Fix some bashisms
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 8 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are more or less the same as in the previous patches, except for a few exceptions). Note that the patch is not complete, in the sense that there may remain further bashisms in the source even after applying the patch (like use of arrays, which is visible even from this patch).
Regards
Jiri PalecekSigned-off-by Jiri Palecek <[email protected]>
Hmm,. Some of them failed to apply. Can you please resend only the error part(s):
patching file testcases/kernel/fs/fs-bench/modaltr.sh Hunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching file testcases/kernel/fs/fs_di/fs_di patching file testcases/kernel/fs/mongo/test.sh Hunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01 Hunk #1 FAILED at 77.
1 out of 1 hunk FAILED -- saving rejects to file testcases/network/tcp_cmds/netstat/netstat01.rej
Regards--
Subrata
Quoted 612 lines
--- .../controllers/cgroup_fj/cgroup_fj_utility.sh | 2 +- .../controllers/cgroup_fj/run_cgroup_test_fj.sh | 8 ++-- .../controllers/cpuctl_fj/run_cpuctl_test_fj.sh | 51 ++++++++++---------- .../memcg/functional/memcg_function_test.sh | 6 +- .../controllers/memcg/stress/memcg_stress_test.sh | 18 ++++---- testcases/kernel/fs/acls/acl_test01 | 12 ++-- testcases/kernel/fs/fs-bench/modaltr.sh | 2 +- testcases/kernel/fs/fs_di/fs_di | 4 +- testcases/kernel/fs/mongo/test.sh | 8 +-- .../hotplug/cpu_hotplug/functional/hotplug02.sh | 6 +- .../hotplug/cpu_hotplug/functional/hotplug03.sh | 16 +++--- .../hotplug/cpu_hotplug/functional/hotplug04.sh | 10 ++-- .../cpu_hotplug/functional/hotplug06.sar.sh | 6 +- .../cpu_hotplug/functional/hotplug06.top.sh | 6 +- .../hotplug/cpu_hotplug/functional/hotplug07.sh | 6 +- .../network/can/filter-tests/run_ltp-can_tests.sh | 2 +- testcases/network/tcp_cmds/netstat/netstat01 | 2 +- 17 files changed, 82 insertions(+), 83 deletions(-)testcases/kernel/controllers/cgroup_fj/cgroup_fj_utility.sh +1
-1@@ -547,7 +547,7 @@ reclaim_foundling() if [ -e /proc/$cur_pid/ ];then echo "ERROR: pid $cur_pid reclaimed" do_echo 0 1 "$cur_pid" "/dev/cgroup/tasks" - let "foundlings = $foundlings + 1" + : $((foundlings += 1)) fi donetestcases/kernel/controllers/cgroup_fj/run_cgroup_test_fj.sh +4
-4@@ -33,7 +33,7 @@ export TMPFILE=$TESTROOT/tmp_tasks export CASENO1=0 export CASENO2=0 -if [ "$UID" -ne 0 ]; then +if [ "$(id -ru)" -ne 0 ]; then echo "***error***:you must use root to test" exit -1 fi@@ -107,17 +107,17 @@ do if [ $? -ne 0 ]; then case $CASETYPE in "function" ) - let "CASENO1 = $CASENO1 + 1" + : $(( CASENO1 += 1 )) export CASENO1=$CASENO1 $TESTROOT/cgroup_fj_function.sh $CASECMD ;; "function2" ) - let "CASENO1 = $CASENO1 + 1" + : $(( CASENO1 += 1 )) export CASENO1=$CASENO1 $TESTROOT/cgroup_fj_function2.sh $CASECMD ;; "stress" ) - let "CASENO2 = $CASENO2 + 1" + : $(( CASENO2 += 1 )) export CASENO2=$CASENO2 $TESTROOT/cgroup_fj_stress.sh $CASECMD ;;testcases/kernel/controllers/cpuctl_fj/run_cpuctl_test_fj.sh +26
-25@@ -58,8 +58,8 @@ setup() # Write the cleanup function cleanup() { - mount | grep "$CPUCTL" &> /dev/null || { - rm -rf "$CPUCTL" &> /dev/null + mount | grep "$CPUCTL" > /dev/null 2>&1 || { + rm -rf "$CPUCTL" > /dev/null 2>&1 return 0 }@@ -73,22 +73,19 @@ cleanup() done umount $CPUCTL || return 1 - rmdir $CPUCTL &> /dev/null + rmdir $CPUCTL > /dev/null 2>&1 } creat_process() { cat /dev/zero > /dev/null 2>/dev/null & - taskset -p 1 $! &>/dev/null + taskset -p 1 $! >/dev/null 2>&1 echo $! } get_cpu_usage() { - top=($(top -b -n 1 -p $1 | tail -2 | head -1)) - top=${top[8]} - top=`echo $top | awk -F "." '{print $1}'` - echo "$top" + top -b -n 1 -p $1 | tail -2 | head -1 | awk 'print $9' | awk -F "." '{print $1}' } kill_all_pid()@@ -424,7 +421,7 @@ case12 () wait $pid } -max_shares=$((2**18)) +max_shares=$((1 << 18)) # echo MAX_SHARES into shares case13 ()@@ -589,7 +586,7 @@ case19() echo "pid $pid cpu_usage $cpu_usage" kill -9 $pid $pid_other - expr 96 \< "$cpu_usage" \& "$cpu_usage" \< 103 &> /dev/null || return 1 + expr 96 \< "$cpu_usage" \& "$cpu_usage" \< 103 > /dev/null 2>&1 || return 1 return 0 }@@ -605,7 +602,7 @@ case20() echo "pid $pid cpu_usage $cpu_usage" kill -9 $pid $pid_other - expr 96 \< "$cpu_usage" \& "$cpu_usage" \< 103 &> /dev/null || return 1 + expr 96 \< "$cpu_usage" \& "$cpu_usage" \< 103 > /dev/null 2>&1 || return 1 return 0 }@@ -615,7 +612,7 @@ case21() pid=$(creat_process) echo $pid > "$CPUCTL/1/tasks" - while ((1)) + while true do creat_process > "$CPUCTL/1/2/tasks" sleep 1@@ -625,18 +622,20 @@ case21() sleep 10 ret=0 - for ((top_times=0; top_times<10 && ret==0; top_times++)) + top_times=0 + while [ "$top_times" -lt 10 -a "$ret" = 0 ] do cpu_usage=$(get_cpu_usage $pid) echo "pid $pid cpu_usage $cpu_usage" - expr 44 \< "$cpu_usage" \& "$cpu_usage" \< 56 &> /dev/null + expr 44 \< "$cpu_usage" \& "$cpu_usage" \< 56 > /dev/null 2>&1 ret=$? + : $(( top_times+=1 )) done - kill -9 $pid $loop_pid &> /dev/null - wait $pid $loop_pid &>/dev/null + kill -9 $pid $loop_pid > /dev/null 2>&1 + wait $pid $loop_pid >/dev/null 2>&1 sleep 2 - kill_all_pid < "$CPUCTL/1/2/tasks" &>/dev/null + kill_all_pid < "$CPUCTL/1/2/tasks" >/dev/null 2>&1 sleep 2 return $ret }@@ -645,7 +644,7 @@ case22() { mkdir "$CPUCTL/1" "$CPUCTL/1/2" pid=$(creat_process) - while ((1)) + while true do echo $pid > "$CPUCTL/1/tasks" echo $pid > "$CPUCTL/1/2/tasks"@@ -656,26 +655,28 @@ case22() sleep 10 ret=0 - for ((top_times=0; top_times<10 && ret==0; top_times++)) + top_times=0 + while [ "$top_times" -lt 10 -a "$ret" = 0 ] do cpu_usage=$(get_cpu_usage $pid) echo "pid $pid cpu_usage $cpu_usage" - expr 94 \< "$cpu_usage" \& "$cpu_usage" \< 106 &> /dev/null + expr 94 \< "$cpu_usage" \& "$cpu_usage" \< 106 > /dev/null 2>&1 ret=$? + : $(( top_times+=1 )) done - kill -9 $pid $loop_pid &> /dev/null - wait $pid $loop_pid &>/dev/null + kill -s KILL $pid $loop_pid > /dev/null 2>&1 + wait $pid $loop_pid >/dev/null 2>&1 return $ret } -rm -rf "$CPUCTL_TMP" &>/dev/null || exit 1 +rm -rf "$CPUCTL_TMP" >/dev/null 2>&1 || exit 1 mkdir -p "$CPUCTL_TMP" || exit 1 # test do_test () { - for (( i=1; i<=$TST_TOTAL; i++ )) + for i in $(seq 1 $TST_TOTAL) do setup || { tst_resm TFAIL "case$i FAIL"@@ -698,5 +699,5 @@ do_test () do_test -rm -rf "$CPUCTL_TMP" &>/dev/null +rm -rf "$CPUCTL_TMP" >/dev/null 2>&1testcases/kernel/controllers/memcg/functional/memcg_function_test.sh +3
-3@@ -536,8 +536,8 @@ do_mount() } # Run all the test cases -for ((i = 1; i <= $TST_TOTAL; i++)) -{ +for i in $(seq 1 $TST_TOTAL) +do export TST_COUNT=$(( $TST_COUNT + 1 )) cur_id=$i@@ -556,7 +556,7 @@ for ((i = 1; i <= $TST_TOTAL; i++)) rmdir /dev/memcg/$i cleanup; -} +done echo $shmmax > /proc/sys/kernel/shmmaxtestcases/kernel/controllers/memcg/stress/memcg_stress_test.sh +9
-9@@ -66,29 +66,29 @@ run_stress() { do_mount; - for ((i = 0; i < $1; i++)) - { + for i in $(seq 0 $(($1-1))) + do mkdir /dev/memcg/$i 2> /dev/null ./memcg_process_stress $2 $3 & pid[$i]=$! echo ${pid[$i]} > /dev/memcg/$i/tasks - } + done - for ((i = 0; i < $1; i++)) - { + for i in $(seq 0 $(($1-1))) + do /bin/kill -s SIGUSR1 ${pid[$i]} 2> /dev/null - } + done sleep $4 - for ((i = 0; i < $1; i++)) - { + for i in $(seq 0 $(($1-1))) + do /bin/kill -s SIGINT ${pid[$i]} 2> /dev/null wait ${pid[$i]} rmdir /dev/memcg/$i 2> /dev/null - } + done cleanup; }testcases/kernel/fs/acls/acl_test01 +6
-6@@ -51,7 +51,7 @@ if [ "z$iam" = "z$TEST_USER1" ] then echo "" else -if [ $UID != 0 ] +if [ $(id -ru) != 0 ] then echo "FAILED: Must have root access to execute this script" exit 1@@ -76,7 +76,7 @@ if [ $? != 0 ] then echo "" echo "FAILED: [ losetup ] Must have loop device support by kernel" - echo -e "\t to execute this script" + printf "\t to execute this script\n" exit 1 fi@@ -90,8 +90,8 @@ then then echo "" echo "FAILED: [ mount ] Make sure that ACL (Access Control List)" - echo -e "\t and Extended Attribute are built into the kernel" - echo -e "\t Can not mount ext3 file system with acl and user_xattr options" + printf "\t and Extended Attribute are built into the kernel\n" + printf "\t Can not mount ext3 file system with acl and user_xattr options\n" exit 1 fi else@@ -102,8 +102,8 @@ else then echo "" echo "FAILED: [ mount ] Make sure that ACL (Access Control List)" - echo -e "\t and Extended Attribute are built into the kernel" - echo -e "\t Can not mount ext2 file system with acl and user_xattr options" + printf "\t and Extended Attribute are built into the kernel\n" + printf "\t Can not mount ext2 file system with acl and user_xattr options\n" exit 1 fi fitestcases/kernel/fs/fs-bench/modaltr.sh +1
-1@@ -43,7 +43,7 @@ LOG_DIR1=/tmp/log1 HOME_DIR=/home BLOCK_DIR=/dev/mtdblock export PATH=$PATH:/sbin - if [ $UID -ne 0 ]; + if [ $(id -ru) -ne 0 ]; then echo "must be root to run this" exittestcases/kernel/fs/fs_di/fs_di +2
-2@@ -201,10 +201,10 @@ $trace_logic end_testcase "Could not create testfile of size ${SIZE}Mb" fi RANDOM_DEPTH=$RANDOM - let "RANDOM_DEPTH %= 500" + : $(( RANDOM_DEPTH %= 500 )) RANDOM_LENGTH=$RANDOM - let "RANDOM_LENGTH %= 500" + : $(( RANDOM_LENGTH %= 500 )) RANDOM_LENGTH=$(( $RANDOM_LENGTH / 10 )) NameCount=0testcases/kernel/fs/mongo/test.sh +3
-5@@ -26,7 +26,7 @@ TEST_DIR=testdir #should be root to execute this script . - if [ $UID -ne 0 ]; then + if [ $(id -ru) -ne 0 ]; then echo "This script must be run as root" exit fi@@ -52,8 +52,7 @@ modprobe loop fi #run the mongo test on reiserfs file system type -function reiserfs - +reiserfs() { cat > fs.sh <<EOF echo "performing mongo on reiserfs"@@ -71,8 +70,7 @@ EOF #To run on jfs file system type -function JFS - +JFS() { cat >> fs.sh <<EOF echo "performing mongo on jfs file system"testcases/kernel/hotplug/cpu_hotplug/functional/hotplug02.sh +3
