You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When compiling a multi-platform target today you can access the platform being compiled as an ARG inside a docker build via TARGETPLATFORM. I need the same information however higher up in the bake file to appropriately assign a cache per platform. Today I cannot do that and due to locking restrictions on debian's package cache as a cache mount, it means that it's not possible to parallelize the particular RUN stage accessing the cache. Additionally having such a magic variable would also allow for better context interpolation as well. If I have separate parent contexts per platform today, I have to mount both to each target and then use a RUN+TARGETPLATFORM switch to copy the right architecture files into place during a multiplatform build.
The text was updated successfully, but these errors were encountered:
Description
When compiling a multi-platform target today you can access the platform being compiled as an ARG inside a docker build via TARGETPLATFORM. I need the same information however higher up in the bake file to appropriately assign a cache per platform. Today I cannot do that and due to locking restrictions on debian's package cache as a cache mount, it means that it's not possible to parallelize the particular RUN stage accessing the cache. Additionally having such a magic variable would also allow for better context interpolation as well. If I have separate parent contexts per platform today, I have to mount both to each target and then use a RUN+TARGETPLATFORM switch to copy the right architecture files into place during a multiplatform build.
The text was updated successfully, but these errors were encountered: