[bpt/emacs.git] / lib-src / rcs2log

#!/bin/sh

# RCS to ChangeLog generator

#	$Id: rcs2clog,v 1.2 1992/02/05 04:31:18 eggert Exp eggert $

# Generate a change log prefix from RCS/* and the existing ChangeLog (if any).
# Output the new prefix to standard output.
# You can edit this prefix by hand, and then prepend it to ChangeLog.


# Parse options.

# defaults
indent=8 # indent of log line
length=79 # suggested max width of log line
tabwidth=8 # width of horizontal tab

while :
do
	case $1 in
	-i)	indent=${2?};;
	-l)	length=${2?};;
	-t)	tabwidth=${2?};;
	-*)	echo >&2 "$0: usage: $0 [-i indent] [-l length] [-t tabwidth] [file ...]"
		exit 1;;
	*)	break
	esac
	shift; shift
done


# Log into $rlogout the revisions checked in since the first ChangeLog entry.

datearg=-d'>1970'
if test -s ChangeLog
then
	e='s/^\(...\) \(...\) \(..\) \(..:..:..\) \(....\) .*/\1 \2 \3 \4 \5/p; 1q'
	date=`sed -n "$e" <ChangeLog` || exit
	case $date in
	?*) datearg="-d>$date"
	esac
fi

rlogout=/tmp/chg$$
trap exit 1 2 13 15
trap 'rm -f $rlogout; exit 1' 0

case $# in
0) set RCS/*
esac

rlog "$datearg" "$@" >$rlogout || exit


# Get the full name of each author the logs mention, and set initialize_fullname
# to awk code that initializes the `fullname' awk associative array.
# Warning: foreign authors (i.e. not known in the passwd file) are mishandled;
# you have to fix the resulting output by hand.

authors=`
	sed -n 's|^date: *[0-9]*/[0-9][0-9]/[0-9][0-9] [0-9][0-9]:[0-9][0-9]:[0-9][0-9]; *author: *\([^; ]*\).*|\1|p' <$rlogout |
	sort -u
`

initialize_fullname=
for author in $authors
do
	fullname=`
		(grep "^$author:" /etc/passwd || ypmatch "$author" passwd) |
		sed -n 's/^[^:]*:[^:]*:[^:]*:[^:]*:\([^,:]*\).*$/\1/;p;q'
	`
	case $fullname in
	*\&*)
		User=`
			expr " $author" : ' \(.\)' |
			tr abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ
		``
			expr " $author" : ' .\(.*\)'
		`
		fullname=`echo "$fullname" | sed "s:&:$User:"`
	esac
	case $fullname in
	?*)
		initialize_fullname="$initialize_fullname
			fullname[\"$author\"] = \"$fullname\""
	esac
done


# 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 \r).
# `files' contains the affected files (each preceded by a space).
printlogline='{

	# Following the GNU coding standards, rewrite
	#	* file: (function): comment
	# to
	#	* file (function): comment
	if (Log ~ /^\([^)]*\): /) {
		i = index(Log, ")")
		files = files " " substr(Log, 1, i)
		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, "\r")) sep = "\n" indent_string

	# Print the label.
	printf "%s*%s:", indent_string, files

	# Print each line of the log, transliterating \r to \n.
	while ((i = index(Log, "\r")) != 0) {
		printf "%s%s\n", sep, substr(Log, 1, i-1)
		sep = indent_string
		Log = substr(Log, i+1)
	}

	printf "\n"
}'

hostname=`(
	hostname || cat /etc/whoami || uuname -l || uname -n
) 2>/dev/null` || {
	echo >&2 "$0: cannot deduce hostname"
	exit 1
}


# Process the rlog output, generating ChangeLog style entries.

# First, reformat the rlog output so that each line contains one log entry.
# Transliterate \n to \r so that multiline entries fit on a single line.
# Discard irrelevant rlog output.
awk <$rlogout '
	/^Working file:/ { filename = $3 }
	/^date: /, /^(-----------*|===========*)$/ {
		if ($0 ~ /^branches: /) { next }
		if ($0 ~ /^date: [0-9][ /0-9:]*;/) {
			time = substr($3, 1, length($3)-1)
			author = substr($5, 1, length($5)-1)
			printf "%s %s %s %s \r", filename, $2, time, author
			next
		}
		if ($0 ~ /^(-----------*|===========*)/) { print ""; next }
		{ printf "%s\r", $0 }
	}
' |

# Now each line is of the form
# FILENAME YYYY/MM/DD HH:MM:SS AUTHOR \rLOG
#	where \r stands for a carriage return,
#	and each line of the log is terminated by \r instead of \n.
# Sort the log entries, first by date (in reverse order),
# then by author, then by log entry, and finally by file name (just in case).
sort +1 -2r +3 +0 |

# Finally, reformat the sorted log entries.
awk '
	BEGIN {

		# Initialize the fullname associative array.
		'"$initialize_fullname"'

		# Initialize indent string.
		indent_string = ""
		i = '"$indent"'
		if (0 < '"$tabwidth"')
			for (;  '"$tabwidth"' <= i;  i -= '"$tabwidth"')
				indent_string = indent_string "\t"
		while (1 <= i--)
			indent_string = indent_string " "

		# Set up date conversion tables.
		# RCS uses a nice, clean, sortable format,
		# but ChangeLog wants the traditional, ugly ctime format.

