APPROVE COMMIT 21.5 RECOMMEND 21.4 Because of this email I received from Jens I'm updating our rcs2log. I didn't see anything that was (X)Emacs specific, so I'm going to just drop it straight in. I'll revert if anyone yells. Thanks very much for letting us know about this, Jens. Steve T/Vin: This should go into 21.4 From: Jens Petersen <petersen(a)redhat.com&gt; Subject: rcs2log tmp race Keywords: emacs To: Vin Shelton <acs(a)xemacs.org&gt;, "Stephen J.Turnbull" <stephen(a)xemacs.org&gt;, Steve Youngs <youngs(a)xemacs.org&gt;, Yoshiki Hayashi <yoshiki(a)xemacs.org&gt; Cc: Ville Skyttä <ville.skytta(a)xemacs.org&gt;, Mark J Cox <mjc(a)redhat.com&gt; Date: 16 Jan 2003 22:04:02 +0900 X-Sent: 1 week, 4 days, 14 hours, 8 minutes, 23 seconds ago Hi Guys, The rcs2log that is distributed with XEmacs hasn't been updated for a long time, and has a potential tmp file race issue, which was fixed in Emacs 21.1 and cvs-1.11.2 in 2001. It would be good if the copy shipped with XEmacs could also be updated to fix this issue. It was published on bugtraq last year apparently. Below is a pointer to the Emacs changelog entry for reference, though I guess it would probably be better just to sync it with the master copy in Emacs. http://savannah.gnu.org/cgi-bin/viewcvs/emacs/emacs/lib-src/rcs2log?only_... Best regards, Jens -- Jens Petersen Linux OS Development, Red Hat ---------- --- NOTE: This patch has been committed. 21.5 patch: ChangeLog files diff command: cvs -q diff -U 0 Files affected: lib-src/ChangeLog Source files diff command: cvs -q diff -uN Files affected: lib-src/rcs2log Index: lib-src/ChangeLog =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/lib-src/ChangeLog,v retrieving revision 1.152 diff -u -U0 -r1.152 ChangeLog --- lib-src/ChangeLog 5 Jan 2003 05:13:32 -0000 1.152 +++ lib-src/ChangeLog 28 Jan 2003 03:13:48 -0000 ＠＠ -0,0 +1,4 ＠＠ +2003-01-28 Steve Youngs <youngs(a)xemacs.org&gt; + + * rcs2log: Synch to GNU version. + Index: lib-src/rcs2log =================================================================== RCS file: /pack/xemacscvs/XEmacs/xemacs/lib-src/rcs2log,v retrieving revision 1.2 diff -u -u -r1.2 rcs2log --- lib-src/rcs2log 9 Jul 1997 04:31:03 -0000 1.2 +++ lib-src/rcs2log 28 Jan 2003 03:13:12 -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). ＠＠ -26,11 +27,12 ＠＠ --help Output help. --version Output version number. -Report bugs to <bug-gnu-emacs(a)prep.ai.mit.edu&gt;.&#39; +Report bugs to <bug-gnu-emacs(a)gnu.org&gt;.&#39; -Id='$Id: rcs2log,v 1.2 1997/07/09 04:31:03 steve Exp $' +Id='$Id: rcs2log,v 1.50 2002/02/03 17:31:31 eggert Exp $' -# Copyright 1992, 1993, 1994, 1995, 1996, 1997 Free Software Foundation, Inc. +# Copyright (C) 1992, 1993, 1994, 1995, 1996, 1997, 1998, 2001, 2002 +# Free Software Foundation, Inc. # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by ＠＠ -47,13 +49,29 ＠＠ # Free Software Foundation, Inc., 59 Temple Place - Suite 330, # Boston, MA 02111-1307, USA. -Copyright='Copyright 1997 Free Software Foundation, Inc. +Copyright='Copyright (C) 2002 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 <eggert(a)twinsun.com&gt;&#39; +# 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=' ' ＠＠ -61,8 +79,8 ＠＠ # Parse options. # defaults -: ${AWK=awk} -: ${TMPDIR=/tmp} +AWK=${AWK-awk} +TMPDIR=${TMPDIR-/tmp} changelog=ChangeLog # change log file name datearg= # rlog date option hostname= # name of local host (if empty, will deduce it later) ＠＠ -74,6 +92,7 ＠＠ recursive= # t if we want recursive rlog revision= # t if we want revision numbers rlog_options= # options to pass to rlog +rlogfile= # log file to read from tabwidth=8 # width of horizontal tab while : ＠＠ -83,14 +102,16 ＠＠ -i) indent=${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' '$4': tabs, newlines not allowed" - exit 1 + exit 1;; esac - loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$2$tab$3$tab$4 + login=$2 + lfm=$2$tab$3$tab$4 shift; shift; shift;; -u) # If $2 is not tab-separated, use colon for separator. ＠＠ -101,24 +122,41 ＠＠ *"$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 '$2': not enough fields" - exit 1 + exit 1;; esac - loginFullnameMailaddrs=$loginFullnameMailaddrs$nl$2 - shift + shift;; esac - logins=$logins$nl$login - ;; - -r) rlog_options=$rlog_options$nl${2?