Closed a-gave closed 8 months ago
Thanks! It looks great :)
And it also allows us to remove one step from the compilation instructions, right? I refer to this one: "Deselect erroneous feeds locations:
Image configuration → Separate feed repositories → Enable feed libremesh
Image configuration → Separate feed repositories → Enable feed profiles"
from this page: https://libremesh.org/development.html#compiling_libremesh_from_source_code
Thanks!
-- Inviato dal mio smartstone
Thanks! It looks great :) And it also allows us to remove one step from the compilation instructions, right? I refer to this one: "Deselect erroneous feeds locations: Image configuration → Separate feed repositories → Enable feed libremesh Image configuration → Separate feed repositories → Enable feed profiles" from this page: https://libremesh.org/development.html#compiling_libremesh_from_source_code Thanks!
On second thought, it seems more correct to me that it is deselected at build time.
Ok! Both options (to instruct the user to deselect that or to do it automatically) seem good to me.
Thanks!
I apologize for bothering again on this topic, but compiling via imagebuilder once again turns out to be broken.
I did not calculate that a subsequent compilation from the build.yml workflow would delete the subdirectory previously created by the multi-arch-build.yml workflow (e.g. now missing https://github.com/libremesh/lime-feed/tree/gh-pages/master/packages).
This fixes it, telling the multi-arch-build.yml workflow to use a different path for output packages to avoid overwriting. This updates also
packages/lime-system/files/etc/uci-defaults/92_add-lime-repos
andREADME.md
accordingly.This also removes the feeds added by default by buildroot for each feed declared in feeds.conf and resulting in a known opkg error such as the one below: