[DTrace-devel] [PATCH 17/17] usdt: testsuite updates

[Nick Alcock]

On 1 Sep 2022, Eugene Loh via DTrace-devel said:

> I think it'd be better to combine removal of xfail annotations with
> the patches in which the tests start to pass.  That way, you can see
> which test goes with which patch.  Also...

I tried that, but that means they all end up in the preceding "usdt:
DTrace userspace side" commit, which is less pretty than what we have
now. (If they started passing at a wide variety of points, you would be
right.)

> On 8/10/22 18:07, Nick Alcock via DTrace-devel wrote:
>> We also mark three pid tests as broken which need raise().  I have no
>> idea how they ever worked, but I don't think their failing now is a sign
>> of a problem in USDT.
>
> The pid tests should work.

Yeah. Definitely a regression. No idea what's wrong yet, only a big list
of things that aren't the cause :/

>> We add one new test to make sure that you can run a probe-containing
>> program and then repeatedly probe it from multiple dtraces in turn
>> without them messing up each others' probes in any way.  (Doing it from
>> multiple dtraces in *parallel* may well not work yet: no specific code
>> has been written to make it work: that's planned for the next phase.)
>
> Might as well phrase that paragraph in such a way that the word "lingering" is used to tie it to the new test.  Not a big deal, I
> know.

Agreed.

>> diff --git a/test/unittest/usdt/tst.entryreturn.sh b/test/unittest/usdt/tst.entryreturn.sh
>> +# @@xfail: dtv2, needs system()
>> -# @@xfail: dtv2
>
> No:  system() works just fine, so far as I can tell.

Oh! Then I bet this is the same failure that is crippling the pid args
tests -- tst.entryreturn.sh has multiple probes in it... and oh look
none are firing.

Removed the @@xfail for the next round, so this is failing as it should.

On 1 Sep 2022, Eugene Loh via DTrace-devel spake thusly:

> And...
>
> Should files have their copyright years updated when touched?

Yes. I repeatedly forget to do that. I'll go through them all (uggggh)
and do so. (Anyone got a script...)

> On 8/10/22 18:07, Nick Alcock via DTrace-devel wrote:
>> diff --git a/test/unittest/usdt/tst.fork.sh b/test/unittest/usdt/tst.lingering.sh
>>   if [ $# != 1 ]; then
>>   	echo expected one argument: '<'dtrace-path'>'
>>   	exit 2
>>   fi
>
> I think this $#!=1 check has been deprecated.  Kris says that if this is going to fail, there will be widespread outages anyhow.  So
> no point in checking all the time.  (Yeah, inherited from tst.fork.sh.)

Added a new commit that drops it from the whole of usdt/ and pid/.

>> diff --git a/test/unittest/usdt/tst.reeval.sh b/test/unittest/usdt/tst.reeval.sh
>> -# @@xfail: dtv2
>> +# @@xfail: dtv2, no args on usdt probes yet
>
> The message looks wrong.

Oh right, yeah, this is failing because it needs sdt (proc).  Adjusted
the message.

-- 
NULL && (void)