-3@@ -15,8 +15,8 @@ fi # Includes: LHCS_PATH=${LHCS_PATH:-".."} -source $LHCS_PATH/include/testsuite.fns -source $LHCS_PATH/include/hotplug.fns +. $LHCS_PATH/include/testsuite.fns +. $LHCS_PATH/include/hotplug.fns echo "Name: $CASE" echo "Date: `date`"@@ -73,7 +73,7 @@ until [ $loop = 0 ]; do # Turn the CPU back online just to see what happens. online_cpu ${CPU_TO_TEST} - let "loop = loop - 1" + : $(( loop -= 1 )) done sleep 2testcases/kernel/hotplug/cpu_hotplug/functional/hotplug03.sh +8
-8@@ -15,8 +15,8 @@ fi # Includes: LHCS_PATH=${LHCS_PATH:-".."} -source $LHCS_PATH/include/testsuite.fns -source $LHCS_PATH/include/hotplug.fns +. $LHCS_PATH/include/testsuite.fns +. $LHCS_PATH/include/hotplug.fns echo "Name: $CASE" echo "Date: `date`"@@ -48,7 +48,7 @@ do_clean() # Turn off the CPUs that were off before the test start until [ $cpu = 0 ];do offline_cpu ${on[${cpu}]} - let "cpu = cpu - 1" + : $(( cpu -= 1 )) done }@@ -60,10 +60,10 @@ until [ $loop = 0 ]; do for i in $( get_all_cpus ); do online_cpu $1 if [ $? = 0 ]; then - let "cpu = cpu + 1" + : $(( cpu += 1 )) on[${cpu}]=$i fi - let "number_of_cpus = number_of_cpus + 1" + : $(( number_of_cpus += 1 )) done offline_cpu ${CPU_TO_TEST}@@ -76,11 +76,11 @@ until [ $loop = 0 ]; do # CPUs we have. This is to help ensure we've got enough processes # that at least one will migrate to the new CPU. Store the PIDs # so we can kill them later. - let "number_of_cpus = number_of_cpus * 2" + : $(( number_of_cpus *= 2 )) until [ $number_of_cpus = 0 ]; do $LHCS_PATH/tools/do_spin_loop > /dev/null 2>&1 & echo $! >> /var/run/hotplug4_$$.pid - let "number_of_cpus = number_of_cpus - 1" + : $(( number_of_cpus -= 1 )) done ps aux | head -n 1@@ -111,7 +111,7 @@ until [ $loop = 0 ]; do do_clean - let "loop = loop -1" + : $(( loop -=1 )) done exit_cleantestcases/kernel/hotplug/cpu_hotplug/functional/hotplug04.sh +5
-5@@ -9,8 +9,8 @@ loop=${HOTPLUG04_LOOPS:-1} # Includes: LHCS_PATH=${LHCS_PATH:-".."} -source $LHCS_PATH/include/testsuite.fns -source $LHCS_PATH/include/hotplug.fns +. $LHCS_PATH/include/testsuite.fns +. $LHCS_PATH/include/hotplug.fns echo "Name: $CASE" echo "Date: `date`"@@ -26,7 +26,7 @@ until [ $loop = 0 ]; do online_cpu $i RC=$? if [ $RC != 0 ]; then - let "cpu = cpu + 1" + : $(( cpu += 1 )) on[${cpu}]=$i echo "${on[${cpu}]}" fi@@ -51,7 +51,7 @@ until [ $loop = 0 ]; do # Online the ones that were on initially until [ $cpu = 0 ]; do online_cpu ${on[${cpu}]} - let "cpu = cpu - 1" + : $(( cpu -= 1 )) done # Return CPU 0 to its initial state@@ -61,7 +61,7 @@ until [ $loop = 0 ]; do offline_cpu 0 fi - let "loop = loop -1" + : $(( loop -=1 )) done exit_cleantestcases/kernel/hotplug/cpu_hotplug/functional/hotplug06.sar.sh +3
-3@@ -15,8 +15,8 @@ fi # Includes: LHCS_PATH=${LHCS_PATH:-".."} -source $LHCS_PATH/include/testsuite.fns -source $LHCS_PATH/include/hotplug.fns +. $LHCS_PATH/include/testsuite.fns +. $LHCS_PATH/include/hotplug.fns echo "Name: $CASE" echo "Date: `date`"@@ -81,7 +81,7 @@ until [ $loop = 0 ]; do # Check that SAR registered the change in CPU online/offline states NEW_NUMBER_CPU_OFF=`cat /tmp/log_$$ | grep ${time} | grep -i nan | wc -l` - let "NUMBER_CPU_OFF = NUMBER_CPU_OFF - 1" + : $(( NUMBER_CPU_OFF -= 1 )) if [ $NUMBER_CPU_OFF = $NEW_NUMBER_CPU_OFF ]; then echo "$CASE PASS: CPU was found after turned on." elsetestcases/kernel/hotplug/cpu_hotplug/functional/hotplug06.top.sh +3
-3@@ -15,8 +15,8 @@ fi # Includes: LHCS_PATH=${LHCS_PATH:-".."} -source $LHCS_PATH/include/testsuite.fns -source $LHCS_PATH/include/hotplug.fns +. $LHCS_PATH/include/testsuite.fns +. $LHCS_PATH/include/hotplug.fns echo "Name: $CASE" echo "Date: `date`"@@ -76,7 +76,7 @@ until [ $loop = 0 ]; do exit_clean -1 fi - let "loop = loop - 1" + : $(( loop -= 1 )) donetestcases/kernel/hotplug/cpu_hotplug/functional/hotplug07.sh +3
-3@@ -11,8 +11,8 @@ loop_one=${HOTPLUG07_LOOPS:-1} # Includes: LHCS_PATH=${LHCS_PATH:-".."} -source $LHCS_PATH/include/testsuite.fns -source $LHCS_PATH/include/hotplug.fns +. $LHCS_PATH/include/testsuite.fns +. $LHCS_PATH/include/hotplug.fns echo "Name: HotPlug Test - Test Case 7" echo "Date: `date`"@@ -89,7 +89,7 @@ do RC=$? echo "Onlining cpu${CPU_TO_TEST}: Return Code = ${RC}" - let "loop_one = loop_one - 1" + : $(( loop_one -= 1 )) donetestcases/network/can/filter-tests/run_ltp-can_tests.sh +1
-1@@ -19,7 +19,7 @@ ## ## ################################################################################ -if [ $UID -ne 0 ]; then +if [ $(id -ru) -ne 0 ]; then echo You need to be root to execute these tests exit 1 fitestcases/network/tcp_cmds/netstat/netstat01 +0
