* [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled
@ 2014-08-04 18:38 Ramsay Jones
2014-09-27 4:29 ` Christopher Li
0 siblings, 1 reply; 6+ messages in thread
From: Ramsay Jones @ 2014-08-04 18:38 UTC (permalink / raw)
To: Christopher Li; +Cc: Sparse Mailing-list
Signed-off-by: Ramsay Jones <ramsay@ramsay1.demon.co.uk>
---
validation/test-suite | 25 +++++++++++++++++++++++--
1 file changed, 23 insertions(+), 2 deletions(-)
diff --git a/validation/test-suite b/validation/test-suite
index 3c011c6..0142701 100755
--- a/validation/test-suite
+++ b/validation/test-suite
@@ -7,12 +7,18 @@ default_cmd="sparse \$file"
tests_list=`find . -name '*.c' | sed -e 's#^\./\(.*\)#\1#' | sort`
prog_name=`basename $0`
+if [ ! -x "$default_path/sparse-llvm" ]; then
+ disabled_cmds="sparsec sparsei sparse-llvm"
+fi
+
# counts:
# - tests that have not been converted to test-suite format
+# - tests that are disabled
# - tests that passed
# - tests that failed
# - tests that failed but are known to fail
unhandled_tests=0
+disabled_tests=0
ok_tests=0
ko_tests=0
known_ko_tests=0
@@ -80,6 +86,7 @@ echo " help prints usage"
# - 0 if the test passed,
# - 1 if it failed,
# - 2 if it is not a "test-suite" test.
+# - 3 if the test is disabled.
do_test()
{
test_failed=0
@@ -95,8 +102,6 @@ do_test()
fi
test_name=$last_result
- echo " TEST $test_name ($file)"
-
# does the test provide a specific command ?
cmd=`eval echo $default_path/$default_cmd`
get_value "check-command" $file
@@ -104,6 +109,19 @@ do_test()
last_result=`echo $last_result | sed -e 's/^ *//'`
cmd=`eval echo $default_path/$last_result`
fi
+
+ # check for disabled commands
+ for i in $disabled_cmds; do
+ case "$cmd" in
+ *$i*)
+ disabled_tests=`expr $disabled_tests + 1`
+ return 3
+ ;;
+ esac
+ done
+
+ echo " TEST $test_name ($file)"
+
verbose "Using command : $cmd"
# grab the expected exit value
@@ -168,6 +186,9 @@ do_test_suite()
if [ "$unhandled_tests" -ne "0" ]; then
echo "$unhandled_tests tests could not be handled by $prog_name"
fi
+ if [ "$disabled_tests" -ne "0" ]; then
+ echo "$disabled_tests tests were disabled"
+ fi
}
##
--
2.0.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled
2014-08-04 18:38 [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled Ramsay Jones
@ 2014-09-27 4:29 ` Christopher Li
2014-09-27 13:37 ` Ramsay Jones
0 siblings, 1 reply; 6+ messages in thread
From: Christopher Li @ 2014-09-27 4:29 UTC (permalink / raw)
To: Ramsay Jones; +Cc: Sparse Mailing-list
On Tue, Aug 5, 2014 at 2:38 AM, Ramsay Jones <ramsay@ramsay1.demon.co.uk> wrote:
> +if [ ! -x "$default_path/sparse-llvm" ]; then
> + disabled_cmds="sparsec sparsei sparse-llvm"
> +fi
I think this can be simplified as if the test command was not found,
then that test will be disabled. It doesn't need to be sparse llvm, it
can be other
test program.
We don't need to set an disable_cmds list then check against that list.
Chris
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled
2014-09-27 4:29 ` Christopher Li
@ 2014-09-27 13:37 ` Ramsay Jones
2014-09-30 3:26 ` Christopher Li
0 siblings, 1 reply; 6+ messages in thread
From: Ramsay Jones @ 2014-09-27 13:37 UTC (permalink / raw)
To: Christopher Li; +Cc: Sparse Mailing-list
On 27/09/14 05:29, Christopher Li wrote:
> On Tue, Aug 5, 2014 at 2:38 AM, Ramsay Jones <ramsay@ramsay1.demon.co.uk> wrote:
>> +if [ ! -x "$default_path/sparse-llvm" ]; then
>> + disabled_cmds="sparsec sparsei sparse-llvm"
>> +fi
>
> I think this can be simplified as if the test command was not found,
> then that test will be disabled. It doesn't need to be sparse llvm, it
> can be other
> test program.
Unfortunately, sparcec and sparsei are the commands that are
used by the tests (I included sparse-llvm in the list in case
any _future_ tests used it directly), and they do exist and
are executable. They simply will not work if sparse-llvm is
not built.
I did think about removing the x-bit on sparse{c,i} in git,
enabling it again in the worktree if sparse-llvm was built,
but I was concerned that would not work on cygwin and it
would possibly complicate installation. ie I didn't try it
out seriously. This patch seemed the best solution to me.
