gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefan Bodewig <bode...@apache.org>
Subject Re: [MO-java-dev] compilation problem for mockobjects
Date Wed, 04 Feb 2004 13:05:59 GMT
On Wed, 4 Feb 2004, Nat Pryce <nat.pryce@b13media.com> wrote:

> Considering what Gump is meant to be used for, would it be better
> for it to be "edge triggered" rather than "level triggered".

It depends on the community you want to reach.  Sometimes it takes
repeated nagging to get the message through.  Sometimes you generate
the opposite effect by nagging, i.e. mails get ignored.

> That is, should it send a message when a project fails to build, and
> then sends a message when the project builds successfully again,
> instead of sending a message every time it fails to build a broken
> project?

Yes, this would certainly be an option, and I think much of the
statistics sthe new Python implementation of Gump collects can be used
for that.  We also generate RSS feeds that contain exactly these types
of status changes IIRC.

The email based nagging system has been Gump's traditional approach
and it has done an outstanding job in many cases.  I agree that it is
plain annoying if it comes unwanted.

Maybe we can improve the nagging part in the Python implementation to
be configurable on a project level.  Some projects may want a nightly
report even on successful builds as confirmation, others may only want
to get notified on status changes (as you describe it).

Stefan

Mime
View raw message