-0
testcases/network/tcp_cmds/netstat/netstat01 do_test() { · line 83
Quoted 8 lines
testcases/network/tcp_cmds/netstat/netstat01 +1
-1@@ -77,7 +77,7 @@ do_test() { do for COMMAND in "netstat -s" "netstat -rn" "netstat -i" "netstat -gn" "netstat -apn" do - $COMMAND 2>&1 1>/dev/null + $COMMAND 1>/dev/null 2>&1 [ $? -eq 0 ] || end_testcase "$COMMAND failed" done
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Quoted 5 lines · Subrata Modak
Quoted 4 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are more or less the same as in the previous patches, except for a few exceptions). Note that the patch is not complete, in the sense that there may remain further bashisms in the source even after applying the patch (like use of arrays, which is visible even from this patch).
Quoted 7 lines
Quoted 4 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <[email protected]>Hmm,. Some of them failed to apply. Can you please resend only the error part(s):
Sorry for not catching this sooner. netstat01 (and a lot of other items under tcp_cmds) were rewritten to 1) remove bashisms, 2) make the testing procedures more uniform, and 3) (the most important) remove duplicate code in ipv6 and tcp_cmds.
Thanks,
-Garrett
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are more or
less the same as in the previous
Quoted 1 line
patches, except for a few exceptions). Note that the patch is not complete,
in the sense that there may
Quoted 1 line
remain further bashisms in the source even after applying the patch (like use
of arrays, which is visible
Quoted 10 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>Hmm,. Some of them failed to apply. Can you please resend only the error part(s):
according to my git repository. these are the missing parts:
--- testcases/kernel/fs/fs-bench/modaltr.sh | 2 +- testcases/kernel/fs/mongo/test.sh | 8 +++----- testcases/network/tcp_cmds/netstat/netstat01 | 2 +- 3 files changed, 5 insertions(+), 7 deletions(-)
testcases/kernel/fs/fs- renamed from testcases/kernel/fs/fs-bench/modaltr.sh +0 -0
bench/modaltr.sh
index 85d2c30..36e6f0c 100755
testcases/kernel/fs/fs-bench/modaltr.sh +1 -1
@@ -43,7 +43,7 @@ LOG_DIR1=/tmp/log1 HOME_DIR=/home BLOCK_DIR=/dev/mtdblock export PATH=$PATH:/sbin - if [ $UID -ne 0 ]; + if [ $(id -ru) -ne 0 ]; then echo "must be root to run this" exit
testcases/kernel/fs/mongo/ renamed from testcases/kernel/fs/mongo/test.sh +0 -0
test.sh
index fe75664..eccb079 100755
testcases/kernel/fs/mongo/test.sh +3 -5
@@ -26,7 +26,7 @@ TEST_DIR=testdir #should be root to execute this script . - if [ $UID -ne 0 ]; then + if [ $(id -ru) -ne 0 ]; then echo "This script must be run as root" exit fi
@@ -52,8 +52,7 @@ modprobe loop fi #run the mongo test on reiserfs file system type -function reiserfs - +reiserfs() { cat > fs.sh <<EOF echo "performing mongo on reiserfs"
@@ -71,8 +70,7 @@ EOF #To run on jfs file system type -function JFS - +JFS() { cat >> fs.sh <<EOF echo "performing mongo on jfs file system"
testcases/network/ renamed from testcases/network/tcp_cmds/netstat/netstat01 +0 -0
tcp_cmds/netstat/netstat01
index 6a197de..365c142 100755
testcases/network/tcp_cmds/netstat/netstat01 +1 -1
@@ -58,7 +58,7 @@ do_test() { while [ $TST_COUNT -le $NUMLOOPS ]; do for flag in "-s" "-rn" "-i" "-gn" "-apn"; do - if ! netstat $flag 2>&1 1>/dev/null; then + if ! netstat $flag 1>/dev/null 2>&1; then end_testcase "$COMMAND failed" fi done
Signature · 15 lines
1.6.4.3 ------------------------------------------------------------------------------ Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now! http://p.sf.net/sfu/devconference _______________________________________________ Ltp-list mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ltp-list
Quoted 21 lines · JiříPaleček
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are more or
less the same as in the previous
Quoted 1 line
patches, except for a few exceptions). Note that the patch is not complete,
in the sense that there may
Quoted 1 line
remain further bashisms in the source even after applying the patch (like use
of arrays, which is visible
Quoted 7 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>
Quoted 7 lines
Quoted 3 lines
Hmm,. Some of them failed to apply. Can you please resend only the error part(s):
according to my git repository. these are the missing parts:
Signed-off-by Jiri Palecek <[email protected]>
Then something wrong with my CVS:
patching file testcases/kernel/fs/fs-bench/modaltr.sh Hunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching file testcases/kernel/fs/mongo/test.sh Hunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01
Can you please verify ?
