incubator-cassandra-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jonathan Ellis <jbel...@gmail.com>
Subject Re: Random Checkins
Date Thu, 02 Apr 2009 23:13:24 GMT
If you absolutely cannot afford regressions you should be working in a
branch.  If you want to run the latest & greatest you need to be
prepared to report bugs.  Nobody is immune to this, that is part of
running trunk.  This is why I created the cassandra-0.2 release branch
for you when I started committing: I suspected you might need a stable
branch.

We in the community cannot run tests that are not checked in.  You
have said that you test by directing some traffic to the new code, and
that's fine, but blaming us for breaking tests that we cannot run is
not reasonable.

I have to disagree that each checkin to trunk should be
production-tested.  That is not what trunk is for.  Trunk is for rapid
development.  I'm doing my best to build a test suite that can run
with ant test.  If we all add tests that catch regressions as we come
across them this will progress rapidly.  If it is just me it will
progress slower. :)

-Jonathan

[replying to both Avinash and Prashant.]

On Thu, Apr 2, 2009 at 4:52 PM, Avinash Lakshman
<avinash.lakshman@gmail.com> wrote:
> Compactions etc are breaking now in our tests. I am going to revert all new
> commits that have gone in. Are you sure these are tested changes? We need to
> have a review process for changing code that is functioning. This is very
> very frustrating.
>
> Avinash
>

On Thu, Apr 2, 2009 at 4:57 PM, Prashant Malik <pmalik@gmail.com> wrote:
> Hey Jonathan ,
>
>  Are you testing these changes ?
>  If we are going to make changes like this without much of a reason we
> better test it as this code needs to run in
>  our production environment.
>  You need to test the changes at a reasonable scale.
>
> - Prashant

Mime
View raw message