Proposal to sign all commits

David Lang david at lang.hm
Thu May 5 10:50:11 PDT 2016


On Thu, 5 May 2016, Kus wrote:

> My original thought was to allow us to grant many people commit access.


This is a good policy question to bring up.

Linux Kernel, with MANY more contributers than OpenWRT/LEDE has one person with 
commit access.

Admittedly, Linus is paid to do kernel work full time, so there is a good 
argument that if you don't have someone doing the merges full time, you need to 
have a small group of part-timers instead.

But with git, why do you need to have a lot of people with the ability to update 
the core? they can all run (and publish) their own tree(s) with whatever work 
they are doing.

The kernel has various people who do this, and they have someone running a tree 
that combines all the pending work until the next merge cycle starts, etc.



Which also brings up the question, what is the workflow that LEDE is planning? 
is it to merge patches continually until close to release time? or do you plan 
to have regular cycles where you merge patches for a bit, stabilize things, and 
repeat?

I'm a huge fan of the latter. It doesn't need to be a strict time-based approach 
like some projects do, but everywhere that I've seen take this approach has seen 
both more stability and more change.

David Lang



More information about the Lede-dev mailing list