Regards--
Subrata
Quoted 60 lines
--- testcases/kernel/fs/fs-bench/modaltr.sh | 2 +- testcases/kernel/fs/mongo/test.sh | 8 +++----- testcases/network/tcp_cmds/netstat/netstat01 | 2 +- 3 files changed, 5 insertions(+), 7 deletions(-)testcases/kernel/fs/fs- renamed from testcases/kernel/fs/fs-bench/modaltr.sh +0
-0
bench/modaltr.sh
index 85d2c30..36e6f0c 100755testcases/kernel/fs/fs-bench/modaltr.sh +1
-1@@ -43,7 +43,7 @@ LOG_DIR1=/tmp/log1 HOME_DIR=/home BLOCK_DIR=/dev/mtdblock export PATH=$PATH:/sbin - if [ $UID -ne 0 ]; + if [ $(id -ru) -ne 0 ]; then echo "must be root to run this" exittestcases/kernel/fs/mongo/ renamed from testcases/kernel/fs/mongo/test.sh +0
-0
test.sh
index fe75664..eccb079 100755testcases/kernel/fs/mongo/test.sh +3
-5@@ -26,7 +26,7 @@ TEST_DIR=testdir #should be root to execute this script . - if [ $UID -ne 0 ]; then + if [ $(id -ru) -ne 0 ]; then echo "This script must be run as root" exit fi@@ -52,8 +52,7 @@ modprobe loop fi #run the mongo test on reiserfs file system type -function reiserfs - +reiserfs() { cat > fs.sh <<EOF echo "performing mongo on reiserfs"@@ -71,8 +70,7 @@ EOF #To run on jfs file system type -function JFS - +JFS() { cat >> fs.sh <<EOF echo "performing mongo on jfs file system"testcases/network/ renamed from testcases/network/tcp_cmds/netstat/netstat01 +0
-0tcp_cmds/netstat/netstat01
index 6a197de..365c142 100755
testcases/network/tcp_cmds/netstat/netstat01 do_test() { · line 64
Quoted 8 lines
testcases/network/tcp_cmds/netstat/netstat01 +1
-1@@ -58,7 +58,7 @@ do_test() { while [ $TST_COUNT -le $NUMLOOPS ]; do for flag in "-s" "-rn" "-i" "-gn" "-apn"; do - if ! netstat $flag 2>&1 1>/dev/null; then + if ! netstat $flag 1>/dev/null 2>&1; then end_testcase "$COMMAND failed" fi done
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Quoted 48 lines · Subrata Modak
Quoted 24 lines · JiříPaleček
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are
more or
less the same as in the previousQuoted 1 line
patches, except for a few exceptions). Note that the patch is not
complete,
in the sense that there mayQuoted 1 line
remain further bashisms in the source even after applying the patch
(like use
of arrays, which is visibleQuoted 7 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>Quoted 8 lines
Quoted 1 line
Hmm,. Some of them failed to apply. Can you please resend only the
error
Quoted 2 lines
part(s):
according to my git repository. these are the missing parts:
Signed-off-by Jiri Palecek <[email protected]>Then something wrong with my CVS:
patching file
testcases/kernel/fs/fs-bench/modaltr.shHunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching filetestcases/kernel/fs/mongo/test.shHunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01
Quoted 1 line
Can you please verify ?
It works for me in CVS, so I guess it was just a linewrap, whitespace or something issue. See attachment for the original (uncrippled) patch.
Regards
Jiri Palecek
Quoted 69 lines
Quoted 68 lines
--- testcases/kernel/fs/fs-bench/modaltr.sh | 2 +- testcases/kernel/fs/mongo/test.sh | 8 +++----- testcases/network/tcp_cmds/netstat/netstat01 | 2 +- 3 files changed, 5 insertions(+), 7 deletions(-)testcases/kernel/fs/fs- renamed from testcases/kernel/fs/fs-bench/modaltr.sh +0
-0
bench/modaltr.sh
index 85d2c30..36e6f0c 100755testcases/kernel/fs/fs-bench/modaltr.sh +1
-1@@ -43,7 +43,7 @@ LOG_DIR1=/tmp/log1 HOME_DIR=/home BLOCK_DIR=/dev/mtdblock export PATH=$PATH:/sbin - if [ $UID -ne 0 ]; + if [ $(id -ru) -ne 0 ]; then echo "must be root to run this" exittestcases/kernel/fs/mongo/ renamed from testcases/kernel/fs/mongo/test.sh +0
-0
test.sh
index fe75664..eccb079 100755testcases/kernel/fs/mongo/test.sh +3
-5@@ -26,7 +26,7 @@ TEST_DIR=testdir #should be root to execute this script . - if [ $UID -ne 0 ]; then + if [ $(id -ru) -ne 0 ]; then echo "This script must be run as root" exit fi@@ -52,8 +52,7 @@ modprobe loop fi #run the mongo test on reiserfs file system type -function reiserfs - +reiserfs() { cat > fs.sh <<EOF echo "performing mongo on reiserfs"@@ -71,8 +70,7 @@ EOF #To run on jfs file system type -function JFS - +JFS() { cat >> fs.sh <<EOF echo "performing mongo on jfs file system"testcases/network/ renamed from testcases/network/tcp_cmds/netstat/netstat01 +0
-0tcp_cmds/netstat/netstat01
index 6a197de..365c142 100755testcases/network/tcp_cmds/netstat/netstat01 +1
-1@@ -58,7 +58,7 @@ do_test() { while [ $TST_COUNT -le $NUMLOOPS ]; do for flag in "-s" "-rn" "-i" "-gn" "-apn"; do - if ! netstat $flag 2>&1 1>/dev/null; then + if ! netstat $flag 1>/dev/null 2>&1; then end_testcase "$COMMAND failed" fi done
Signature · 1 line
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
From 8ba003dbc80f0280f1d68bf512329ddbe1a07dd0 Mon Sep 17 00:00:00 2001
From: Jiri Palecek <[email protected]>
Date: Wed, 21 Oct 2009 02:19:58 +0200
Subject: [PATCH] Fix some bashisms
--- testcases/kernel/fs/fs-bench/modaltr.sh | 2 +- testcases/kernel/fs/mongo/test.sh | 8 +++----- testcases/network/tcp_cmds/netstat/netstat01 | 2 +- 3 files changed, 5 insertions(+), 7 deletions(-)
testcases/kernel/fs/fs-bench/modaltr.sh +1 -1
@@ -43,7 +43,7 @@ LOG_DIR1=/tmp/log1 HOME_DIR=/home BLOCK_DIR=/dev/mtdblock export PATH=$PATH:/sbin - if [ $UID -ne 0 ]; + if [ $(id -ru) -ne 0 ]; then echo "must be root to run this" exit
