Exploded layer proof of concept #203
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
schneems
force-pushed
the
schneems/get-set-exploded-layer
branch
from
d2d0233 to
4974006
Compare
schneems
force-pushed
the
schneems/logging-state-machine-continued
branch
2 times, most recently
from
34ea214 to
02758b4
Compare
schneems
force-pushed
the
schneems/logging-state-machine-continued
branch
3 times, most recently
from
64c7e62 to
ce52872
Compare
schneems
force-pushed
the
schneems/logging-state-machine-continued
branch
from
0ebcbe9 to
7c85b5e
Compare
schneems
force-pushed
the
schneems/logging-state-machine-continued
branch
2 times, most recently
from
d645702 to
b3ccd76
Compare
Base automatically changed from
schneems/logging-state-machine-continued
to
main
October 30, 2023 15:11
The current design of the Layer trait encourages developers to put lots of logic into their layer. One impact of this, is it forces stateful logging to jump through hoops that I would call less ergonomic heroku/libcnb.rs#669. This proof of concept is a re-imagining of how we might want to interact with a layer using only the existing layer primitives from libcnb. It encourages operations to be performed outside of a layer, because it does not require the developer to implement Layer on their own structs. Instead it treats metadata as a primitive and makes "read" and "write" layer information their own discrete actions. At a high level the flow of interacting with a layer now becomes: - Read cached data (get LayerData) - Do stuff (modify path, log, etc.) - Write new data (write LayerResult) In this process I reimagined the flow for cache invalidation to fit within a `MetadataDiff` enum. This essentially encourages a set of "reasonable defaults" for equality based metadata, but does not restrict to a specific use cases. Worth noting that I've got an idea of a way to better handle metadata that cannot be deserialized, but it would be distracting from the primary purpose of this PoC. ## Known The code in `main.rs` is quite long and would likely be moved to its own function if we continued down this path. ## Discussion/Feedback The purpose of this PR is to get feedback on this approach, both high level (concept) and low level (implementation).
schneems
force-pushed
the
schneems/get-set-exploded-layer
branch
from
4974006 to
cbaecb1
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The current design of the Layer trait encourages developers to put lots of logic into their layer. One impact of this, is it forces stateful logging to jump through hoops that I would call less ergonomic heroku/libcnb.rs#669.
This proof of concept is a re-imagining of how we might want to interact with a layer using only the existing layer primitives from libcnb. It encourages operations to be performed outside of a layer because it does not require the developer to implement
Layeron their own structs. Instead, it treats metadata as a primitive and makes "read" and "write" layer information their own discrete actions.At a high level the flow of interacting with a layer now becomes:
In this process, I reimagined the flow for cache invalidation to fit within a
MetadataDiffenum. This essentially encourages a set of "reasonable defaults" for equality-based metadata, but does not restrict to a specific use cases.Worth noting that I've got an idea of a way to better handle metadata that cannot be deserialized, but it would be distracting from the primary purpose of this PoC.
Known
The code in
main.rsis quite long and would likely be moved to its own function if we continued down this path.Discussion/Feedback
The purpose of this PR is to get feedback on this approach, both high level (concept) and low level (implementation).