On Wed, 3 Aug 2022 at 23:02, Chris Laplante via lists.openembedded.org <chris.laplante=agilent....@lists.openembedded.org> wrote: > But IMHO bundling the layer setup tool inside bitbake itself feels a bit like > the chicken-and-egg problem. I'd like my spec file (or JSON file, kas > configuration file, whatever) to be a complete manifest of things to > download. But with bitbake-layers, first I need to > know the right version > of poky and where to get it. Also, I think it would be hard to implement 'yb > sync'-like functionality there, specifically if we wanted to be able to > change between poky branches. > > Overall the thing I like about yb is that it is independent of Yocto/BitBake > and easier to setup. Like an IDE, I can install it once and use it across > multiple projects.
The bitbake-layers command is needed only to generate the layer config file (in json format) out of an already existing yocto setup. It does not perform the actual layer fetch/setup from the json. That's the feature I like the most in my proposal: none of the other tools bootstrap the config file in this way, you need to always write it by hand. The actual layer setup (using the json) is performed by an independent, self-contained python script (copied from a template in oe-core), which you can place anywhere you want (e.g. in a product layer repo, or in a 'config stream repo'), and which does not require an active yocto environment or a poky checkout. The equivalant of 'yb sync' would be to pull that repo to get latest revisions of the script and the json, then run the script pointing to an existing checkout - it will sync everything to the revisions in json. Alex
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#168828): https://lists.openembedded.org/g/openembedded-core/message/168828 Mute This Topic: https://lists.openembedded.org/mt/92801786/21656 Group Owner: openembedded-core+ow...@lists.openembedded.org Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
