1
0
Fork 0
mirror of https://git.tukaani.org/xz.git synced 2024-04-04 12:36:23 +02:00
xz-archive/src/scripts/xzgrep.in

300 lines
10 KiB
Text
Raw Normal View History

#!@POSIX_SHELL@
# SPDX-License-Identifier: GPL-2.0-or-later
# xzgrep -- a wrapper around a grep program that decompresses files as needed
# Adapted from a version sent by Charles Levert <charles@comm.polymtl.ca>
# Copyright (C) 1998, 2001, 2002, 2006, 2007 Free Software Foundation
# Copyright (C) 1993 Jean-loup Gailly
# Modified for XZ Utils by Andrew Dudman and Lasse Collin.
# 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
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
@enable_path_for_scripts@
#SET_PATH - This line is a placeholder to ease patching this script.
# Instead of unsetting XZ_OPT, just make sure that xz will use file format
# autodetection. This way memory usage limit and thread limit can be
# specified via XZ_OPT. With gzip, bzip2, and lzop it's OK to just unset the
# environment variables.
xz='@xz@ --format=auto'
unset GZIP BZIP BZIP2 LZOP
case ${0##*/} in
*egrep*) prog=xzegrep; grep=${GREP:-grep -E};;
*fgrep*) prog=xzfgrep; grep=${GREP:-grep -F};;
*) prog=xzgrep; grep=${GREP:-grep};;
esac
version="$prog (@PACKAGE_NAME@) @VERSION@"
usage="Usage: ${0##*/} [OPTION]... [-e] PATTERN [FILE]...
Look for instances of PATTERN in the input FILEs, using their
uncompressed contents if they are compressed.
OPTIONs are the same as for '$grep'.
Report bugs to <@PACKAGE_BUGREPORT@>."
# sed script to escape all ' for the shell, and then (to handle trailing
# newlines correctly) turn trailing X on last line into '.
escape='
s/'\''/'\''\\'\'''\''/g
$s/X$/'\''/
'
operands=
have_pat=0
files_with_matches=0
files_without_matches=0
no_filename=0
with_filename=0
# See if -H and --label options are supported (GNU and *BSDs).
if test f:x = "$(eval "echo x | $grep -H --label=f x 2> /dev/null")"; then
grep_supports_label=1
else
grep_supports_label=0
fi
while test $# -ne 0; do
option=$1
shift
optarg=
case $option in
(-[0123456789abcdEFGhHiIKlLnoPqrRsTuUvVwxyzZ]*[!0123456789]*)
# Something like -Fiv was specified, that is, $option contains more
# than one option of which the first option (in this example -F)
# doesn't take an argument. Split the first option into a standalone
# argument and continue parsing the rest of the options (in this example,
# replace -Fiv with -iv in the argument list and set option=-F).
#
# If there are digits [0-9] they are treated as if they were a single
# option character because this syntax is an alias for -C for GNU grep.
# For example, "grep -25F" is equivalent to "grep -C25 -F". If only
# digits are specified like "grep -25" we don't get here because the
# above pattern in the case-statement doesn't match such strings.
arg2=-\'$(LC_ALL=C expr "X${option}X" : 'X-.[0-9]*\(.*\)' |
LC_ALL=C sed "$escape")
eval "set -- $arg2 "'${1+"$@"}'
option=$(LC_ALL=C expr "X$option" : 'X\(-.[0-9]*\)');;
(--binary-*=* | --[lm]a*=* | --reg*=*)
# These options require an argument and an argument has been provided
# with the --foo=argument syntax. All is good.
;;
(-[ABCDefmX] | --binary-* | --file | --[lm]a* | --reg*)
