Log cache decisions in order to better understand cache misses

[gautaz]

Hello,

I did not find any existing issue related to this, my apologies if this is not the right place.

Many times, I find myself wondering why buildah did not use the cached layers during a build while I had not directly modified any file in a project.
It seems I am not the only one dealing with this kind of issue, see: https://github.com/MShekow/directory-checksum/.

Would it be hard to add some logs in buildah which would detail how it decides to use or not cached layers?
This way, one would understand why a cache miss occurred and probably fix it by ignoring the right files.

[flouthoc]

I think we can document a workflow first, when and how cache-layers are detected by buildah ? @gautaz would that help ?

[gautaz]

Hello @flouthoc, it would certainly help to understand how buildah manages its cache 👍, documentation never hurts as long as it can be maintained (which is in fact an additional burden to consider).
If this documentation could also give hints to the reader on ways to better diagnose and avoid certain caching issues, that would be great.
It can probably be organized based on the Dockerfile command used at the step where the user expects the caching system to operate but does not.
Ultimately, as I said earlier, this documentation can be an additional burden for buildah developers/maintainers when modifying/improving the caching system, so I am not entirely sure it will satisfy everyone...