}; shift;; + case $logins in + '') logins=$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${2?};; + esac + shift;; -R) recursive=t;; -t) tabwidth=${2?}; shift;; -v) revision=t;; ＠＠ -130,9 +168,9 ＠＠ -*) echo >&2 "Usage: $0 [OPTION]... [FILE ...]$nl$Help" case $1 in --help) exit 0;; - *) exit 1 + *) exit 1;; esac;; - *) break + *) break;; esac shift done ＠＠ -144,148 +182,182 ＠＠ m[9]="Oct"; m[10]="Nov"; m[11]="Dec" ' +logdir=$TMPDIR/rcs2log$$ +llogout=$logdir/l +trap exit 1 2 13 15 +trap "rm -fr $logdir 2>/dev/null" 0 +(umask 077 && exec mkdir $logdir) || exit -# 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 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 - e=' - /^[0-9]+-[0-9][0-9]-[0-9][0-9]/{ - # ISO 8601 date - print $1 - exit + extractTZ=' + /^.*change-log-time-zone-rule['"$tab"' ]*:['"$tab"' ]*"\([^"]*\)".*/{ + s//\1/; p; q } - /^... ... [ 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 = $3 - printf "%d-%02d-%02d\n", year, i+1, dd - exit + /^.*change-log-time-zone-rule['"$tab"' ]*:['"$tab"' ]*t.*/{ + s//UTC0/; p; q } ' - d=`$AWK "$e" <"$changelog"` || exit - case $d in - ?*) datearg="-d>$d" + logTZ=`tail "$changelog" | sed -n "$extractTZ"` + case $logTZ in + ?*) TZ=$logTZ; export TZ;; 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 -# 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 - ;; + # Use $rlog's -zLT option, if $rlog supports it. + case `$rlog -zLT 2>&1` in + *' option'*) ;; *) - # 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 + case $rlog_options in + '') rlog_options=-zLT;; + ?*) rlog_options=-zLT$nl$rlog_options;; + esac;; esac -fi - -# Use $rlog's -zLT option, if $rlog supports it. -case `$rlog -zLT 2>&1` in -*' option'*) ;; -*) rlog_options=-zLT$nl$rlog_options -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 + # 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 "$file" || continue;; esac - find . -name '*,v' -print - } | - sort -u | - sed "$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 - esac - files=$files$nl$file - done - case $files in - '') exit 0 + case $files in + '') files=$file;; + ?*) files=$files$nl$file;; + esac + done + case $files in + '') exit 0;; + esac;; esac - esac - set x $files - shift - IFS=$oldIFS + set x $files + shift + IFS=$oldIFS;; + esac;; esac + + case $datearg in + ?*) $rlog $rlog_options "$datearg" ${1+"$＠"} >$rlogfile;; + '') $rlog $rlog_options ${1+"$＠"} >$rlogfile;; + esac || exit;; esac -llogout=$TMPDIR/rcs2log$$l -rlogout=$TMPDIR/rcs2log$$r -trap exit 1 2 13 15 -trap "rm -f $llogout $rlogout; exit 1" 0 -case $datearg in -?*) $rlog $rlog_options "$datearg" ${1+"$＠"} >$rlogout;; -'') $rlog $rlog_options ${1+"$＠"} >$rlogout -esac || exit +# Prefer the POSIX-style -k options, since POSIX 1003.1-2001 prohibits +# support for the traditional-style +M -N options. +SORT_K_OPTIONS='-k 3,4r -k 5 -k 1' +sort $SORT_K_OPTIONS </dev/null 2>/dev/null || SORT_K_OPTIONS='+2 -4r +4 +0' # Get the full name of each author the logs mention, and set initialize_fullname ＠＠ -303,17 +375,14 ＠＠ sed 's/["\\]/\\&/g' >$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=$3 ＠＠ -323,25 +392,26 ＠＠ initialize_mailaddr="$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 ?