On Thu, 17 Aug 2023 at 06:41, Mikko Rapeli <[email protected]> wrote: > One of the major problems with yocto major updates is that it becomes a big > bang, > everything must be updated at once. This can be split down and individual > layers updated > one at a time while keeping poky on the old version. This requires working > around > layer compatibility, which in many cases is quite artificial and just > indiciates that layer > is not tested with the other poky versions but it might still just work. > > I've done this so many times now that I don't see any reason to hide it, > also because many non-core layers are simply not maintaining other than > master branch > exposing users to obsole and insecure SW if they stick to the branch name.
Getting around LAYERSERIES_COMPAT or using BBMASK in order to update layers one at a time is not the problem here. The problem here is that a newcomer wants to have an older boost (which is in poky) with a newer everything else (which is also in poky). What I'm trying to say, nobody stopped and asked, what are they actually trying to do? What made them think they need an older boost? Why isn't a newer boost suitable? Maybe it is, with a bit of work? You should know very well that making two versions of the same thing coexist in the same build is painful and often impossible if they clash in sysroots or packaging. Rolling back to an older version is also painful, but for a different reason: the rest of the stack may actually have moved on, requiring newer APIs and features. Older version may also be insecure. So can we try to do the right thing first? Alex
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#60815): https://lists.yoctoproject.org/g/yocto/message/60815 Mute This Topic: https://lists.yoctoproject.org/mt/100776894/21656 Group Owner: [email protected] Unsubscribe: https://lists.yoctoproject.org/g/yocto/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