		# January 1, 0 AD (Gregorian) was Saturday = 6
		EPOCH_WEEKDAY = 6
		# Of course, there was no 0 AD, but the algorithm works anyway.

		w[0]="Sun"; w[1]="Mon"; w[2]="Tue"; w[3]="Wed"
		w[4]="Thu"; w[5]="Fri"; w[6]="Sat"

		m[0]="Jan"; m[1]="Feb"; m[2]="Mar"
		m[3]="Apr"; m[4]="May"; m[5]="Jun"
		m[6]="Jul"; m[7]="Aug"; m[8]="Sep"
		m[9]="Oct"; m[10]="Nov"; m[11]="Dec"

		# days in non-leap year thus far, indexed by month (0-12)
		mo[0]=0; mo[1]=31; mo[2]=59; mo[3]=90
		mo[4]=120; mo[5]=151; mo[6]=181; mo[7]=212
		mo[8]=243; mo[9]=273; mo[10]=304; mo[11]=334
		mo[12]=365
	}
	{
		newlog = substr($0, 1 + index($0, "\r"))
		if (Log != newlog || date != $2 || author != $4) {
			# The previous log and this log differ.
			# Print the old one.
			if (date != "") '"$printlogline"'

			# Get ready for the next log.
			Log = newlog
			if (files != "")
				for (i in filesknown)
					filesknown[i] = 0
			files = ""
		}
		if (date != $2  ||  author != $4) {
			# The previous date+author and this date+author differ.
			# Print the new one.
			date = $2
			author = $4

			# Convert nice RCS date like "1992/01/03 00:03:44"
			# into ugly ctime date like "Fri Jan  3 00:03:44 1992".
			# Calculate day of week from Gregorian calendar.
			i = index($2, "/")
			year = substr($2, 1, i-1)
			monthday = substr($2, i+1)
			i = index(monthday, "/")
			month = substr(monthday, 1, i-1)
			day = substr(monthday, i+1)
			leap = 0
			if (2 <= month && year%4 == 0 && (year%100 != 0 || year%400 == 0)) leap = 1
			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

			# Print "date  fullname  (email address)" if the fullname is known;
			# print "date  author" otherwise.
			# Get the fullname from the associative array.
			# The email address is just author@thishostname.
			printf "%s %s %2d %s %d  ", w[days_since_Sunday_before_epoch%7], m[month-1], day, $3, year
			if (fullname[author])
				printf "%s  (%s@%s)\n\n", fullname[author], author, "'"$hostname"'"
			else
				printf "%s\n\n", author
		}
		if (! filesknown[$1]) {
			filesknown[$1] = 1
			files = files " " $1
		}
	}
	END {
		# Print the last log.
		if (date != "") '"$printlogline"'
	}
' &&


