Too many lords, not enough stewards
Too many lords, not enough stewards
Posted Feb 3, 2018 22:29 UTC (Sat) by jani (subscriber, #74547)In reply to: Too many lords, not enough stewards by neilbrown
Parent article: Too many lords, not enough stewards
Yes, but who in their right mind would want to? The community keeps preaching everyone who cares to listen, and probably many who don't care, about working upstream. Carrying local patches is expensive. I don't think forking the kernel is a viable option to bypass maintainers at any level.
Or I just didn't understand what you mean by forking in this context.
(Log in to post comments)
Too many lords, not enough stewards
Posted Feb 4, 2018 5:47 UTC (Sun) by neilbrown (subscriber, #359) [Link]
Yes it is - long term. Short term it is very cheap.
Apparently working with certain maintainers is also expensive (with others it is a joy).
Depending on your particular needs at a particular time it makes sense to perform a cost/benefit analysis and decide what the best course of action is.
Maybe the best approach is to persist with the maintainer.
Maybe it is to try to route around them.
Maybe it is to fork and maintain a separate tree for a while, and then try to merge again in 6-12 months when circumstances might have changed.
Maybe it is to fork permanently.
Key point is that you have options and you can take control within the parameters of those options. You cannot force other people to change their behaviour, but you can choose how you will behave, and it is valuable to have a clear view of all of the options.
How a person chooses to behave typically speaks more loudly than unsolicited opinions they might choose to present.
Too many lords, not enough stewards
Posted Feb 4, 2018 8:36 UTC (Sun) by rodgerd (guest, #58896) [Link]
> Yes, but who in their right mind would want to?
Are the people who work at or on Red Hat, SuSE, Google, Amazon, Debian and Oracle? Alan Cox was insane (news to the people who found years of -ac kernels vastly better than Linus kernels)?
The hyper-majority of Linux users are no-where near a mainline kernel, and many developers only touch it to pull patches into their own trees.
Too many lords, not enough stewards
Posted Feb 4, 2018 14:20 UTC (Sun) by jani (subscriber, #74547) [Link]
> Alan Cox was insane (news to the people who found years of -ac kernels vastly better than Linus kernels)?
>
> The hyper-majority of Linux users are no-where near a mainline kernel, and many developers
> only touch it to pull patches into their own trees.
Yet most prefer being as close to upstream as possible, carrying local patches for their chosen stable release, downstream, perhaps to provide "value add" for their customers. Contrast this with, say, hardware vendors or individual developers forking upstream to bypass the maintainer structure, and trying to convince the above mentioned downstreams to carry their out-of-tree patches to deliver to the end users.
I suppose you can argue some level of downstream forking happens all the time, but I just don't see it as a relevant argument in the discussion at hand.
Too many lords, not enough stewards
Posted Feb 4, 2018 21:26 UTC (Sun) by neilbrown (subscriber, #359) [Link]
It is not unheard of for a hardware vendor to partner with an distro to work on getting hardware support upstream - each side brings different skills for mutual benefit. They work together on a fork when circumstances prevent them from working together upstream.
If an individual developer is having trouble getting a patch upstream, it may make perfect sense to submit a bug report/feature-request to their favourite distro and say "I have a bug, I have a fix, I cannot git it upstream, could you take it directly?". This distro maintainer might do that, or might help get it upstream, or might do both.
This is all part of "routing around".