forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Thorsten Scherler <>
Subject to branch or not to branch part II (was Re: Dispatcher rc1)
Date Wed, 01 Feb 2006 19:48:32 GMT
El mié, 01-02-2006 a las 15:37 +0000, Ross Gardler escribió:
> To Branch or not to Branch?
> ===========================
> [ASIDE] Don't worry Thorsten I'm not going to ask you to change anything 
> . I heard your frustration at the circular discussion about "to branch 
> or not to branch", that is why I stayed quite in the end, I was silently 
> agreeing to you just getting it done

Thanks for that, I really appreciate what you have done because I could
do the job.

> This lack of ability to support both V2 and V3 during this work, which 
> Thorsten had previously warned us about, was the reason I recommended a 
> branch.

Actually the fault is lying in
which normally should not be the standard situation for plugins and our
new properties system looks very promising to overcome the need for such

IMO Forrest should support different versions of plugins which should be
configurable on a project base. That would make it easy to develop
different versions of plugins without the need of a branch.

> For the future reference of the community (not for Thorsten or any other 
> individual), if a development will temporarily break existing 
> functionality we should branch. 

IMO only for core code. 

> This was one of the criteria we 
> identified for branching in our previous discussions on this subject.
> With a branch we could just do an svn switch and it would work, then we 
> could just switch back to trunk to work with v2 stuff if necessary.

You are totally right and regarding core code I am with you to branch.

One problem is that switching branches as well leads sometimes to
conflicts due to e.g. locally modified sources. If we do not deal with
core code I prefer a new version of a plugin rather then a branch.

> Thorsten did point us at this patch within the large number of commit 
> emails and Jira issues he generated in his mammoth weekend attack on 
> this work. 

:) thx for the compliment. ;-)

> However, it seems both David and I missed the importance of 
> that message and have therefore wasted our time, and Thorstens, in 
> debugging the issue.

No, you have not wasted my time, but helped pointing out that the
instructions on the FOR-796 have been confusing (I am changing them
right now).



"Together we stand, divided we fall!" 
Hey you (Pink Floyd)

View raw message