To be extra clear: it's not "one gadget, one image". The model assertion is really what defines what goes into the image, including the gadget itself.
So it's really "One model, one image." That's a very nice invariant, considering that the model is actually signed by the brand. On Tue, Oct 11, 2016 at 7:49 PM, Gustavo Niemeyer < [email protected]> wrote: > > > On Tue, Oct 11, 2016 at 6:45 PM, Oliver Grawert <[email protected]> wrote: > >> > You can use the same image for whatever you need. One image, one >> > gadget. > > > > if you define no image size today and use ubuntu-image today you >> (should) get a 300MB image that you dd to a device ... on startup, >> before the first mount occurs the image gets resized to the full disk >> size and it gets made sure that your GPT backup table sits in the right >> place at the end of the disk. >> >> if you start to hard code sizes you lose all this flexibility, you will >> already not be able to use the pc image on a 2GB USB key today simply >> because we hard code the size for it to ~3GB just for the kvm use case. >> > > Please re-read the point I made above. I didn't talk about resizing at all. > > One gadget, one image. If it makes sense for a category of devices to > resize, let's figure the best way to represent that in gadget.yaml. > > > (...) > >> so you think we should have a pc.img, a pc-kvm.img, pc-vmware.img and a >> > pc-cloud.img using the same kernel the same core snap and apart from >> one line the same gadget ? >> > > If all of these environments can run with the exact same image, by all > means let's have a single image. > > > gustavo @ http://niemeyer.net > -- gustavo @ http://niemeyer.net
-- Devices mailing list [email protected] Modify settings or unsubscribe at: https://lists.snapcraft.io/mailman/listinfo/devices
