fix: Add parent dirs to mtree #96
Conversation
|
|
| results.append(_mtree_line(parts, type = "dir", mode = "0755")) | ||
|
|
||
| return results | ||
|
|
||
| # Copied from aspect-bazel-lib/lib/private/tar.bzl |
There was a problem hiding this comment.
Looking at this comment and your the code you are adding, is this something we could simplify by using some rules from https://github.com/bazel-contrib/tar.bzl/blob/main/docs/mtree.md?
There was a problem hiding this comment.
What part are you looking to simplify exactly?
The mtree_spec rule simply creates an mtree file for the files passed in srcs, it won't create entries for parent dirs of these files. mtree_mutate has many properties but none that would help either except having a whole different awk_script.
There was a problem hiding this comment.
Didn't look too much into the details. I just saw this comment and would have guessed that this could be implemented already somewhere else. Nevermind.
There was a problem hiding this comment.
I agree we should try to de-duplicate these if we can.
There was a problem hiding this comment.
Sorry I didn't realize this is related to the comment.
What about making _mtree_line public in @tar.bzl//tar:mtree.bzl, similarly to what is done for s3_sync in rules_aws ?
There was a problem hiding this comment.
Yeah that's the obvious answer. However I suspect there's a somewhat more substantial code sharing possible here, since https://github.com/aspect-build/rules_py/blob/main/py/private/py_image_layer.bzl was created after this rule is sketched out. @thesayyn can confirm - or if you have a few minutes to compare the implementations, maybe it's obvious how to do it.
Nested files/folders in image layers result in
InvalidImage(MissingParentDirectory: Parent directory does not exist for filewhen deploying on AWS if the parent folders aren't listed in the layer archive mtree.Changes are visible to end-users: no
Test plan
aws_py_lambdawith internal and external dependencies, pushed to ECR and deployed on AWS Lambda. Successfully ran code using internal and external dependencies.//examples/python_lambda:tarballbazel run //examples/python_lambda:tarball # Then push, deploy and run the image on AWS Lambda