# These options require an argument which should now be in $1.
# If it isn't, display an error and exit.
case ${1?"$option option requires an argument"} in
(*\'*)
optarg=" '"$(printf '%sX\n' "$1" | LC_ALL=C sed "$escape");;
(*)
optarg=" '$1'";;
esac
shift;;
(--)
break;;
(-?*)
;;
(*)
case $option in
(*\'*)
operands="$operands '"$(printf '%sX\n' "$option" |
LC_ALL=C sed "$escape");;
(*)
operands="$operands '$option'";;
esac
${POSIXLY_CORRECT+break}
continue;;
esac
case $option in
(-[drRzZ] | --di* | --exc* | --inc* | --rec* | --nu*)
printf >&2 '%s: %s: Option not supported\n' "$0" "$option"
exit 2;;
(-[ef]* | --file | --file=* | --reg*)
have_pat=1;;
(--h | --he | --hel | --help)
printf '%s\n' "$usage" || exit 2
exit;;
(-H | --wi | --wit | --with | --with- | --with-f | --with-fi \
| --with-fil | --with-file | --with-filen | --with-filena | --with-filenam \
| --with-filename)
with_filename=1
continue;;
(-l | --files-with-*)
files_with_matches=1
continue;;
(-L | --files-witho*)
files_without_matches=1
continue;;
(-h | --no-f*)
no_filename=1;;
(-V | --v | --ve | --ver | --vers | --versi | --versio | --version)
printf '%s\n' "$version" || exit 2
exit;;
esac
case $option in
(*\'?*)
option=\'$(printf '%sX\n' "$option" | LC_ALL=C sed "$escape");;
(*)
option="'$option'";;
esac
grep="$grep $option$optarg"
done
eval "set -- $operands "'${1+"$@"}'
if test $have_pat -eq 0; then
case ${1?"Missing pattern; try '${0##*/} --help' for help"} in
(*\'*)
grep="$grep -e '"$(printf '%sX\n' "$1" | LC_ALL=C sed "$escape");;
(*)
grep="$grep -e '$1'";;
esac
shift
fi
if test $# -eq 0; then
set -- -
fi
exec 3>&1
# res=1 means that no file matched yet
res=1
for i; do
case $i in
*[-.][zZ] | *_z | *[-.]gz | *.t[ag]z) uncompress="gzip -cdf";;
*[-.]bz2 | *[-.]tbz | *.tbz2) uncompress="bzip2 -cdf";;
*[-.]lzo | *[-.]tzo) uncompress="lzop -cdf";;
*[-.]zst | *[-.]tzst) uncompress="zstd -cdfq";; # zstd needs -q.
*) uncompress="$xz -cdfqQ";; # -qQ to ignore warnings like unsupp. check.
esac
# xz_status will hold the decompressor's exit status.
# Exit status of grep (and in rare cases, printf or sed) is
# available as the exit status of this assignment command.
xz_status=$(
exec 5>&1
($uncompress -- "$i" 5>&-; echo $? >&5) 3>&- |
if test $files_with_matches -eq 1; then
eval "$grep -q" && { printf '%s\n' "$i" || exit 2; }
elif test $files_without_matches -eq 1; then
eval "$grep -q" || {
r=$?
if test $r -eq 1; then
printf '%s\n' "$i" || r=2
fi
exit $r
}
elif test $with_filename -eq 0 &&
{ test $# -eq 1 || test $no_filename -eq 1; }; then
eval "$grep"
elif test $grep_supports_label -eq 1; then
# The grep implementation in use allows us to specify the filename
# that grep will prefix to the output lines. This is faster and
# less prone to security bugs than the fallback method that uses sed.
# This also avoids confusing output with GNU grep >= 3.5 (2020-09-27)
# which prints "binary file matches" to stderr instead of stdout.
#
# If reading from stdin, let grep use whatever name it prefers for
# stdin. With GNU grep it is a locale-specific translated string.
if test "x$i" = "x-"; then
eval "$grep -H"
else
eval "$grep -H --label \"\$i\""
fi
else
xzgrep: Fix escaping of malicious filenames (ZDI-CAN-16587). Malicious filenames can make xzgrep to write to arbitrary files or (with a GNU sed extension) lead to arbitrary code execution. xzgrep from XZ Utils versions up to and including 5.2.5 are affected. 5.3.1alpha and 5.3.2alpha are affected as well. This patch works for all of them. This bug was inherited from gzip's zgrep. gzip 1.12 includes a fix for zgrep. The issue with the old sed script is that with multiple newlines, the N-command will read the second line of input, then the s-commands will be skipped because it's not the end of the file yet, then a new sed cycle starts and the pattern space is printed and emptied. So only the last line or two get escaped. One way to fix this would be to read all lines into the pattern space first. However, the included fix is even simpler: All lines except the last line get a backslash appended at the end. To ensure that shell command substitution doesn't eat a possible trailing newline, a colon is appended to the filename before escaping. The colon is later used to separate the filename from the grep output so it is fine to add it here instead of a few lines later. The old code also wasn't POSIX compliant as it used \n in the replacement section of the s-command. Using \<newline> is the POSIX compatible method. LC_ALL=C was added to the two critical sed commands. POSIX sed manual recommends it when using sed to manipulate pathnames because in other locales invalid multibyte sequences might cause issues with some sed implementations. In case of GNU sed, these particular sed scripts wouldn't have such problems but some other scripts could have, see: info '(sed)Locale Considerations' This vulnerability was discovered by: cleemy desu wayo working with Trend Micro Zero Day Initiative Thanks to Jim Meyering and Paul Eggert discussing the different ways to fix this and for coordinating the patch release schedule with gzip.
