Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [cdt-dev] Task Tags on Project Properties Page/Time for Voting?


> Two points,
> 1) Sean, please send messages to this list as plain text.  The digest 
> messed up when you don't.


> 2) Good point on the TODO tags.  This is another feature that went in 
> without proper review.  Also, whatever you are doing Sean with natures 
> needs to be reviewed (every C/C++ project should have the cnature, if 
> need new natures, do that instead).

If this is a question directed towards me in particular, I have not done 
anything to the cnature in any of my work on the standard build system. 
All of the info that was managed by the cnature before my patch is still 
managed by it now, and the interfaces to access that information is still 
the same. You are quite right that every new nature will have to have a 
requires-nature dependency on cnature, so I will have to modify the UI 
slightly to see if there are any other builders associated with the 
project. If there are, then I will not show the tabs for includes paths 
and defined symbols. That will solve the problem in the short term. 

> I'm starting to think that it is time to start voting on patches.  New 
> features are getting in through the patch list without review.  With so 
> many contributors putting work in, and I appreciate everyone's efforts 
> this, but I think the committers need to start clamping down on what 
> are committing.  In the end, the CDT needs to be of the highest quality 
> since we are including it in commercial products and to do that we need 
> show constraint.


> I would suggest that each patch needs the agreement of at least two 
> committers (preferably from different companies) with no descenters. 
> Essentially every committer has a veto.  The PMC should have the right 
> overrule any vetoes to make sure we don't get stuck in an impasse.  I'm 
> not sure if this is the eclipse way, but it'll ease us into a bit of 
> formality here.
> Thoughts?

We need redundancy then, so that critical path patches get reviewed and 
submitted promptly even if committers are away or unavailable.

Sean Evoy
Rational Software - IBM Software Group
Ottawa, Ontario, Canada

Back to the top