> > I'm not going to reinvent the wheel analyzing how hard it is to say > how many features make it worth releasing new version. A lot of big > (and actively developed) projects are using cycling releases now > (including Linux kernel). And come one, OpenWrt is getting a lot of > patches almost daily. > I like 6-monthly (of similar) release cycle, it's just a matter of > fact of finding some ppl leading that releases. > > > -- > Rafał
So it doesn't sound as if there is any currently defined release or maintenance cycle. I don't have an appreciation for how long a branch would be maintained, but I would imagine that it is a function dependent on the number maintainers available. I would also add that looking at the openwrt revision history, updates to branches seem to stop after about 2 - 2.5 years. (last update to 8.09 was 3 years ago, last update to 10.03 was 8 months ago) I also gather that the general consensus is that the releases are too far apart, however I think that 2 years is just right for inter-LTS release spacing, it matches what Ubuntu is doing for what it is worth. Additionally, the historical data points out that there hasn't ever been more than one active maintenance branch at a time. (the last update to 8.09 happened before 10.03 and the last update to 10.03 was before 12.10) Is this due to a lack of available maintainers or a decision that after 2-2.5 years the branch is no longer relevant? I understand that all of this is a delicate balance of bringing in newer kernels with new drivers/features vs. longer upstream support, and that it is all divisible by the available number of maintainers. It is a difficult problem. I don't claim to have an answer, but I am of the opinion that the rate at which new openwrt-oriented hardware is released, the length of time the hardware is available, and the rate at which most people replace their openwrt hardware is a longer period of time when compared to the desktop and server market. This opinion also makes me want to continue to see openwrt releases that I can rely on for 2+ years. Given the above Openwrt historical data and the current kernel LTS outlook given at: https://www.kernel.org/releases.html even if BB was released tomorrow on the most recent LTS kernel (3.4) it would only have 1.5 years of upstream kernel updates. In my eyes, that's a little short.
_______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
