[torquedev] Big Bang vs Evolution (was: IPv6 port)

Glen Beane glen.beane at gmail.com
Fri Dec 26 20:10:50 MST 2008


On Fri, Dec 26, 2008 at 7:03 PM, Matti Bickel <cat5 at minet.uni-jena.de> wrote:
> Hi,
>
>  i'm slowly chipping away at the ipv6 branch here. But as time goes on
>  the branch diverts more and more from trunk. It's hopeless to play
>  catch up here, i'm moving too slowly.

I thought Garrick (or was it someone else?)  had volunteered to keep
your brancy in synch with trunk?


>  Therefore, i'm trying to backport my fixes to trunk and break them
>  up into small, manageable chunks. Because i don't want to bring IPv6
>  to the trunk just yet (too scary), i'll try to be more gentle and make
>  changes such that each one faciliates ipv6 usage, but will not
>  introduce anything related to ipv6.
>
>  If you like the new[tm] way, please apply the attached patch after
>  review. It's nothing spectacular, but uses pbs_net_t to hide away all
>  the pain that is associated with changing the common ip type later on.
>
>  To make the changeset smaller, i haven't touched the port handling
>  code or converted any of the ulong ip address uses to pbs_net_t.
>
>  Dunno how you want to handle this kind of change or if you want it at
>  all. My target is to produce a changeset approximatly each day and
>  change a few lines of code each iteration. The iteration will compile
>  but not necessarily work as i don't have the means to test everything.
>
>  Please comment on this idea and give me a "no, thanks" or "yes, please
>  keep it coming"...

In principal I think this is a good idea,   but I dont know about a
few lines of untested code eacy day.  I think a slightly larger patch
once every week or so that has had at least some (not necessarilly
exhaustive) testing would be more manageable.


More information about the torquedev mailing list