On Fri, Apr 23, 2021 at 11:35 AM Steve Wills <steve_at_mouf.net> wrote: > Hi, > > On 4/22/21 6:00 PM, Alan Somers wrote: > > Because we have a policy of never releasing anything even a little bit > > backwards-incompatible in a minor release. > > I believe we have made backwards incompatible changes in minor branches > in the fairly recent past, at least compiler changes and VM changes that > affected video drivers. But I'm not sure, please correct me if I'm wrong. > Userland guarantees are easier to achieve than private kernel interfaces. Compilers have been binary compatible always, though newer versions were pickier at times which caused some minor grief (though that had nothing to do with the video driver stuff). By and large, things are broadly compatible. in the minor releases. And that's a good thing since we need packages for fewer release trains. But it also means things can take longer to get into a release. It's a balancing act, really, that we set the balance for decades ago and we may benefit from a close re-examination to see if a different model would let us use resources better and/or deliver features more quickly. WarnerReceived on Fri Apr 23 2021 - 16:08:10 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:28 UTC