testcases/kernel/fs/mongo/test.sh +3 -5
@@ -26,7 +26,7 @@ TEST_DIR=testdir #should be root to execute this script . - if [ $UID -ne 0 ]; then + if [ $(id -ru) -ne 0 ]; then echo "This script must be run as root" exit fi
@@ -52,8 +52,7 @@ modprobe loop fi #run the mongo test on reiserfs file system type -function reiserfs - +reiserfs() { cat > fs.sh <<EOF echo "performing mongo on reiserfs"
@@ -71,8 +70,7 @@ EOF #To run on jfs file system type -function JFS - +JFS() { cat >> fs.sh <<EOF echo "performing mongo on jfs file system"
testcases/network/tcp_cmds/netstat/netstat01 +1 -1
@@ -58,7 +58,7 @@ do_test() { while [ $TST_COUNT -le $NUMLOOPS ]; do for flag in "-s" "-rn" "-i" "-gn" "-apn"; do - if ! netstat $flag 2>&1 1>/dev/null; then + if ! netstat $flag 1>/dev/null 2>&1; then end_testcase "$COMMAND failed" fi done
Signature · 1 line
1.6.4.3
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Quoted 51 lines · Jiří Paleček
Quoted 48 lines · Subrata Modak
Quoted 24 lines · JiříPaleček
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are
more or
less the same as in the previousQuoted 1 line
patches, except for a few exceptions). Note that the patch is not
complete,
in the sense that there mayQuoted 1 line
remain further bashisms in the source even after applying the patch
(like use
of arrays, which is visibleQuoted 7 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>Quoted 8 lines
Quoted 1 line
Hmm,. Some of them failed to apply. Can you please resend only the
error
Quoted 2 lines
part(s):
according to my git repository. these are the missing parts:
Signed-off-by Jiri Palecek <[email protected]>Then something wrong with my CVS:
patching file
testcases/kernel/fs/fs-bench/modaltr.shHunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching filetestcases/kernel/fs/mongo/test.shHunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01
Quoted 4 lines
Quoted 1 line
Can you please verify ?
It works for me in CVS, so I guess it was just a linewrap, whitespace or something issue. See attachment for the original (uncrippled) patch.
Yes, it works fine now :-)
patching file testcases/kernel/fs/fs-bench/modaltr.sh patching file testcases/kernel/fs/mongo/test.sh patching file testcases/network/tcp_cmds/netstat/netstat01
Regards--
Subrata
Quoted 69 lines
Regards
Jiri PalecekQuoted 65 lines
Quoted 65 lines
--- testcases/kernel/fs/fs-bench/modaltr.sh | 2 +- testcases/kernel/fs/mongo/test.sh | 8 +++----- testcases/network/tcp_cmds/netstat/netstat01 | 2 +- 3 files changed, 5 insertions(+), 7 deletions(-)testcases/kernel/fs/fs- renamed from testcases/kernel/fs/fs-bench/modaltr.sh +0
-0
bench/modaltr.sh
index 85d2c30..36e6f0c 100755testcases/kernel/fs/fs-bench/modaltr.sh +1
-1@@ -43,7 +43,7 @@ LOG_DIR1=/tmp/log1 HOME_DIR=/home BLOCK_DIR=/dev/mtdblock export PATH=$PATH:/sbin - if [ $UID -ne 0 ]; + if [ $(id -ru) -ne 0 ]; then echo "must be root to run this" exittestcases/kernel/fs/mongo/ renamed from testcases/kernel/fs/mongo/test.sh +0
-0
test.sh
index fe75664..eccb079 100755testcases/kernel/fs/mongo/test.sh +3
-5@@ -26,7 +26,7 @@ TEST_DIR=testdir #should be root to execute this script . - if [ $UID -ne 0 ]; then + if [ $(id -ru) -ne 0 ]; then echo "This script must be run as root" exit fi@@ -52,8 +52,7 @@ modprobe loop fi #run the mongo test on reiserfs file system type -function reiserfs - +reiserfs() { cat > fs.sh <<EOF echo "performing mongo on reiserfs"@@ -71,8 +70,7 @@ EOF #To run on jfs file system type -function JFS - +JFS() { cat >> fs.sh <<EOF echo "performing mongo on jfs file system"testcases/network/ renamed from testcases/network/tcp_cmds/netstat/netstat01 +0
-0tcp_cmds/netstat/netstat01
index 6a197de..365c142 100755testcases/network/tcp_cmds/netstat/netstat01 +0
-0@@ -58,7 +58,7 @@ do_test() { while [ $TST_COUNT -le $NUMLOOPS ]; do
Quoted 12 lines
Quoted 7 lines
Quoted 6 lines
testcases/network/tcp_cmds/netstat/netstat01 +1
-1for flag in "-s" "-rn" "-i" "-gn" "-apn"; do - if ! netstat $flag 2>&1 1>/dev/null; then + if ! netstat $flag 1>/dev/null 2>&1; then end_testcase "$COMMAND failed" fi doneSignature · 1 line
Using Opera's revolutionary e-mail client: http://www.opera.com/mail/
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Quoted 54 lines · Subrata Modak
Quoted 53 lines · Jiří Paleček
Quoted 49 lines · Subrata Modak
Quoted 24 lines · JiříPaleček
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are
more or
less the same as in the previousQuoted 1 line
patches, except for a few exceptions). Note that the patch is not
complete,
in the sense that there mayQuoted 1 line
remain further bashisms in the source even after applying the patch
(like use
of arrays, which is visibleQuoted 7 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>Quoted 8 lines
Quoted 1 line
Hmm,. Some of them failed to apply. Can you please resend only the
error
Quoted 2 lines
part(s):
according to my git repository. these are the missing parts:
Signed-off-by Jiri Palecek <[email protected]>Then something wrong with my CVS:
patching file
testcases/kernel/fs/fs-bench/modaltr.shHunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching filetestcases/kernel/fs/mongo/test.shHunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01Can you please verify ?
