[asterisk-dev] [Code Review]: Preserve DTMF length in main/features.c

[Terry Wilson]

> I would propose a 2-tier committers system:
> 
> - A committer (someone with commit access)
> - A reviewer (a committer that can sign off on non-trivial changes on
> reviewboard)
> 
> Then you have to answer how one can become a committer or a reviewer.
> I'll kick off the topic with proposing:
> 
> - A contributer can become a committer if backed by a reviewer.
> - A committer can become a reviewer if backed by 2 other reviewers.

I'm perfectly ok with committer==reviewer. If we can trust someone to commit code to the repo, surely we can trust them to look at someone else's code to review it before committing it. Commit access without review access is pretty worthless. I see commit access as a way to reduce bottleneck problems of getting code into the tree. A monkey could almost be trained to commit code if it didn't have any right to decide whether the code should be committed at all.

All non-trivial patches should be reviewed no matter who the developer is (Digium employee, community committer, or random person submitting a patch). In my opinion, any developer who could commit a patch should be able to review someone else's patch and commit it. I think guilt and public shame for putting something through that is broken should be enough of a deterrent to sloppy reviewing. Nobody's perfect, and its all a learning process.

Terry