2022-03-29 18:19:12 +02:00
# Append a colon so that the last character will never be a newline
# which would otherwise get lost in shell command substitution.
i="$i:"
# Escape & \ | and newlines only if such characters are present
# (speed optimization).
case $i in
(*'
'* | *'&'* | *'\'* | *'|'*)
# If sed fails, set i to a known safe string to ensure that
# failing sed did not create a half-escaped dangerous string.
i=$(printf '%s\n' "$i" | LC_ALL=C sed 's/[&\|]/\\&/g; $!s/$/\\/') ||
i='(unknown filename):';;
esac
xzgrep: Fix escaping of malicious filenames (ZDI-CAN-16587). Malicious filenames can make xzgrep to write to arbitrary files or (with a GNU sed extension) lead to arbitrary code execution. xzgrep from XZ Utils versions up to and including 5.2.5 are affected. 5.3.1alpha and 5.3.2alpha are affected as well. This patch works for all of them. This bug was inherited from gzip's zgrep. gzip 1.12 includes a fix for zgrep. The issue with the old sed script is that with multiple newlines, the N-command will read the second line of input, then the s-commands will be skipped because it's not the end of the file yet, then a new sed cycle starts and the pattern space is printed and emptied. So only the last line or two get escaped. One way to fix this would be to read all lines into the pattern space first. However, the included fix is even simpler: All lines except the last line get a backslash appended at the end. To ensure that shell command substitution doesn't eat a possible trailing newline, a colon is appended to the filename before escaping. The colon is later used to separate the filename from the grep output so it is fine to add it here instead of a few lines later. The old code also wasn't POSIX compliant as it used \n in the replacement section of the s-command. Using \<newline> is the POSIX compatible method. LC_ALL=C was added to the two critical sed commands. POSIX sed manual recommends it when using sed to manipulate pathnames because in other locales invalid multibyte sequences might cause issues with some sed implementations. In case of GNU sed, these particular sed scripts wouldn't have such problems but some other scripts could have, see: info '(sed)Locale Considerations' This vulnerability was discovered by: cleemy desu wayo working with Trend Micro Zero Day Initiative Thanks to Jim Meyering and Paul Eggert discussing the different ways to fix this and for coordinating the patch release schedule with gzip.
2022-03-29 18:19:12 +02:00
# $i already ends with a colon so do not add it here.
xzgrep: Fix escaping of malicious filenames (ZDI-CAN-16587). Malicious filenames can make xzgrep to write to arbitrary files or (with a GNU sed extension) lead to arbitrary code execution. xzgrep from XZ Utils versions up to and including 5.2.5 are affected. 5.3.1alpha and 5.3.2alpha are affected as well. This patch works for all of them. This bug was inherited from gzip's zgrep. gzip 1.12 includes a fix for zgrep. The issue with the old sed script is that with multiple newlines, the N-command will read the second line of input, then the s-commands will be skipped because it's not the end of the file yet, then a new sed cycle starts and the pattern space is printed and emptied. So only the last line or two get escaped. One way to fix this would be to read all lines into the pattern space first. However, the included fix is even simpler: All lines except the last line get a backslash appended at the end. To ensure that shell command substitution doesn't eat a possible trailing newline, a colon is appended to the filename before escaping. The colon is later used to separate the filename from the grep output so it is fine to add it here instead of a few lines later. The old code also wasn't POSIX compliant as it used \n in the replacement section of the s-command. Using \<newline> is the POSIX compatible method. LC_ALL=C was added to the two critical sed commands. POSIX sed manual recommends it when using sed to manipulate pathnames because in other locales invalid multibyte sequences might cause issues with some sed implementations. In case of GNU sed, these particular sed scripts wouldn't have such problems but some other scripts could have, see: info '(sed)Locale Considerations' This vulnerability was discovered by: cleemy desu wayo working with Trend Micro Zero Day Initiative Thanks to Jim Meyering and Paul Eggert discussing the different ways to fix this and for coordinating the patch release schedule with gzip.
