS file: / {
+		if (pository != "") {
+			filename = substr($0, 11)
+			if (substr(filename, 1, length(pository)) == pository) {
+				filename = substr(filename, length(pository) + 1)
 			}
 			if (filename ~ /,v$/) {
 				filename = substr(filename, 1, length(filename) - 2)
@@ -526,7 +574,7 @@
 		}
 		rev = "?"
 	}
-	/^Working file:/ { if (repository == "") filename = $3 }
+	/^Working file: / { if (repository == "") filename = substr($0, 15) }
 	/'"$rlog_revision_pattern"'/, /^(-----------*|===========*)$/ {
 		line = $0
 		if (line ~ /'"$rlog_revision_pattern"'/) {
@@ -547,7 +595,7 @@
 			}
 			time = substr($3, 1, length($3) - 1)
 			author = substr($5, 1, length($5)-1)
-			printf "%s %s %s %s %s %c", filename, rev, date, time, author, 1
+			printf "%s%s%s%s%s%s%s%s%s%s", filename, SOH, rev, SOH, date, SOH, time, SOH, author, SOH
 			rev = "?"
 			next
 		}
@@ -556,29 +604,25 @@
 		if (line == "Initial revision" || line ~ /^file .+ was initially added on branch .+\.$/) {
 			line = "New file."
 		}
-		printf "%s%c", line, 1
+		printf "%s%s", line, SOH
 	}
-' |
+' <"$rlogfile" |
 
 # Now each line is of the form
-# FILENAME REVISION YYYY-MM-DD HH:MM:SS[+-TIMEZONE] AUTHOR \001LOG
-#	where \001 stands for a carriage return,
-#	and each line of the log is terminated by \001 instead of \n.
+# FILENAME@REVISION@YYYY-MM-DD@HH:MM:SS[+-TIMEZONE]@AUTHOR@LOG
+#	where @ stands for an SOH (octal code 001),
+#	and each line of LOG is terminated by SOH instead of \n.
 # Sort the log entries, first by date+time (in reverse order),
 # then by author, then by log entry, and finally by file name and revision
 # (just in case).
-sort +2 -4r +4 +0 |
+sort -t"$SOH" +2 -4r +4 +0 |
 
 # Finally, reformat the sorted log entries.
-$AWK '
+$AWK -F"$SOH" '
 	BEGIN {
 		logTZ = "'"$logTZ"'"
 		revision = "'"$revision"'"
 
-		# Some awk variants do not understand "\001", so we have to
-		# put the char directly in the file.
-		SOH="
" # <-- There is a single SOH (octal code 001) here.
-
 		# Initialize the fullname and mailaddr associative arrays.
 		'"$initialize_fullname"'
 		'"$initialize_mailaddr"'
@@ -594,7 +638,8 @@
 	}
 
 	{
-		newlog = substr($0, 1 + index($0, SOH))
+		newlog = ""
+		for (i = 6; i < NF; i++) newlog = newlog $i "\n"
 
 		# Ignore log entries prefixed by "#".
 		if (newlog ~ /^#/) { next }