* [LTP] [PATCH v3] runltp: add -R option for randomize test order
@ 2016-09-24 12:35 Zorro Lang
2016-09-25 9:38 ` Li Wang
0 siblings, 1 reply; 4+ messages in thread
From: Zorro Lang @ 2016-09-24 12:35 UTC (permalink / raw)
To: ltp
Generally we write test entries into runtests/${testfile}, then
runltp reads and runs lines sequentially everytime, nearly no one
will change the test order.
Recently we find bugs by change the test order,different test
combination maybe find more bugs. So I add this new option -R to
randomize test order.
Signed-off-by: Zorro Lang <zlang@redhat.com>
---
V2 use 'sort -R' to instead of 'randlines.awk'.
V3 remove change log and someone useless comment
Thanks,
Zorro
runltp | 11 ++++++++++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git a/runltp b/runltp
index bb2e5d8..33e23ab 100755
--- a/runltp
+++ b/runltp
@@ -153,6 +153,7 @@ usage()
-p Human readable format logfiles.
-q Print less verbose output to screen.
-r LTPROOT Fully qualified path where testsuite is installed.
+ -R Randomize test order.
-s PATTERN Only run test cases which match PATTERN.
-S SKIPFILE Skip tests specified in SKIPFILE
-t DURATION Execute the testsuite for given duration. Examples:
@@ -213,12 +214,13 @@ main()
local EMAIL_TO=""
local TAG_RESTRICT_STRING=""
local PAN_COMMAND=""
+ local RANDOMRUN=0
local DEFAULT_FILE_NAME_GENERATION_TIME=`date +"%Y_%m_%d-%Hh_%Mm_%Ss"`
local scenfile=
version_date=$(cat "$LTPROOT/Version")
- while getopts a:c:C:T:d:D:f:F:ehi:I:K:g:l:m:M:Nno:pqr:s:S:t:T:w:x:b:B:z:Z: arg
+ while getopts a:c:C:T:d:D:f:F:ehi:I:K:g:l:m:M:Nno:pqr:Rs:S:t:T:w:x:b:B:z:Z: arg
do case $arg in
a) EMAIL_TO=$OPTARG
ALT_EMAIL_OUT=1;;
@@ -410,6 +412,8 @@ main()
r) LTPROOT=$OPTARG;;
+ R) RANDOMRUN=1;;
+
s) TAG_RESTRICT_STRING=$OPTARG;;
S) case $OPTARG in
@@ -740,6 +744,11 @@ main()
done
fi
+ if [ "$RANDOMRUN" != "0" ]; then
+ sort -R ${TMP}/alltests > ${TMP}/alltests.temp
+ cat ${TMP}/alltests.temp > ${TMP}/alltests
+ fi
+
[ ! -z "$QUIET_MODE" ] && { echo "INFO: Test start time: $(date)" ; }
PAN_COMMAND="${LTPROOT}/bin/ltp-pan $QUIET_MODE -e -S $INSTANCES $DURATION -a $$ \
-n $$ $PRETTY_PRT -f ${TMP}/alltests $LOGFILE $OUTPUTFILE $FAILCMDFILE $TCONFCMDFILE"
--
2.7.4
^ permalink raw reply related [flat|nested] 4+ messages in thread* [LTP] [PATCH v3] runltp: add -R option for randomize test order 2016-09-24 12:35 [LTP] [PATCH v3] runltp: add -R option for randomize test order Zorro Lang @ 2016-09-25 9:38 ` Li Wang 2016-09-25 13:47 ` Zorro Lang 0 siblings, 1 reply; 4+ messages in thread From: Li Wang @ 2016-09-25 9:38 UTC (permalink / raw) To: ltp On Sat, Sep 24, 2016 at 8:35 PM, Zorro Lang <zlang@redhat.com> wrote: > Generally we write test entries into runtests/${testfile}, then > runltp reads and runs lines sequentially everytime, nearly no one > will change the test order. > > Recently we find bugs by change the test order,different test > combination maybe find more bugs. So I add this new option -R to > randomize test order. > > Signed-off-by: Zorro Lang <zlang@redhat.com> > --- > > V2 use 'sort -R' to instead of 'randlines.awk'. > V3 remove change log and someone useless comment > > Thanks, > Zorro > > runltp | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) > > diff --git a/runltp b/runltp > index bb2e5d8..33e23ab 100755 > --- a/runltp > +++ b/runltp > @@ -153,6 +153,7 @@ usage() > -p Human readable format logfiles. > -q Print less