Note that you effectively have a baseline of every delivery, in that you
can pick any change-set in the history of a stream, discard all of
change-sets after that stream, and recreate the configuration at the
time immediately after that delivery.

So like Ralph, I'd be interested in hearing what use case motivates the
creation of these baselines, since there may be a simpler way to achieve
it in RTC without automatically creating these baselines.

Cheers,
Geoff

Hi Andrea,

the reason I was asking is that this requirement seems not to add any benefit from my point of view. If every single check in is in it own baseline it seems to be the baseline is abused to simulate a form of tagging. And since then the system is swamped with baselines I see no potential benefit in having them at all.

The Idea of a baseline (from my perspective of course) is that some capable person for instance an integrator can baseline something considered special e.g. tested, integrated, in general some useful state in development. Since not every delivery will lead to such a special state it seems to be not worth a baseline. The change sets and the history are enough for that.

Since Streams and repository workspaces can be used to automatically rebase on the current state and to create special collections of change sets that are worth being baselined I would not recommend to do what you propose above. I have never felt the need to create a baseline for every change. Typically I baseline only if something is really done to a point that satisfying across a component. Also, sharing change sets can be done via various means. Even in other SCM systems I have never seen each and every change being tagged with all other constant data.

So from my perspective it seems odd to do what you want. But maybe there is a reason why you need all that baselines and I just don't get it. You describe below you want users to be able to pass baselines. For what purpose? Especially if they are working against the same stream and see the other's changes.

Hi Andrea,

yes, in this special case it might actually make sense to do that, just because you need to be able to refer to the specific baseline. But in this case I assume the deliveries are at least not typically really just one simple change set, but rather one that represents many changes and the delivery/accept is more an integration task.

Thanks for the clarification.

So it sounds like you not only want to create a baseline, but you want
it to be created with a specific name. Where would the automated
"make-baseline" get that name from?

Cheers,
Geoff