Frank Schönheit - Sun Microsystems Germany wrote:
Hi Christian,
Hi Frank,
That's in fact a problem. Those tinderbox stati in EIS have been
introduced without extensive explanation.
Wrong. It has been announced here and on at least the tools-ML.
http://www.mail-archive.com/[email protected]/msg03936.html
Okay, in this mail was more than I remembered. It could have been
phrased more explicitly ("ignore the error counts in all cases, only
look for colors"), but the idea to look for "red" was in fact there.
What's still missing in this announcement is the explanation of the
three items. What's in there is "if it's red, it's bad, and can have
three causes". What's *not* in there - and what is *needed* to actually
work with this data - is how to differ those three possible causes. I'm
am completely alone in deciding this.
In case tinderbox status red and CWS is ready for QA the QA Rep should
get in contact with the owner.
In case the owner or some other developer member of the CWS believes
that the state of the CWS is that it should be buildable (which it
should be when it´s already being set to ready for qa at least) and the
tinderbox status is red he/she should follow the link to the tinderbox
that is above the status messages in EIS. He/she than would have to
analyse the LOG file that is available in tinderbox (Klick on one of the
"l" "L" or "C" letters and a popup will open that
provides links to the summary log and the full log.) to check wether it
was a network problem with collab.net or a real build problem. If you
find cvs checkout errors etc. it was not a build problem for example. In
case of a build problem the developer would than have to compare with
the tinderbox status of the MWS the ChildWorkspace is based on (
navigate to that via EIS/Misc/Tinderbox MainPage and than look for the
master ) to see wether the problem was introduced by his/her CWS or
already on the master it is based on.
Following the link above the status messages once also gives an
explanation about what the other colors do mean.
[...]
Kind regards,
Bernd Eilers
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
