While I would like to see devfs removed from the kernel in the (near) future, Oliver had a point.
I don't think he meant stable as in 'not crashing'. He meant stable as in 'unchanging'.
No matter how buggy devfs may or may not be, the fact is that there are a lot of end users who are still depending on it. If it gets removed in the stable tree from one version to the next, any of these users who unwittingly decide to upgrade could suddenly be wondering where/dev/* went.
Such a change that has the potential to cause significant disruption among end users should probably be withheld for an unstable tree. In any case, it should be withheld *at least* until all the major vendors no longer depend on it.
It doesn't matter whether removing the code would make the kernel more stable to run. Any major change is a disruption, and *that* is what should be avoided halfway through a stable series.
Slashdot Mirror
While I would like to see devfs removed from the kernel in the (near) future, Oliver had a point.
/dev/* went.
I don't think he meant stable as in 'not crashing'. He meant stable as in 'unchanging'.
No matter how buggy devfs may or may not be, the fact is that there are a lot of end users who are still depending on it. If it gets removed in the stable tree from one version to the next, any of these users who unwittingly decide to upgrade could suddenly be wondering where
Such a change that has the potential to cause significant disruption among end users should probably be withheld for an unstable tree. In any case, it should be withheld *at least* until all the major vendors no longer depend on it.
It doesn't matter whether removing the code would make the kernel more stable to run. Any major change is a disruption, and *that* is what should be avoided halfway through a stable series.