Marcel Offermans wrote:
On Jul 9, 2007, at 19:52 , Richard S. Hall wrote:
Since the sandbox is for conducting experiments on code that might
never be officially released into the trunk, perhaps we shouldn't be
using the official package space...for example, we could use
something like "org.apache.felix.sandbox.*" or something.
Of course, if you are experimenting with a branch of something that
is already in the trunk, then this wouldn't be necessary, but for
unbranched experiments it seems like we are opening up the potential
for naming clashes.
You say yourself that we cannot always use this convention, so I'd say
it's not necessary to have it. I think we should just agree that code
in the sandbox can never by released officially (neither as snapshots,
nor as parts of official releases). That eliminates most problems if
you ask me.
Well, it won't be possible to really enforce this since any arbitrary
person could make it available. And I view branching an existing project
for experimentation differently than a completely new experiment (i.e,
it doesn't violate the convention since it is a branch)...
However, if the consensus is that this is not an issue, then I can live
with it.
-> richard
Further, we need to have some overall agreement when assigning
official package names, so this probably shouldn't be done in
someone's sandbox. If an experiment moves from someone's sandbox into
the trunk, then we should decide what its official package name
should be.
Yes, I agree, that is the time to formally review the code and package
naming is one of the issues that should be reviewed.
Greetings, Marcel
