[Insight-developers] dashboard tag

[Blezek, Daniel J (CRD)]

My take on this is to remove the human from the loop.  The consortium decides on a metric, and when
the metric is met, we tag, but not before.  At a minimum, all tests running and passing should be the
start, then we add the criteria of errors/warnings/coverage/purify as needed.

It's very easy to tag the source each night.  In fact, this would make the building of a nightly
much, much easier from the testing framework point of view.

-dan

-----Original Message-----
From: Bill Hoffman [mailto:bill.hoffman@kitware.com]
Sent: Friday, May 18, 2001 2:35 PM
To: insight-Developers
Subject: [Insight-developers] dashboard tag


I was talking with Will and Brad about this tag for the "Last Known Good Build" and
we came up with some ideas.

1. In the testing stuff, use a moving tag on the entire source tree:
NIGHTLY_DASHBOARD_TAG
This is moved forward on all files each night.

2. Have another tag LAST_KNOWN_STABLE_BUILD, this is on every file
in the system.   It is set by hand.   When someone (Bill L. ?) blesses a
dashboard build, the LAST_KNOWN_STABLE_BUILD is moved to the
NIGHTLY_DASHBOARD_TAG on all files.


This means that if you are running nightly builds, you will only
have to do a cvs update -d ( after an initial cvs update -r NIGHTLY_DASHBOARD_TAG ).
The files will move forward each night.

If you want a stable build, you can checkout Insight with the
LAST_KNOWN_STABLE_BUILD tag the first time, and assume that you can
do an update at any point, and it will build clean.


-Bill


_______________________________________________
Insight-developers mailing list
Insight-developers@public.kitware.com
http://public.kitware.com/mailman/listinfo/insight-developers