Anders

Ideally one could just add a check to `validate_manifest()` that throws an error when `minimize_size` has `shrink` enabled and the image type is `raw`.

Hm, I think what we need to do here is fix how the workspace path is handled. I never thought of using a relative path for that part. To keep...

> please give me some hints how you would like to get the workspace path relative to the manifest done. Oh, that's easy. Just use [`rel_path`](https://github.com/andsens/bootstrap-vz/blob/7913246c2d1d4071a0cd11a535e0f509acf861af/bootstrapvz/plugins/ansible/tasks.py#L26) like all other tasks...

Ah yes, I remember this, there was quite a bit of hassle with those labels. This: ``` 'LABEL description='"'"'Hashbang shell server, based on Debian jessie'"'"'' ``` looks really bad. What...

@liori, oh that is interesting! I am however puzzled as to why this would change the permissions on `/tmp`, since we explicitly mount bind `/tmp` *after* bootstrapping, where all the...

Sounds like a great idea. Relative paths have been introduced in other parts of the manifest as well. We should stay with the practice of making them relative to the...

No, not really. I also think that defaults which change depending on other variables can be quite unpredictable, better to go with just showing what to set it to in...

That is a great idea. I love the concept of `commit`s between phases. We would have to leverage dockers built-in concept of caching in order to reuse layers, which is...

> that way, the same layers get built, get the same id (as the id is a hash of the layer) and get only stored once. You won't get the...

Well f**\* :-/ I had no clue and assumed that if James shipped them with the official images they'd be _good enough_. I'll review them and replace the current files....