2022-03-29 18:19:12 +02:00
sed_script="s|^|$i|"
# If grep or sed fails, pick the larger value of the two exit statuses.
# If sed fails, use at least 2 since we use >= 2 to indicate errors.
r=$(
exec 4>&1
xzgrep: Fix escaping of malicious filenames (ZDI-CAN-16587). Malicious filenames can make xzgrep to write to arbitrary files or (with a GNU sed extension) lead to arbitrary code execution. xzgrep from XZ Utils versions up to and including 5.2.5 are affected. 5.3.1alpha and 5.3.2alpha are affected as well. This patch works for all of them. This bug was inherited from gzip's zgrep. gzip 1.12 includes a fix for zgrep. The issue with the old sed script is that with multiple newlines, the N-command will read the second line of input, then the s-commands will be skipped because it's not the end of the file yet, then a new sed cycle starts and the pattern space is printed and emptied. So only the last line or two get escaped. One way to fix this would be to read all lines into the pattern space first. However, the included fix is even simpler: All lines except the last line get a backslash appended at the end. To ensure that shell command substitution doesn't eat a possible trailing newline, a colon is appended to the filename before escaping. The colon is later used to separate the filename from the grep output so it is fine to add it here instead of a few lines later. The old code also wasn't POSIX compliant as it used \n in the replacement section of the s-command. Using \<newline> is the POSIX compatible method. LC_ALL=C was added to the two critical sed commands. POSIX sed manual recommends it when using sed to manipulate pathnames because in other locales invalid multibyte sequences might cause issues with some sed implementations. In case of GNU sed, these particular sed scripts wouldn't have such problems but some other scripts could have, see: info '(sed)Locale Considerations' This vulnerability was discovered by: cleemy desu wayo working with Trend Micro Zero Day Initiative Thanks to Jim Meyering and Paul Eggert discussing the different ways to fix this and for coordinating the patch release schedule with gzip.
2022-03-29 18:19:12 +02:00
(eval "$grep" 4>&-; echo $? >&4) 3>&- |
LC_ALL=C sed "$sed_script" >&3 4>&-
) || {
sed_status=$?
test "$sed_status" -lt 2 && sed_status=2
test "$r" -lt "$sed_status" && r=$sed_status
}
exit $r
fi >&3 5>&-
)
r=$?
# If grep or sed or other non-decompression command failed with a signal,
# exit immediately and ignore the possible remaining files.
#
# NOTE: Instead of 128 + signal_number, some shells use
# 256 + signal_number (ksh) or 384 + signal_number (yash).
# This is fine for us since their "exit" and "kill -l" commands take
# this into account. (At least the versions I tried do but there is
# a report of an old ksh variant whose "exit" truncates the exit status
# to 8 bits without any special handling for values indicating a signal.)
test "$r" -ge 128 && exit "$r"
if test -z "$xz_status"; then
# Something unusual happened, for example, we got a signal and
# the exit status of the decompressor was never echoed and thus
# $xz_status is empty. Exit immediately and ignore the possible
# remaining files.
exit 2
elif test "$xz_status" -ge 128; then
# The decompressor died due to a signal. SIGPIPE is ignored since it can
# occur if grep exits before the whole file has been decompressed (grep -q
# can do that). If the decompressor died with some other signal, exit
# immediately and ignore the possible remaining files.
test "$(kill -l "$xz_status" 2> /dev/null)" != "PIPE" && exit "$xz_status"
elif test "$xz_status" -gt 0; then
# Decompression failed but we will continue with the remaining
# files anwyway. Set exit status to at least 2 to indicate an error.
test "$r" -lt 2 && r=2
fi
# Since res=1 is the initial value, we only need to care about
# matches (r == 0) and errors (r >= 2) here; r == 1 can be ignored.
if test "$r" -ge 2; then
# An error occurred in decompressor, grep, or some other command. Update
# res unless a larger error code has been seen with an earlier file.
test "$res" -lt "$r" && res=$r
elif test "$r" -eq 0; then
# grep found a match and no errors occurred. Update res if no errors have
# occurred with earlier files.
test "$res" -eq 1 && res=0
fi
done
# 0: At least one file matched and no errors occurred.
# 1: No matches were found and no errors occurred.
# >=2: Error. It's unknown if matches were found.
exit "$res"