# Exit successfully.

exec rm -f $rlogout
Commit	Line	Data
b86e8615 PE	1	#!/bin/sh
	2
	3	# RCS to ChangeLog generator
	4
ff17867b	5	# $Id: rcs2clog,v 1.2 1992/02/05 04:31:18 eggert Exp eggert $
b86e8615 PE	6
	7	# Generate a change log prefix from RCS/* and the existing ChangeLog (if any).
	8	# Output the new prefix to standard output.
	9	# You can edit this prefix by hand, and then prepend it to ChangeLog.
	10
	11
4ca2c4b4 PE	12	# Parse options.
	13
	14	# defaults
	15	indent=8 # indent of log line
	16	length=79 # suggested max width of log line
	17	tabwidth=8 # width of horizontal tab
	18
	19	while :
	20	do
	21	case $1 in
	22	-i) indent=${2?};;
	23	-l) length=${2?};;
	24	-t) tabwidth=${2?};;
	25	-*) echo >&2 "$0: usage: $0 [-i indent] [-l length] [-t tabwidth] [file ...]"
	26	exit 1;;
	27	*) break
	28	esac
	29	shift; shift
	30	done
	31
	32
b86e8615 PE	33	# Log into $rlogout the revisions checked in since the first ChangeLog entry.
	34
	35	datearg=-d'>1970'
	36	if test -s ChangeLog
	37	then
ff17867b PE	38	e='s/^\(...\) \(...\) \(..\) \(..:..:..\) \(....\) .*/\1 \2 \3 \4 \5/p; 1q'
	39	date=`sed -n "$e" <ChangeLog` \|\| exit
	40	case $date in
	41	?*) datearg="-d>$date"
	42	esac
b86e8615 PE	43	fi
	44
	45	rlogout=/tmp/chg$$
	46	trap exit 1 2 13 15
	47	trap 'rm -f $rlogout; exit 1' 0
	48
4ca2c4b4 PE	49	case $# in
	50	0) set RCS/*
	51	esac
	52
	53	rlog "$datearg" "$@" >$rlogout \|\| exit
b86e8615 PE	54
	55
	56	# Get the full name of each author the logs mention, and set initialize_fullname
	57	# to awk code that initializes the `fullname' awk associative array.
	58	# Warning: foreign authors (i.e. not known in the passwd file) are mishandled;
	59	# you have to fix the resulting output by hand.
	60
	61	authors=`
	62	sed -n 's\|^date: [0-9]/[0-9][0-9]/[0-9][0-9] [0-9][0-9]:[0-9][0-9]:[0-9][0-9]; author: \([^; ]\).\|\1\|p' <$rlogout \|
	63	sort -u
	64	`
	65
	66	initialize_fullname=
	67	for author in $authors
	68	do
	69	fullname=`
	70	(grep "^$author:" /etc/passwd \|\| ypmatch "$author" passwd) \|
	71	sed -n 's/^[^:]:[^:]:[^:]:[^:]:\([^,:]\).$/\1/;p;q'
	72	`
	73	case $fullname in
	74	\&)
	75	User=`
	76	expr " $author" : ' \(.\)' \|
	77	tr abcdefghijklmnopqrstuvwxyz ABCDEFGHIJKLMNOPQRSTUVWXYZ
	78	``
	79	expr " $author" : ' .\(.*\)'
	80	`
	81	fullname=`echo "$fullname" \| sed "s:&:$User:"`
	82	esac
4ca2c4b4 PE	83	case $fullname in
	84	?*)
	85	initialize_fullname="$initialize_fullname
	86	fullname[\"$author\"] = \"$fullname\""
	87	esac
b86e8615 PE	88	done
	89
	90
	91	# Function to print a single log line.
	92	# We don't use awk functions, to stay compatible with old awk versions.
	93	# `Log' is the log message (with \n replaced by \r).
	94	# `files' contains the affected files (each preceded by a space).
b86e8615 PE	95	printlogline='{
	96
	97	# Following the GNU coding standards, rewrite
	98	# * file: (function): comment
	99	# to
	100	# * file (function): comment
	101	if (Log ~ /^\([^)]*\): /) {
	102	i = index(Log, ")")
	103	files = files " " substr(Log, 1, i)
	104	Log = substr(Log, i+3)
	105	}
	106
	107	# If "label: comment" is too long, break the line after the ":".
	108	sep = " "
4ca2c4b4	109	if ('"$length"' <= '"$indent"' + 1 + length(files) + index(Log, "\r")) sep = "\n" indent_string
b86e8615 PE	110
b86e8615 PE	111	# Print the label.
4ca2c4b4	112	printf "%s*%s:", indent_string, files
b86e8615 PE	113
	114	# Print each line of the log, transliterating \r to \n.
	115	while ((i = index(Log, "\r")) != 0) {
	116	printf "%s%s\n", sep, substr(Log, 1, i-1)
4ca2c4b4	117	sep = indent_string
b86e8615 PE	118	Log = substr(Log, i+1)
	119	}
	120
	121	printf "\n"
	122	}'
	123
	124	hostname=`(
	125	hostname \|\| cat /etc/whoami \|\| uuname -l \|\| uname -n
	126	) 2>/dev/null` \|\| {
	127	echo >&2 "$0: cannot deduce hostname"
	128	exit 1
	129	}
	130
	131
	132	# Process the rlog output, generating ChangeLog style entries.
	133
	134	# First, reformat the rlog output so that each line contains one log entry.
	135	# Transliterate \n to \r so that multiline entries fit on a single line.
