I think the work over the last year does count as "special circumstances", ie a major re-architecture of the graphics stack.The development work started at least a year earlier, but during the last year UMS and KMS needed to co-exist side-by-side in the same code base.
Starting with the spring 2010 distro releases I think pretty much everyone will be moving to KMS, although we'll want to keep the UMS paths going for a while until other OSes are able to implement the required KMS/GEM/TTM/DRI2 functionality and get away from UMS completely. There are other architectural transitions coming, eg moving Mesa from classic HW drivers to Gallium3D-based drivers, but nothing with the same bottom-to-top impact.
So yeah, special circumstances.
Most of the communication between developers is archived (IRC, mailing lists and Phoronix) so whether or not I control what developers say is irrelevent. If a different message went out in the past there should be a record of it.
Probably best for airlied or agd5f to comment on expected # of RCs. There are still reports of problems with the KMS paths but the recent ones don't seem to repro on the developer's systems so not sure how they will shake out.
I agree that hastily thrown together release candidates appearing only after vigorous campaigning would look shoddy, but since your first post about the 6.13 release came *after* the first RC was announced I assume that is a hypothetical question.
Starting with the spring 2010 distro releases I think pretty much everyone will be moving to KMS, although we'll want to keep the UMS paths going for a while until other OSes are able to implement the required KMS/GEM/TTM/DRI2 functionality and get away from UMS completely. There are other architectural transitions coming, eg moving Mesa from classic HW drivers to Gallium3D-based drivers, but nothing with the same bottom-to-top impact.
So yeah, special circumstances.
Most of the communication between developers is archived (IRC, mailing lists and Phoronix) so whether or not I control what developers say is irrelevent. If a different message went out in the past there should be a record of it.
Probably best for airlied or agd5f to comment on expected # of RCs. There are still reports of problems with the KMS paths but the recent ones don't seem to repro on the developer's systems so not sure how they will shake out.
I agree that hastily thrown together release candidates appearing only after vigorous campaigning would look shoddy, but since your first post about the 6.13 release came *after* the first RC was announced I assume that is a hypothetical question.
Comment