From mboxrd@z Thu Jan 1 00:00:00 1970 From: Cyril Hrubis Date: Mon, 31 Aug 2020 19:49:48 +0200 Subject: [LTP] [PATCH] [RFC] runtest/quickhit: Remove. In-Reply-To: References: <20200831094617.7764-1-chrubis@suse.cz> Message-ID: <20200831174948.GA15731@yuki.lan> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ltp@lists.linux.it Hi! > Thinking about this a bit more.... > > Since this test is used (at least by me), more as a test of the testing > infrastructure rather than the hardware/software under test, > maybe it make sense to repurpose the test and adjust its contents > based on this new purpose. What would be nice is a test that > exercises a bunch of different possible LTP behaviors or outputs, > to test whether CI systems calling LTP can handle them all correctly. > > So, there are 2 attributes of the test that are important to me: > - it runs quickly (more quickly than a "full" ltp) > - it runs a variety of individual LTP test programs > > So it might be good to have this be a test that includes items that > behave strangely (but quickly). That would make this test more > useful for the purpose I'm actually using it for. > > It might even make sense to rename it to reflect this change of purpose > (if it *is* a change of purpose). For example, maybe name it > 'smoketest' or 'weirdstuff' or 'selftest'. But 'quickhit' at least captures > one attribute that is important - that this test is used as a quick > check that basic LTP functionality is working. > > Just some more ideas.... I would vote for removing this one and adding either a smoketest or selftest. The quickhit name is way too confusing. Also if we want something as a selftest we can also throw in a few test test library sanity tests that are not even installed at this point. -- Cyril Hrubis chrubis@suse.cz