verbose output to screen. > -r LTPROOT Fully qualified path where testsuite is installed. > + -R Randomize test order. > -s PATTERN Only run test cases which match PATTERN. > -S SKIPFILE Skip tests specified in SKIPFILE > -t DURATION Execute the testsuite for given duration. Examples: > @@ -213,12 +214,13 @@ main() > local EMAIL_TO="" > local TAG_RESTRICT_STRING="" > local PAN_COMMAND="" > + local RANDOMRUN=0 > local DEFAULT_FILE_NAME_GENERATION_TIME=`date +"%Y_%m_%d-%Hh_%Mm_%Ss"` > local scenfile= > > version_date=$(cat "$LTPROOT/Version") > > - while getopts a:c:C:T:d:D:f:F:ehi:I:K:g:l:m:M:Nno:pqr:s:S:t:T:w:x:b:B:z:Z: arg > + while getopts a:c:C:T:d:D:f:F:ehi:I:K:g:l:m:M:Nno:pqr:Rs:S:t:T:w:x:b:B:z:Z: arg > do case $arg in > a) EMAIL_TO=$OPTARG > ALT_EMAIL_OUT=1;; > @@ -410,6 +412,8 @@ main() > > r) LTPROOT=$OPTARG;; > > + R) RANDOMRUN=1;; > + > s) TAG_RESTRICT_STRING=$OPTARG;; > > S) case $OPTARG in > @@ -740,6 +744,11 @@ main() > done > fi > > + if [ "$RANDOMRUN" != "0" ]; then > + sort -R ${TMP}/alltests > ${TMP}/alltests.temp > + cat ${TMP}/alltests.temp > ${TMP}/alltests How about combine them in one line? shuf ${TMP}/alltests -o ${TMP}/alltests or, sort -R ${TMP}/alltests -o ${TMP}/alltests > + fi > + > [ ! -z "$QUIET_MODE" ] && { echo "INFO: Test start time: $(date)" ; } > PAN_COMMAND="${LTPROOT}/bin/ltp-pan $QUIET_MODE -e -S $INSTANCES $DURATION -a $$ \ > -n $$ $PRETTY_PRT -f ${TMP}/alltests $LOGFILE $OUTPUTFILE $FAILCMDFILE $TCONFCMDFILE" > -- > 2.7.4 > > > -- > Mailing list info: https://lists.linux.it/listinfo/ltp -- Regards, Li Wang Email: liwang@redhat.com ^ permalink raw reply [flat|nested] 4+ messages in thread
* [LTP] [PATCH v3] runltp: add -R option for randomize test order 2016-09-25 9:38 ` Li Wang @ 2016-09-25 13:47 ` Zorro Lang 2016-09-26 8:25 ` Cyril Hrubis 0 siblings, 1 reply; 4+ messages in thread From: Zorro Lang @ 2016-09-25 13:47 UTC (permalink / raw) To: ltp On Sun, Sep 25, 2016 at 05:38:07PM +0800, Li Wang wrote: > On Sat, Sep 24, 2016 at 8:35 PM, Zorro Lang <zlang@redhat.com> wrote: > > Generally we write test entries into runtests/${testfile}, then > > runltp reads and runs lines sequentially everytime, nearly no one > > will change the test order. > > > > Recently we find bugs by change the test order,different test > > combination maybe find more bugs. So I add this new option -R to > > randomize test order. > > > > Signed-off-by: Zorro Lang <zlang@redhat.com> > > --- > > > > V2 use 'sort -R' to instead of 'randlines.awk'. > > V3 remove change log and someone useless comment > > > > Thanks, > > Zorro > > > > runltp | 11 ++++++++++- > > 1 file changed, 10 insertions(+), 1 deletion(-) > > > > diff --git a/runltp b/runltp > > index bb2e5d8..33e23ab 100755 > > --- a/runltp > > +++ b/runltp > > @@ -153,6 +153,7 @@ usage() > > -p Human readable format logfiles. > > -q Print less verbose output to screen. > > -r LTPROOT Fully qualified path where testsuite is installed. > > + -R Randomize test order. > > -s PATTERN Only run test cases which match PATTERN. > > -S SKIPFILE Skip tests specified in SKIPFILE > > -t DURATION Execute the testsuite for given duration. Examples: > > @@ -213,12 +214,13 @@ main() > > local EMAIL_TO="" > > local TAG_RESTRICT_STRING="" > > local PAN_COMMAND="" > > + local RANDOMRUN=0 > > local DEFAULT_FILE_NAME_GENERATION_TIME=`date +"%Y_%m_%d-%Hh_%Mm_%Ss"` > > local scenfile= > > > > version_date=$(cat "$LTPROOT/Version") > > > > - while getopts a:c:C:T:d:D:f:F:ehi:I:K:g:l:m:M:Nno:pqr:s:S:t:T:w:x:b:B:z:Z: arg > > + while getopts a:c:C:T:d:D:f:F:ehi:I:K:g:l:m:M:Nno:pqr:Rs:S:t:T:w:x:b:B:z:Z: arg > > do case $arg in > > a) EMAIL_TO=$OPTARG > > ALT_EMAIL_OUT=1;; > > @@ -410,6 +412,8 @@ main() > > > > r) LTPROOT=$OPTARG;; > > > > + R) RANDOMRUN=1;; > > + > > s) TAG_RESTRICT_STRING=$OPTARG;; > > > > S) case $OPTARG in > > @@ -740,6 +744,11 @@ main() > > done > > fi > > > > + if [ "$RANDOMRUN" != "0" ]; then > > + sort -R ${TMP}/alltests > ${TMP}/alltests.temp > > + cat ${TMP}/alltests.temp > ${TMP}/alltests > > How about combine them in one line? > > shuf ${TMP}/alltests -o ${TMP}/alltests > or, > sort -R ${TMP}/alltests -o ${TMP}/alltests Thanks Li:) Hm, it works, even works for large file (>pagesize). I thought it maybe make mistake if read and rewrite same file at same time when I sent this patch. Now I just read the code of sort.c, it'll read the file into buffer and temp file at first, then close the fopen(r) file and go to fopen(w) the file: if (buf.eof && !nfiles && !ntemps && !buf.left) { xfclose (fp, file); tfp = xfopen (output_file, "w"); ... ... So it looks safe to read and write the same file together ... But I don't know if it worth changing this patch (the old way looks safer:). Can maintainers choose a favorite way? Thanks, Zorro > > > + fi > > + > > [ ! -z "$QUIET_MODE" ] && { echo "INFO: Test start time: $(date)" ; } > > PAN_COMMAND="${LTPROOT}/bin/ltp-pan $QUIET_MODE -e -S $INSTANCES $DURATION -a $$ \ > > -n $$ $PRETTY_PRT -f ${TMP}/alltests $LOGFILE $OUTPUTFILE $FAILCMDFILE $TCONFCMDFILE" > > -- > > 2.7.4 > > > > > > -- > > Mailing list info: https://lists.linux.it/listinfo/ltp > > > > -- > Regards, > Li Wang > Email: liwang@redhat.com ^ permalink raw reply [flat|nested] 4+ messages in thread
* [LTP] [PATCH v3] runltp: add -R option for randomize test order 2016-09-25 13:47 ` Zorro Lang @ 2016-09-26 8:25 ` Cyril Hrubis 0 siblings, 0 replies; 4+ messages in thread From: Cyril Hrubis @ 2016-09-26 8:25 UTC (permalink / raw) To: ltp Hi! > Hm, it works, even works for large file (>pagesize). I thought it maybe make > mistake if read and rewrite same file at same time when I sent this patch. > Now I just read the code of sort.c, it'll read the file into buffer and temp > file at first, then close the fopen(r) file and go to fopen(w) the file: > > if (buf.eof && !nfiles && !ntemps && !buf.left) > { > xfclose (fp, file); > tfp = xfopen (output_file, "w"); > ... > ... > > So it looks safe to read and write the same file together ... But I don't know > if it worth changing this patch (the old way looks safer:). Can maintainers > choose a favorite way? The POSIX specification for sort explicitly states: -o output Specify the name of an output file to be used instead of the standard output. This file can be the same as one of the input files. So I would say that using the same filename for input and output is pretty much safe. -- Cyril Hrubis chrubis@suse.cz ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2016-09-26 8:25 UTC | newest] Thread overview: 4+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2016-09-24 12:35 [LTP] [PATCH v3] runltp: add -R option for randomize test order Zorro Lang 2016-09-25 9:38 ` Li Wang 2016-09-25 13:47 ` Zorro Lang 2016-09-26 8:25 ` Cyril Hrubis
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox