Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 4 Oct 2005 11:44:57 +0100 (BST)
From:      Robert Watson <rwatson@FreeBSD.org>
To:        Mike Silbersack <silby@silby.com>
Cc:        cvs-src@FreeBSD.org, Don Lewis <truckman@FreeBSD.org>, src-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/sys/ufs/ffs ffs_alloc.c
Message-ID:  <20051004114119.G69774@fledge.watson.org>
In-Reply-To: <20051004032213.R11746@odysseus.silby.com>
References:  <200510040452.j944q0rI020011@gw.catspoiler.org> <20051004032213.R11746@odysseus.silby.com>

next in thread | previous in thread | raw e-mail | index | archive | help

On Tue, 4 Oct 2005, Mike Silbersack wrote:

> I've cc'd rwatson, since he seems to be the regression test king, and 
> may have some input on this idea.

King, eh? :-)

There's definitely a niche here for someone who wants to pick up where Nik 
left off integrating the regression test suite a bit more with a unified 
run and reporting system.  And for someone to take ownership of the system 
in terms of running it regularly to actually detect regressions, generate 
reports, and so on.  While I've been adding a lot of regression tests, I 
know relatively little about testing, and so while the tests work, they 
probably leave quite a bit to be desired.  For example, they tend to fail 
entirely on first detected problem, whereas it might be desirable to run 
more like configure and test various elements and report on each 
independently, so that tests aren't skipped when a first failure is hit.

Also, I'd love to see some instructions somewhere about how it would be 
best to format output for a test, and possible a libtest(3) to link 
against in order to get formatted output routines and automated cleanup 
routines for when tests end (i.e., please rm -Rf /tmp/testdir.XXXXXX, even 
if the test process dies).  Finally, a number of tests require similar 
pieces of infrastructure: a UFS file system image, or a configured network 
interface with a local IP address, etc.  Things that could be provided by 
a framework in a modular way, rather than reproducing them in every 
regression test.

Robert N M Watson



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051004114119.G69774>