Quoted 7 lines
Quoted 2 lines
It works for me in CVS, so I guess it was just a linewrap, whitespace or something issue. See attachment for the original (uncrippled) patch.
Yes, it works fine now :-)
patching filetestcases/kernel/fs/fs-bench/modaltr.shpatching filetestcases/kernel/fs/mongo/test.shpatching file testcases/network/tcp_cmds/netstat/netstat01
Sorry for replying late again, but just for future reference, there's a function called is_root in $(abs_top_srcdir)/testcases/lib/cmdlib.sh [in the install tree -- it gets put in $(abs_top_builddir)/testcases/bin] that can be leveraged s.t. you don't have to do $(id -ru) -eq 0 everywhere. There are some other handy constructs and functions too that can be used for expediting setup, execution, and teardown. The only thing that might be a detractor for some legacy scripts is that I explicitly did set -u to avoid QA issues with unset variables, so all variables must be defined before including the cmd library.
Using that script though would be helpful as it cuts down on a lot of duplicate code and aims to be completely POSIX compliant. Some other pluses:
1. incr_tst_count - increments $TST_COUNT appropriately.
2. TCID is automatically set if not already provided.
3. exists is a function which ensures that commands exist on a given system before executing a test.
4. Common SHELL_DEBUG logic.
Just trying to make life a little easier for test writers and maintainers if possible :]...
Cheers,
-Garrett
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Quoted 82 lines · Garrett Cooper
Quoted 56 lines · Subrata Modak
Quoted 55 lines · Jiří Paleček
Quoted 10 lines · Subrata Modak
Quoted 9 lines · JiříPaleček
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes
are
Quoted 6 lines
Quoted 6 lines
more or
less the same as in the previousQuoted 1 line
patches, except for a few exceptions). Note that the patch is not
complete,
in the sense that there mayQuoted 1 line
remain further bashisms in the source even after applying the
patch
Quoted 33 lines
Quoted 9 lines
(like use
of arrays, which is visibleQuoted 7 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>Quoted 8 lines
Quoted 1 line
Hmm,. Some of them failed to apply. Can you please resend only the
error
Quoted 2 lines
part(s):
according to my git repository. these are the missing parts:
Signed-off-by Jiri Palecek <[email protected]>Then something wrong with my CVS:
patching file
testcases/kernel/fs/fs-bench/modaltr.shHunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching filetestcases/kernel/fs/mongo/test.shHunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01Can you please verify ?
Quoted 8 lines
Quoted 3 lines
It works for me in CVS, so I guess it was just a linewrap, whitespace or
something issue. See attachment for the original (uncrippled) patch.Yes, it works fine now :-)
patching filetestcases/kernel/fs/fs-bench/modaltr.shpatching filetestcases/kernel/fs/mongo/test.shpatching file testcases/network/tcp_cmds/netstat/netstat01Sorry for replying late again, but just for future reference, there's a function called
is_rootin $(abs_top_srcdir)/testcases/lib/cmdlib.sh [in the install tree -- it gets put in $(abs_top_builddir)/testcases/bin] that can be leveraged s.t. you don't have to do $(id -ru) -eq 0 everywhere. There are some other handy constructs and functions too that can be used for expediting setup, execution, and teardown. The only thing that might be a detractor for some legacy scripts is that I explicitly did set -u to avoid QA issues with unset variables, so all variables must be defined before including the cmd library.Using that script though would be helpful as it cuts down on a lot of duplicate code and aims to be completely POSIX compliant. Some other pluses:
Quoted 8 lines
1.
incr_tst_count- increments $TST_COUNT appropriately.
2. TCID is automatically set if not already provided.
3. exists is a function which ensures that commands exist on a given system before executing a test.
4. CommonSHELL_DEBUGlogic.Just trying to make life a little easier for test writers and maintainers if possible :]...
Thanks for noting it. I think, however, that it would be much morer valuable to create exact duplicate of the C test API (or a subset of it) for shell scripts, as this would permit to remove much of the hacky stuff about maintaining TST_COUNT and return value from the shell scripts.
Looking at the file you mentioned, I have just a few remarks:
tst_cleanup()
{
# Disable the trap EXIT handler.
trap '' EXIT
# To ensure set -u passes...
TCtmp=${TCtmp:=}
Why? Shouldn't it catch invalid use (eg. without doing tst_setup first)?
# Nuke the testcase temporary directory if it exists.
[ -d "$TCtmp" ] && rm -rf "$TCtmp" (1)
}
tst_setup() {
TST_COUNT=1
TST_TOTAL=${TST_TOTAL:=1} (2)
export TCID TST_COUNT TST_TOTAL
for varname in TST_TOTAL; do
if eval "test -z \"\$${varname}\""; then
end_testcase "You must set ${varname} before calling setup()."
fi
doneWhat is the sense of doing "test -z "$TST_TOTAL"" in such an elaborate way, especially when you reset TST_TOTAL in (2) so test -z $TST_TOTAL can never be true here?
...
TCtmp=${TCtmp:=$TEMPDIR/$TC$$}
# User wants a temporary sandbox to play with.
if [ -n "$TCtmp" -a "$TCtmp" != "$TEMPDIR/$$" ] ; then
test -d "$TCtmp" || mkdir -p "$TCtmp"What is so special about "$TEMPDIR/$$" that you refuse to create a directory with this name, but still delete it in (1)?
...
}
incr_tst_count()
{
: $(( TST_COUNT + 1 ))
}Was this really meant this way (ie. as a code with no effect), or should it be "+="?
