Split up FfiConverter #1766
Merged
Split up FfiConverter #1766
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
bendk
force-pushed
the
split-up-fficonverter
branch
2 times, most recently
from
ee94c6c to
767d02c
Compare
jplatte
approved these changes
Sep 26, 2023
Comment on lines
350
to
351
| // Note that because of webidl limitations, | ||
| // the key must always be of the String type. |
mhammond
approved these changes
Sep 26, 2023
| //! | ||
| //! These traits define how to pass values over the FFI in various ways: as arguments or as return | ||
| //! values, from Rust to the foreign side and vice-versa. These traits are mainly used by the | ||
| //! proc-macro generated code. The goal is to allow the proc-macros to go from a type name to the |
There was a problem hiding this comment.
I think this para might be better dropping the word "proc-macro" from "proc-macro generated code"? It almost implies "not by UDL" which I don't think is the intent.
| //! | | | ||
| //! | -------------- | ||
| //! | | | | ||
| //! [Return] [LiftRef] [ForeignReturn] |
There was a problem hiding this comment.
Not too keen on Return and ForeignReturn - it's a little clearer once you see the trait definiton, but less so when reading code referencing the trait without a method. How about LowerReturn/LiftForeignReturn (any maybe even dropping Foreign in the latter?)
bendk
force-pushed
the
split-up-fficonverter
branch
from
767d02c to
06f1d2a
Compare
Added new traits for specific operations (lift, lower, return, etc) and use those in the macro code. Added the `derive_ffi_traits!` macro to derive impls for those traits for types that define FfiConverter. The new system allows us have more fine grained control over what types implement. We now can implement Lower without Lift, Return without Lower, etc. This simplifies callback error handling. No more need for the handle_unknown_callback_error attribute. Errors that implement From<UnexpectedUniFFICallbackError> work with callback interfaces and errors that don't will result in a compile-time error. Many exported functions that used to cause UniFFI panics when called, like functions that inputted errors or results will now fail to compile. This required some changes to the keywords fixtures. It also means that we don't need to implement `FfiConverter` methods with a body that just panics.
bendk
force-pushed
the
split-up-fficonverter
branch
from
06f1d2a to
496ed73
Compare
3 tasks
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.
We've been discussing this for a while, here's my attempt at it. This makes it so the macro code stops using
FfiConverterand starts using more specialized traitsReturn,Lift, etc. TheFfiConvertertrait lives on, but just as a simple way to implement all the traits that we actually use in the generated code.This is rebased on top of the async-cancellation branch because that one had some changes to how we handled callback returns. Ignore all commits except the last. If we end up not taking that branch, I think it should be pretty easy to rebase over main too.