Re: [Gnumed-devel] Re: Understanding code maintenance

[James Busser]

On 1-Feb-08, at 4:19 PM, Karsten Hilbert wrote:

> > Do commits into the "main trunk" (or "head" ?) happen at multiple
> > points, even in between official releases and maybe even with just
> > bug fixes?
> No. They always happen upwards of the current official release  
> before what will become the branch root for the next official release.

Is "main trunk" same as HEAD?

When one does checkins, one specifies a directory where the checkin  
is to go, i.e. into any one particular branch, or into HEAD, or into  
a personal area?

When one supplies a patch for a bug fix, one still has to checkin,  
and then commit, but it would be customary to commit right away if  
there is confidence in the patch, so that the availability of the fix  
is not delayed?

When one checks new code into HEAD, does it get committed at  
intervals, in between branch points, or is it only ever committed as  
part of the act of getting tagged and a new branch point created and  
therefore a release producible?

When something is checked-in (say, to HEAD), but not yet committed,  
then when (for example) I would do anonymous checkout from cvs ---  
and when I would then run ./gm-from-cvs.sh --- would I by default  
only be running the committed-code, or would this shell run any  
checked-in code as well, if there is any value in my testing what had  
been checked in?