*) ＠＠ -402,21 +472,24 ＠＠ ' initialize_fullname=` - ( - cat /etc/passwd - for author in $authors - do nismatch $author passwd.org_dir - done - ypmatch $authors passwd - ) 2>/dev/null | + { + (getent passwd $authors) || + ( + cat /etc/passwd + for author in $authors + do NIS_PATH= nismatch $author passwd.org_dir + done + ypmatch $authors passwd + ) + } 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 \r). +# `Log' is the log message. # `files' contains the affected files. printlogline='{ ＠＠ -424,21 +497,27 ＠＠ # * file: (function): comment # to # * file (function): comment - if (Log ~ /^\([^)]*\): /) { + if (Log ~ /^\([^)]*\):[\t\n ]/) { 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, CR)) 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 \r to \n. - while ((i = index(Log, CR)) != 0) { + # Print each line of the log. + while (i) { logline = substr(Log, 1, i-1) if (logline ~ /[^'"$tab"' ]/) { printf "%s%s\n", sep, logline ＠＠ -447,6 +526,7 ＠＠ } sep = indent_string Log = substr(Log, i+1) + i = index(Log, "\n") } }' ＠＠ -467,24 +547,27 ＠＠ *) domainname=`(domainname) 2>/dev/null` && case $domainname in - *.*) hostname=$hostname.$domainname - esac - esac + *.*) hostname=$hostname.$domainname;; + esac;; + esac;; esac # Process the rlog output, generating ChangeLog style entries. # First, reformat the rlog output so that each line contains one log entry. -# Transliterate \n to \r so that multiline entries fit on a single line. +# Transliterate \n to SOH so that multiline entries fit on a single line. # Discard irrelevant rlog output. -$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) ＠＠ -497,13 +580,14 ＠＠ } rev = "?" } - /^Working file:/ { if (repository == "") filename = $3 } + /^Working file: / { if (repository == "") filename = substr($0, 15) } /'"$rlog_revision_pattern"'/, /^(-----------*|===========*)$/ { - if ($0 ~ /'"$rlog_revision_pattern"'/) { + line = $0 + if (line ~ /'"$rlog_revision_pattern"'/) { rev = $2 next } - if ($0 ~ /^date: [0-9][- +\/0-9:]*;/) { + if (line ~ /^date: [0-9][- +\/0-9:]*;/) { date = $2 if (date ~ /\//) { # This is a traditional RCS format date YYYY/MM/DD. ＠＠ -517,38 +601,34 ＠＠ } 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, 13 + printf "%s%s%s%s%s%s%s%s%s%s", filename, SOH, rev, SOH, date, SOH, time, SOH, author, SOH rev = "?" next } - if ($0 ~ /^branches: /) { next } - if ($0 ~ /^(-----------*|===========*)$/) { print ""; next } - if ($0 == "Initial revision" || $0 ~ /^file .+ was initially added on branch .+\.$/) { - $0 = "New file." - } - printf "%s%c", $0, 13 + if (line ~ /^branches: /) { next } + if (line ~ /^(-----------*|===========*)$/) { print ""; next } + if (line == "Initial revision" || line ~ /^file .+ was initially added on branch .+\.$/) { + line = "New file." + } + printf "%s%s", line, SOH } -' | +' <"$rlogfile" | # Now each line is of the form -# FILENAME REVISION YYYY-MM-DD HH:MM:SS[+-TIMEZONE] AUTHOR \rLOG -# where \r stands for a carriage return, -# and each line of the log is terminated by \r 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" $SORT_K_OPTIONS | # Finally, reformat the sorted log entries. -$AWK ' +$AWK -F"$SOH" ' BEGIN { logTZ = "'"$logTZ"'" revision = "'"$revision"'" - # Some awk variants do not understand "\r" or "\013", so we have to - # put a carriage return directly in the file. - CR=" " # <-- There is a single CR between the " chars here. - # Initialize the fullname and mailaddr associative arrays. '"$initialize_fullname"' '"$initialize_mailaddr"' ＠＠ -564,7 +644,8 ＠＠ } { - newlog = substr($0, 1 + index($0, CR)) + newlog = "" + for (i = 6; i < NF; i++) newlog = newlog $i "\n" # Ignore log entries prefixed by "#". if (newlog ~ /^#/) { next } ＠＠ -588,7 +669,7 ＠＠ newclumpname = substr(newlog, 1, i) while (substr(newlog, i+1) ~ /^['"$tab"' ]/) i++ newlog = substr(newlog, i+1) - if (clumpname == newclumpname) sep = "" + if (clumpname == newclumpname && date == $3 && author == $5) sep = "" } printf sep clumpname = newclumpname ＠＠ -643,7 +724,7 ＠＠ # Exit successfully. -exec rm -f $llogout $rlogout +exec rm -fr $logdir # Local Variables: # tab-width:4 -- |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---| | XEmacs - It's not just an editor. | | It's a way of life. | |------------------------------------&lt;youngs(a)xemacs.org&gt;---|