On Fri, Apr 25, 2008 at 07:42:31PM +0200, Tilman Sauerbeck wrote:
  

Juergen Daubert [2008-04-12 11:16]:
    

On Sat, Apr 12, 2008 at 10:39:15AM +0200, Tilman Sauerbeck wrote:
      

Hi guys,
        

Hello,

      

core ports aren't supposed to have specific maintainers, right?

tilman@brimstone [] > grep Maintainer */Pkgfile|grep -v core-ports
bzip2/Pkgfile:# Maintainer: Tilman Sauerbeck
exim/Pkgfile:# Maintainer:  Juergen Daubert
iproute2/Pkgfile:# Maintainer: Simone Rota
libarchive/Pkgfile:# Maintainer: Tilman Sauerbeck
libusb/Pkgfile:# Maintainer:  Jürgen Daubert
pciutils/Pkgfile:# Maintainer:  Jürgen Daubert
pkg-config/Pkgfile:# Maintainer: Tilman Sauerbeck
tilman@brimstone [/usr/ports/core] > 

(output truncated). These should be fixed, right? :o
        

not really sure about that. As Per retired we decided, as a quick
solution, to introduce the core-ports maintainer. The above ports
are moved from opt to core for some reasons, so the maintainers
are the original ones. 
      

This doesn't apply to all core ports though. eg core/libarchive still
has my name in it because I originally had the port in my private
repo, and when I moved it to core I forgot to update it.

    

We never discussed in deep how to handle the core ports at all,
      

I suggest to put "core-ports" as the maintainer for the remaining core
ports.
    

basically I agree, but see my remarks below.

  

I think that we are talking about the organisation and structure 
of whole project at this point ... 
      

Why? While we normally have a real human being listed as a port's
maintainer, the core ports would then be the exception that proves the
rule ;)
    

to avoid double work and to optimize quality. 

It happens often that I'm working on a port but you are 'faster' 
and has committed the changes already, I bet you've seen it the 
other way around. Strictly it's a good thing if two people are 
looking over the same work, but we should try to organize this a 
little bit better, even more if more than two dev's are involved. 
One possibility might be to commit to a devel-branch and merge 
the changes after some kind of signing by others. But that sounds
like overkill to me.