Date: Fri, 18 Aug 2000 17:29:52 +0800 (+0800) From: Michael Robinson <robinson@netrinsics.com> To: nate@yogotech.com Cc: freebsd-java@freebsd.org Subject: RE: State of Server-Side Java Message-ID: <200008180929.RAA51600@netrinsics.com> In-Reply-To: <200008180640.AAA25895@nomad.yogotech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
>> This may go without saying, but it's probably a good idea to do as much
>> as possible with typescript running. For posterity.
>
>Given the fact that I'm not allowed to share how the tests work or my
>results (other than YES/NO) or how it all works due to licensing, I'm
>not sure why...
I envision a sequence of events that goes something like this:
1. Figure out how to get test t[n] to work.
2. Test t[n] fails.
3. Repeat steps 1, 2 for t[n], t[m], t[o] ...
4. Identify bug causing failure of t[n], t[m], t[o] ...
5. Some time later, Greg fixes bug.
Now, my guess is, with multiple thousands of tests, at this point, you'll
either a) repeat step 1, or b) consult your typescripts. Of course, there
are better forms of documentation than typescripts, but none as painless.
6. Test n passes.
7. Some time later, Greg adds, e.g., native kernel threads.
Now, at this point it would be really nice to have a bunch of shell scripts
that run the whole suite automatically (sort of like "make world"), so that
every time someone touches something, you don't have to start at step 1
again. That is to say, a regression test. Typescripts are an excellent
starting point for developing these kinds of scripts.
Of course, the big fly in the ointment is all the AWT/Swing stuff that
probably can't be scripted easily. Your license probably doesn't even allow
you to say what percent of the tests require human interaction.
-Michael Robinson
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-java" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200008180929.RAA51600>