[Another idea I didn't try: name the sparse{c,i} scripts
sparse{c,i}-in and rename them as part of a successful
build of sparse-llvm]
BTW, the 'validation/backend/sum.c' fails for me on cygwin
(64-bit), but I think it is actually a problem with the
llvm interpreter lli. (it works fine on my old 32-bit Linux
laptop).
ATB,
Ramsay Jones
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled
2014-09-27 13:37 ` Ramsay Jones
@ 2014-09-30 3:26 ` Christopher Li
2014-10-02 3:44 ` Christopher Li
0 siblings, 1 reply; 6+ messages in thread
From: Christopher Li @ 2014-09-30 3:26 UTC (permalink / raw)
To: Ramsay Jones; +Cc: Sparse Mailing-list
On Sat, Sep 27, 2014 at 9:37 PM, Ramsay Jones
<ramsay@ramsay1.demon.co.uk> wrote:
> On 27/09/14 05:29, Christopher Li wrote:
>
> Unfortunately, sparcec and sparsei are the commands that are
> used by the tests (I included sparse-llvm in the list in case
> any _future_ tests used it directly), and they do exist and
> are executable. They simply will not work if sparse-llvm is
> not built.
>
I see. That make sense.
We just need a way to specify the dependency of test commands.
e.g. sparsec depends on sparse-llvm. So if sparsec does not exist,
or sparse-llvm does not exist, the test will be skipped.
I am thinking some thing in the test-suite like:
set_deps sparsec sparse-llvm
set_deps sparsei sparse-llvm
...
get_deps `basename cmd`
dep_cmd=$last_result
> Note that I renamed sparse{c,i} to sparse{c,i}-in. It may have been
> better to name then sparse{c,i}.sh, since they are not really templates
> to be 'processed' to create a build product. dunno.
>
I prefer specify the dependency. Another copy is not necessary.
Chris
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled
2014-09-30 3:26 ` Christopher Li
@ 2014-10-02 3:44 ` Christopher Li
2014-10-02 10:31 ` Ramsay Jones
0 siblings, 1 reply; 6+ messages in thread
From: Christopher Li @ 2014-10-02 3:44 UTC (permalink / raw)
To: Ramsay Jones; +Cc: Sparse Mailing-list
On Tue, Sep 30, 2014 at 11:26 AM, Christopher Li <sparse@chrisli.org> wrote:
> I am thinking some thing in the test-suite like:
>
> set_deps sparsec sparse-llvm
> set_deps sparsei sparse-llvm
>
> ...
> get_deps `basename cmd`
> dep_cmd=$last_result
Actually, I try that, it is a little bit too complicated.
Your first approach is fine.
I end up using your V1 patch, with a small change to do exact match rather than
a wild cast. The wild cast has danger that match a sub word of the command.
e.g. if you try to disable command "sp", it will disable "sparse" as well.
Chris
diff --git a/validation/test-suite b/validation/test-suite
index 1c94aa8..784ff6b 100755
--- a/validation/test-suite
+++ b/validation/test-suite
@@ -111,13 +111,13 @@ do_test()
fi
# check for disabled commands
+ set -- $cmd
+ base_cmd=`basename $1`
for i in $disabled_cmds; do
- case "$cmd" in
- *$i*)
+ if [ "$i" == "$base_cmd" ] ; then
disabled_tests=`expr $disabled_tests + 1`
return 3
- ;;
- esac
+ fi
done
echo " TEST $test_name ($file)"
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled
2014-10-02 3:44 ` Christopher Li
@ 2014-10-02 10:31 ` Ramsay Jones
0 siblings, 0 replies; 6+ messages in thread
From: Ramsay Jones @ 2014-10-02 10:31 UTC (permalink / raw)
To: Christopher Li; +Cc: Sparse Mailing-list
On 02/10/14 04:44, Christopher Li wrote:
> On Tue, Sep 30, 2014 at 11:26 AM, Christopher Li <sparse@chrisli.org> wrote:
>> I am thinking some thing in the test-suite like:
>>
>> set_deps sparsec sparse-llvm
>> set_deps sparsei sparse-llvm
>>
>> ...
>> get_deps `basename cmd`
>> dep_cmd=$last_result
>
> Actually, I try that, it is a little bit too complicated.
>
> Your first approach is fine.
>
> I end up using your V1 patch, with a small change to do exact match rather than
> a wild cast. The wild cast has danger that match a sub word of the command.
> e.g. if you try to disable command "sp", it will disable "sparse" as well.
Ah yes, good catch!
ATB,
Ramsay Jones
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-10-02 10:31 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-08-04 18:38 [PATCH 06/10] don't run sparse{c,i} tests when sparse-llvm is disabled Ramsay Jones
2014-09-27 4:29 ` Christopher Li
2014-09-27 13:37 ` Ramsay Jones
2014-09-30 3:26 ` Christopher Li
2014-10-02 3:44 ` Christopher Li
2014-10-02 10:31 ` Ramsay Jones
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).