Retrofit {Hex,Decimal}NumberExpression from v2 into v0/v1 spec
#657
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
Now that we can't reliably expose this parser starting from DSL v2, let's test individual choices
|
OmarTawfik
approved these changes
Nov 15, 2023
github-merge-queue bot
pushed a commit
that referenced
this pull request
Nov 28, 2023
Part of #638 ### Outline The new definition is used to construct the grammar definition model from the DSL v1, which is then used by the codegen to create the parser and rule/token kinds, like before. The rough translation done here is to translate every struct/enum/repeated/separated/precedence expression as a rule/non-terminal kind (with the exception of choice nodes, which are inlined, as per #650 (comment)); the rest (i.e. trivia, fragment, token, keywords) are translated as "old" tokens/terminals. ### More context In general, this requires more polish and decision which rules are we comfortable with leaving and fixing leftover issues such as explicit `PrecedenceExpression#rule_name` added to fit the old DSL v1 model/codegen better. I specifically didn't want to touch the v1 codegen/model yet, because it helped being able to incrementally migrate to a new definition and I wanted to limit changes done in a single PR, so that it's easier to review. We can modify the codegen/model to fit the v2 model better once we fully migrate and remove the old DSL v1. The translation/construction is done at build "run-time" rather than inside the newly introduced definition "compiler" proc macro. There's not much explicit reason for that, other than the fact that it was quicker for me to work on (quicker dev cycle) and the logic is just plain Rust code. This can be moved inside the proc macro later on, if we decide that it's better that way. ### Differences Because the DSL v2 takes a more named and structured approach (as it's designed to be a model for the strongly typed CST/AST), translating it 1-1 to our current parser structure is impossible or even desired. There are some rules that were introduced and it doesn't make much sense for some to replicate the old behaviour. 1. At the moment, it's impossible to specify an optional *separated* item, so the following ones are introduced as wrappers: - `TupleValue` (used in `TupleValuesList`) - `TupleMemberDeconstruction` (used in `TupleMembersList`) 2. Some sequences in the structs are optional, but these are now introduced and named, rather than inlined as before: - `IndexAccessEnd` - `ImportAlias` - `UsingAlias` - `VariableDeclarationValue` 3. ~Previously inlined sequence parsers now have to be named~ (fixed with #657) - `NumericExpression` is now split into two, now named, choices: `HexNumberExpression` and `DecimalNumberExpression` Moreover, the following was done to bring back the old CST shape as much as possible: - Some of the new rules where renamed back to the old ones, e.g. some repeated nodes have `List` suffix again - The (Yul)Statement are outlined (as in v0/v1) and re-introduced as a struct wrapper - `ArgumentsDeclaration` is outlined (as in v0/v1) as well - `FunctionCallOptions` is outlined again as a `(NamedArgs NamedArgs*)` rather than `NamedArgs | NamedArgs+` to better match the old CST in the singular case (but differs from the old CST in the multiple case) - #653 This was done to unify the definitions where possible and to help with reviewing the CST changes by minimizing the shape differences. Once we move off v0 (#637) and v1 (#638), we will be free to change the final shape, since the old definitions will stop tying us down quite a bit. ### Outstanding issues - [x] Separate outstanding FIXMEs that need to be done after the migration into a task list * LeadingTrivia in v2 is using the v1 definition for now * SourceUnit is hacked to always be followed by Leading Trivia (copies v1; trivia model in v2 is a bit different) * Clean up `PrecedenceExpression#rule_name` and adapt the codegen model to fit the v2 definition better * Stop leaking identifiers by adapting either v1 or v2 models * Keyword trie inclusion should be reworked to not require synthetic rules over all keywords (v1 model) and to properly allow keywords to overlap identifiers (#568) - [x] Fix the Prettier error caused by newly possible nested, empty rules: ``` [error] crates/solidity/testing/snapshots/cst_output/TupleExpression/empty/generated/0.4.11-success.yml: SyntaxError: All collection items must start at the same column (11:9) [error] 9 | - TupleExpression (Rule): # 0..3 "( )" [error] 10 | - OpenParen (Token): "(" # 0..1 [error] > 11 | - TupleValuesList (Rule): [] # 1..1 [error] | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [error] > 12 | - TupleValue (Rule): [] # 1..1 [error] | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ [error] > 13 | - CloseParen (Token): ")" # 2..3 [error] | ^ [error] 14 | ```
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.
Part of #652
Split out of #650