[Boost-testing] New testers: Dont test on release?

From: Darryl Green (darryl.green_at_[hidden])
Date: 2008-07-29 19:28:11

On Dev List Beman wrote:
> Markus Schöpflin wrote:
>> You might want to set up your regression tester for trunk testing first. 
>> If you ironed out the wrinkles and all is working smoothly, contact the 
>> respective release manager before switching to release testing.
>
> Yes, exactly.

Oops. I have done one run on the release branch - GreenLenny (Debian Lenny GCC
4.3). Sorry for not contacting the release manager before doing this.

I had been following boost dev (but not test) - surely common for those not
(yet) doing testing. It seemed some broader test coverage on release was needed.
So I followed the simple and easy to find instructions to run a test on release
at http://www.boost.org/development/running_regression_tests.html but then...

On Dev List Beman wrote:
> Alp Mestan wrote:
>> Great, thanks. In the upcoming days I'll try to get a Gentoo environment to
>> test the latest branch.
>
> Please do any testing against trunk. We don't have the infrastructure to 
> add additional release tests yet, and in any case release testers should 
> only be added at the beginning of a release cycle.

So - lets get this right - boost doesn't want the release branch tested on more
platforms/by more users? I find this fairly surprising and incomprehensible,. At
the very least it should be in big (friendly?) letters on the "running tests"
page.

Regards
Darryl