Hi Eric, *,

On Mon, Jun 30, 2008 at 1:40 PM, eric b <[EMAIL PROTECTED]> wrote:
> Le 30 juin 08 à 13:13, Christian Lohmaier a écrit :
>>> [...]
>>> Eric's point is that having unstable code available to the general public
>> [...]
> A lot of people are subscribed/read the lists (they are public, and often
> listed on gmane),

People who read the lists should know what the builds are about.

> and publy the links on their blogs or whatever other way
> of communication.

And when people do that, then having a seperate server for the regular
snapshots won't help.
Since those people would publish the links from the other server as well.

> [...]
> Again -> the only protection (for the image of the port) would be to use
> another server QA dedicated, and reserve only known-to-be-working builds on
> good-day, as "snapshots".

Well, this sounds a lot different from your initial demand.

But I still don't see it as the "only" protection, and by far not the
most efficient one to prevent this.

One needs to put the efforts to the very bottom of the queue, not at
the top of it.
beware - a car-related analogy will follow - as such it will always
have weak points, but I'll try to make my point clear:
If your car doesn't stop when you hit the break, what will you do:
* Buy a louder horn, so that you can alarm people to get out of your way
* Only drive at 10 km/h, so when you hit stuff, you won't break much
* Fix your breaking system

So I think fixing the breaking system is the best solution (work
towards milestones being what the milestone rules suggest: Always
building, always in a state in such no new problems are introduced)
Of course, driving slowly can help to prevent causes of an accident -
but this doesn't work when running downhill (when a release is near,
you cannot work up all the work that should have been done during the
milestone tests)

ciao
Christian

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to