Regards
Jiri Palecek
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
Hi Jiri!
2009/10/31 Jiří Paleček <[email protected]>:
Quoted 90 lines
Quoted 87 lines · Garrett Cooper
Quoted 56 lines · Subrata Modak
Quoted 54 lines · Jiří Paleček
Quoted 49 lines · Subrata Modak
Quoted 24 lines · JiříPaleček
Hello,
Quoted 5 lines · Subrata Modak
Quoted 3 lines · Jiri Palecek
Hello,
this is another patch fixing bashisms in LTP tests (the fixes are
more or
less the same as in the previousQuoted 1 line
patches, except for a few exceptions). Note that the patch is not
complete,
in the sense that there mayQuoted 1 line
remain further bashisms in the source even after applying the patch
(like use
of arrays, which is visibleQuoted 7 lines
even from this patch).
Quoted 5 lines
Regards
Jiri PalecekSigned-off-by Jiri Palecek <jpalecek@...>Quoted 8 lines
Quoted 1 line
Hmm,. Some of them failed to apply. Can you please resend only the
error
Quoted 2 lines
part(s):
according to my git repository. these are the missing parts:
Signed-off-by Jiri Palecek <[email protected]>Then something wrong with my CVS:
patching file
testcases/kernel/fs/fs-bench/modaltr.shHunk #1 FAILED at 43.
1 out of 1 hunk FAILED -- saving rejects to file testcases/kernel/fs/fs-bench/modaltr.sh.rej patching filetestcases/kernel/fs/mongo/test.shHunk #1 FAILED at 26.
Hunk #2 FAILED at 52.
Hunk #3 FAILED at 70.
3 out of 3 hunks FAILED -- saving rejects to file testcases/kernel/fs/mongo/test.sh.rej
patching file testcases/network/tcp_cmds/netstat/netstat01Can you please verify ?
Quoted 7 lines
Quoted 2 lines
It works for me in CVS, so I guess it was just a linewrap, whitespace or something issue. See attachment for the original (uncrippled) patch.
Yes, it works fine now :-)
patching filetestcases/kernel/fs/fs-bench/modaltr.shpatching filetestcases/kernel/fs/mongo/test.shpatching file testcases/network/tcp_cmds/netstat/netstat01Sorry for replying late again, but just for future reference, there's a function called
is_rootin $(abs_top_srcdir)/testcases/lib/cmdlib.sh [in the install tree -- it gets put in $(abs_top_builddir)/testcases/bin] that can be leveraged s.t. you don't have to do $(id -ru) -eq 0 everywhere. There are some other handy constructs and functions too that can be used for expediting setup, execution, and teardown. The only thing that might be a detractor for some legacy scripts is that I explicitly did set -u to avoid QA issues with unset variables, so all variables must be defined before including the cmd library.Using that script though would be helpful as it cuts down on a lot of duplicate code and aims to be completely POSIX compliant. Some other pluses:
1.
incr_tst_count- increments $TST_COUNT appropriately.
2. TCID is automatically set if not already provided.
3. exists is a function which ensures that commands exist on a given system before executing a test.
4. CommonSHELL_DEBUGlogic.
Quoted 7 lines
Quoted 2 lines
Just trying to make life a little easier for test writers and maintainers if possible :]...
Thanks for noting it. I think, however, that it would be much more valuable to create exact duplicate of the C test API (or a subset of it) for shell scripts, as this would permit to remove much of the hacky stuff about maintaining
TST_COUNTand return value from the shell scripts.
Ok. I'll definitely take that into consideration, but it's up to the test writer to determine how many `subtestcases' are being executed in a master testcase.
Quoted 10 lines
Looking at the file you mentioned, I have just a few remarks:
tst_cleanup()
{
# Disable the trap EXIT handler.
trap '' EXIT
# To ensure set -u passes...
TCtmp=${TCtmp:=}Why? Shouldn't it catch invalid use (eg. without doing
tst_setupfirst)?
Yeah, that's useless if TCtmp is exported in tst_setup and tst_setup is executed to completion (or at least this variable is exported there) first. I may need to adjust the trap(1) to get this to function in a better way.
Quoted 16 lines
# Nuke the testcase temporary directory if it exists. [ -d "$TCtmp" ] && rm -rf "$TCtmp" (1) }
tst_setup(){TST_COUNT=1
TST_TOTAL=${TST_TOTAL:=1} (2) export TCIDTST_COUNTTST_TOTALfor varname in
TST_TOTAL; do
if eval "test -z \"\$${varname}\""; then
end_testcase "You must set ${varname} before callingsetup()." fi
done
Quoted 3 lines
What is the sense of doing "test -z "$TST_TOTAL"" in such an elaborate way, especially when you reset
TST_TOTALin (2) so test -z $TST_TOTAL can never be true here?
I was trying to avoid duplicate logic, but if it makes sense to do:
if test "x$foo" = x ; then
fi
if test "x$foo" = x ; then
fi
and do some basic arithmetic on the values to ensure they're in range, then I'll gladly rewrite it :]. The point was to avoid the nasty `you need to defined variable X.Y.Z' message from tst_res*(3), etc, because this is quick and resolves coding issues with the underlying testcases.
Quoted 9 lines
...
TCtmp=${TCtmp:=$TEMPDIR/$TC$$}
# User wants a temporary sandbox to play with. if [ -n "$TCtmp" -a "$TCtmp" != "$TEMPDIR/$$" ] ; then test -d "$TCtmp" || mkdir -p "$TCtmp"What is so special about "$TEMPDIR/$$" that you refuse to create a directory with this name, but still delete it in (1)?
Good point; this is not very well thought-out logic that I implemented on impulse a while ago (I think that the original logic test was `"$TCtmp" != "$TEMPDIR"').
Quoted 10 lines
...
}
incr_tst_count()
{
: $((TST_COUNT+ 1 ))
}Was this really meant this way (ie. as a code with no effect), or should it be "+="?
That's a really stupid mistake that I made. Fixed now (thanks!).
Thanks for the review!
-Garrett
------------------------------------------------------------------------------
Come build with us! The BlackBerry(R) Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9 - 12, 2009. Register now!
http://p.sf.net/sfu/devconference
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list