Re: [Boost-testing] Request for new 'unresolved issues' view

From: Alisdair Meredith (alisdair.meredith_at_[hidden])
Date: 2005-06-15 08:07:52

Markus Schöpflin wrote:

> I agree with you. So it looks like the regression testers currently
> fulfill the role which I had in mind when speaking of platform
> maintainers.

I believe there could still be a role for platform maintainers, if
there are people interested in stepping up.  It just so happens that
most of the people I would imagine in the role are already regression
testers <g>

The difference is in what you are offering.
A regression runner is offering hardware and resources to regularly run
the tests.

A platform maintainer would be someone intimate with the knowledge of a
platform - its idiosyncracies and workaround for all the commonly
encountered problems.  Ideally they would be the best placed person
mark libraries as unsupportable on a given compiler, identify tests
that are expected to fail, and help library authors port new
submissions onto their platform.

This is a lot easier if you are running regressions, but I don't think
it should be a requirement, and vice versa.

Given the current overlap of the people involved it may not be worth
making the distinction.  OTOH, if it means we get more knowledgable
people involved in actively supporting their platforms, it would be a
good thing.

AlisdairM