On Wed, Aug 05, 2009 at 11:37:34AM +0100, Chris Bowditch wrote:
> Vincent Hennebert wrote:
> >Hi Simon,
> 
> Hi Simon, Vincent,
> 
> >
> >The report looks good. Just a thing about the ChangingIPD hack: actually
> >it is still undecided whether it should be merged back to Trunk or not.
> >
> >My main concern about that is that it will get in the way when
> >implementing the new approach. I will basically have to start with
> >reverting all the changes made for the hack.
> >
> >My idea was to actively maintain the branch in sync with the Trunk.
> >Releases could then be made from the branch. But this may actually turn
> >out to be more work than reverting all the changes in one go. Plus it
> >has been mentioned to me that this may create difficulties for users who
> >regularly take snapshots of the Trunk (the branch will always lag
> >a little bit behind).
> >
> >Any opinion about this is welcome. Meanwhile, I???ll modify the status
> >report.
> 
> Thanks for bringing this up Vincent. I am a little reluctant to be
> working off a branch as we regularly take build snapshots from
> trunk. This creates a lot of extra work keeping the branch in sync.
> 
> I think we all acknowledge that the changing IPD hack is just that
> and therefore code quality may be sub optimal. However, since
> Vincent is working on a totally new layout engine then we expect
> this code to be replaced a year or so down the line anyway.
> Therefore if the only reason for not putting this change into trunk
> is to save effort at the point when the new layout algorithm is
> introduced then I think this hack should be merged with trunk. This
> is because it will be more time consuming to do a 100 or so merges
> over a 1-2 year period than to remove changes from trunk in a one
> time exercise.

I appreciate your point of view. I expect you will announce your
intention to merge the changing IPD branch with trunk when the time is
there. Then it is a good time to hear everyone's opinion.

Simon

-- 
Simon Pepping
home page: http://www.leverkruid.eu

Reply via email to