	136	# Discard irrelevant rlog output.
	137	awk <$rlogout '
	138	/^Working file:/ { filename = $3 }
	139	/^date: /, /^(-----------\|===========)$/ {
	140	if ($0 ~ /^branches: /) { next }
	141	if ($0 ~ /^date: [0-9][ /0-9:]*;/) {
	142	time = substr($3, 1, length($3)-1)
	143	author = substr($5, 1, length($5)-1)
	144	printf "%s %s %s %s \r", filename, $2, time, author
	145	next
	146	}
	147	if ($0 ~ /^(-----------\|===========)/) { print ""; next }
	148	{ printf "%s\r", $0 }
	149	}
	150	' \|
	151
	152	# Now each line is of the form
	153	# FILENAME YYYY/MM/DD HH:MM:SS AUTHOR \rLOG
	154	# where \r stands for a carriage return,
	155	# and each line of the log is terminated by \r instead of \n.
	156	# Sort the log entries, first by date (in reverse order),
	157	# then by author, then by log entry, and finally by file name (just in case).
	158	sort +1 -2r +3 +0 \|
	159
	160	# Finally, reformat the sorted log entries.
	161	awk '
	162	BEGIN {
	163
	164	# Initialize the fullname associative array.
	165	'"$initialize_fullname"'
	166
4ca2c4b4 PE	167	# Initialize indent string.
	168	indent_string = ""
	169	i = '"$indent"'
	170	if (0 < '"$tabwidth"')
	171	for (; '"$tabwidth"' <= i; i -= '"$tabwidth"')
	172	indent_string = indent_string "\t"
	173	while (1 <= i--)
	174	indent_string = indent_string " "
	175
b86e8615 PE	176	# Set up date conversion tables.
	177	# RCS uses a nice, clean, sortable format,
	178	# but ChangeLog wants the traditional, ugly ctime format.
	179
	180	# January 1, 0 AD (Gregorian) was Saturday = 6
	181	EPOCH_WEEKDAY = 6
	182	# Of course, there was no 0 AD, but the algorithm works anyway.
	183
	184	w[0]="Sun"; w[1]="Mon"; w[2]="Tue"; w[3]="Wed"
	185	w[4]="Thu"; w[5]="Fri"; w[6]="Sat"
	186
	187	m[0]="Jan"; m[1]="Feb"; m[2]="Mar"
	188	m[3]="Apr"; m[4]="May"; m[5]="Jun"
	189	m[6]="Jul"; m[7]="Aug"; m[8]="Sep"
	190	m[9]="Oct"; m[10]="Nov"; m[11]="Dec"
	191
	192	# days in non-leap year thus far, indexed by month (0-12)
	193	mo[0]=0; mo[1]=31; mo[2]=59; mo[3]=90
	194	mo[4]=120; mo[5]=151; mo[6]=181; mo[7]=212
	195	mo[8]=243; mo[9]=273; mo[10]=304; mo[11]=334
	196	mo[12]=365
	197	}
	198	{
	199	newlog = substr($0, 1 + index($0, "\r"))
	200	if (Log != newlog \|\| date != $2 \|\| author != $4) {
	201	# The previous log and this log differ.
	202	# Print the old one.
	203	if (date != "") '"$printlogline"'
	204
	205	# Get ready for the next log.
	206	Log = newlog
4ca2c4b4 PE	207	if (files != "")
	208	for (i in filesknown)
	209	filesknown[i] = 0
b86e8615 PE	210	files = ""
	211	}
	212	if (date != $2 \|\| author != $4) {
	213	# The previous date+author and this date+author differ.
	214	# Print the new one.
	215	date = $2
	216	author = $4
	217
	218	# Convert nice RCS date like "1992/01/03 00:03:44"
	219	# into ugly ctime date like "Fri Jan 3 00:03:44 1992".
	220	# Calculate day of week from Gregorian calendar.
	221	i = index($2, "/")
	222	year = substr($2, 1, i-1)
	223	monthday = substr($2, i+1)
	224	i = index(monthday, "/")
	225	month = substr(monthday, 1, i-1)
	226	day = substr(monthday, i+1)
	227	leap = 0
	228	if (2 <= month && year%4 == 0 && (year%100 != 0 \|\| year%400 == 0)) leap = 1
	229	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
	230
4ca2c4b4 PE	231	# Print "date fullname (email address)" if the fullname is known;
4ca2c4b4 PE	232	# print "date author" otherwise.
b86e8615 PE	233	# Get the fullname from the associative array.
b86e8615 PE	234	# The email address is just author@thishostname.
4ca2c4b4 PE	235	printf "%s %s %2d %s %d ", w[days_since_Sunday_before_epoch%7], m[month-1], day, $3, year
	236	if (fullname[author])
	237	printf "%s (%s@%s)\n\n", fullname[author], author, "'"$hostname"'"
	238	else
	239	printf "%s\n\n", author
	240	}
	241	if (! filesknown[$1]) {
	242	filesknown[$1] = 1
	243	files = files " " $1
b86e8615	244	}
b86e8615 PE	245	}
	246	END {
	247	# Print the last log.
	248	if (date != "") '"$printlogline"'
	249	}
	250	' &&
	251
	252
	253	# Exit successfully.
	254
	255	exec rm -f $rlogout