Jim Meyering wrote:
Might be tricky to portably transform that NUL byte into something we
can embed in a command-line-specified search string. Is there even a
notation for that? I don't think so.
But NUL problems aside, this also should work, requiring alternation
in the regexp derived from input with two or more lines, but then
we'll have to escape embedded '|' bytes, too:
How about the attached patch instead? It uses a bigger hammer, which should
address both issues.
>From a5c927ea71ccc86fbb90ab0ea6083bf3cdcd9472 Mon Sep 17 00:00:00 2001
From: Paul Eggert <egg...@cs.ucla.edu>
Date: Thu, 17 Mar 2016 13:08:06 -0700
Subject: [PATCH] zgrep: with -f SPECIAL, read SPECIAL just once
Problem reported by Fulvio Scapin in: http://bugs.gnu.org/22945
* NEWS: Document this.
* tests/zgrep-f: Add a test.
* zgrep.in (with_filename): With -f FILE, if FILE is stdin or not
a regular file, copy it into a temporary and use the temporary.
---
NEWS | 3 +++
tests/zgrep-f | 8 ++++++++
zgrep.in | 45 ++++++++++++++++++++++++++++++++++++++-------
3 files changed, 49 insertions(+), 7 deletions(-)
diff --git a/NEWS b/NEWS
index 6363d71..541ad94 100644
--- a/NEWS
+++ b/NEWS
@@ -35,6 +35,9 @@ GNU gzip NEWS -*- outline -*-
gzip -k -v no longer reports that files are replaced.
[bug present since the beginning]
+ zgrep -f A B C no longer reads A more than once if A is not a regular file.
+ This better supports invocations like 'zgrep -f <(COMMAND) B C' in Bash.
+ [bug introduced in gzip-1.2]
* Noteworthy changes in release 1.6 (2013-06-09) [stable]
diff --git a/tests/zgrep-f b/tests/zgrep-f
index a8eb746..9a86550 100755
--- a/tests/zgrep-f
+++ b/tests/zgrep-f
@@ -29,6 +29,14 @@ zgrep -f - haystack.gz < n > out 2>&1 || fail=1
compare out n || fail=1
+if ${BASH_VERSION+:} false; then
+ set +o posix
+ # This failed with gzip 1.6.
+ cat n n >nn || framework_failure_
+ eval 'zgrep -h -f <(cat n) haystack.gz haystack.gz' >out || fail=1
+ compare out nn || fail=1
+fi
+
# This failed with gzip 1.4.
echo a-b | zgrep -e - > /dev/null || fail=1
diff --git a/zgrep.in b/zgrep.in
index c24be57..06baf38 100644
--- a/zgrep.in
+++ b/zgrep.in
@@ -55,6 +55,7 @@ files_with_matches=0
files_without_matches=0
no_filename=0
with_filename=0
+pattmp=
while test $# -ne 0; do
option=$1
@@ -113,13 +114,34 @@ while test $# -ne 0; do
# The pattern is coming from a file rather than the command-line.
# If the file is actually stdin then we need to do a little
# magic, since we use stdin to pass the gzip output to grep.
- # Turn the -f option into an -e option by copying the file's
- # contents into OPTARG.
- case $optarg in
- (" '-'" | " '/dev/stdin'" | " '/dev/fd/0'")
- option=-e
- optarg=" '"$(sed "$escape") || exit 2;;
- esac
+ # Similarly if it is not a regular file, since it might be read repeatedly.
+ # In either of these two cases, copy the pattern into a temporary file,
+ # and use that file instead. The pattern might contain null bytes,
+ # so we cannot simply switch to -e here.
+ if case $optarg in
+ (" '-'" | " '/dev/stdin'" | " '/dev/fd/0'")
+ :;;
+ (*)
+ eval "test ! -f$optarg";;
+ esac
+ then
+ if test -n "$pattmp"; then
+ eval "cat --$optarg" >>"$pattmp"
+ continue
+ fi
+ trap '
+ test -n "$pattmp" && rm -f "$pattmp"
+ (exit 2); exit 2
+ ' HUP INT PIPE TERM 0
+ if type mktemp >/dev/null 2>&1; then
+ pattmp=$(mktemp -t -- "zgrep.XXXXXX") || exit 2
+ else
+ set -C
+ pattmp=${TMPDIR-/tmp}/zgrep.$$
+ fi
+ eval "cat --$optarg" >"$pattmp"
+ optarg=' "$pattmp"'
+ fi
have_pat=1;;
(--h | --he | --hel | --help)
echo "$usage" || exit 2
@@ -232,5 +254,14 @@ do
test 126 -le $res && break
done
+if test -n "$pattmp"; then
+ rm -f "$pattmp" || {
+ r=$?
+ test $r -lt 2 && r=2
+ test $res -lt $r && res=$r
+ }
+ trap - HUP INT PIPE TERM 0
+fi
+
test 128 -le $res && kill -$(expr $res % 128) $$
exit $res
--
2.5.0
