[Integration] main (b84f0bd) -> swift/main #533
Merged
Conversation
This file contains hidden or 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
…acter classes in unicode scalars mode (swiftlang#511) * [benchmark] Add no-capture version of grapheme breaking exercise * [benchmark] Add cross-engine benchmark helpers * [benchmark] Hangul Syllable finding benchmark * Add debug mode * Fix typo in css regex * Add HTML benchmark * Add email regex benchmarks * Add save/compare functionality to the benchmarker * Clean up compare and add cli flags * Make fixes * oops, remove some leftover code * Fix linux build issue + add cli option for specifying compare file * First ver of bitset character classes * Did a dumb and didn't use the new api I had added... * Fix bug in inverted character sets * Remove nested chararcter class cases * Remove comment it did in fact, not need @escaping * Cleanup handling of isInverted * Cleanup * Remove isCaseInsensitive property It was already being folded into the value on initialization, no reason to keep it * Add tests for special cases * Use switch on ranges instead of if * Rename asciivalue to singleScalarAsciiValue * Properly handle unicode scalars mode in custom character classes * I most definitely did not forget to commit the tests * Cleanup * Add support for testing if compilation contains certain opcodes * Forgot the tests again, twice in one day... * Spelling mistakes * Make expectProgram take sets of opcodes * Add compiler options + validation testing against unoptimized regexes * Cleanup, clear cache of Regex.Program when setting new compile options Co-authored-by: Michael Ilseman <[email protected]>
add comments
Make variadic builder overloads AEIC (take 3)
This stores both a source location, and has the ability to be `nil`, which is necessary to enable parser recovery in cases where we expect a number but parse something that e.g overflows.
Currently we use Swift error handling for parser errors. While this is convenient, it has a number of drawbacks: - Any AST parsed gets thrown away as soon as we encounter an error. This prevents clients from being able to get any useful information from invalid AST (rdar://93677069). - Multiple diagnostics cannot be issued, meaning that e.g a basic syntactic error could obscure a more useful semantic error. - It doesn't extend nicely to e.g warning diagnostics, meaning that we'd eventually end up with 2 ways of emitting diagnostics. - The thrown errors relied on `recordLoc` blocks to annotate them with source location info, which could lead to errors without location info if we forgot to add the appropriate `recordLoc` calls. Additionally, in some cases we want a more fine grained location info than the block would give us. Therefore this commit removes the use of Swift error handling throughout the parser. The parser is now a total function that _always_ returns an AST. If errors are encountered while parsing, they are recorded, and are attached to the resulting AST by the parser. The parser attempts to recover as much of the AST it can when encountering an error. As such, there is now are now `.invalid` atom and character property kinds. Sema then runs and can attach more diagnostics onto the AST. For now, the compiler interface remains the same, and we pick a single error to `throw`, but this will be changed in a later PR to allow multiple errors and warnings, as well as AST recovery. This also means we can better preserve the capture type in the presence of parser errors. Fortunately, in most cases, this is quite a mechanical transformation. It entails: - Moving the lexical analysis methods onto the `Parser`. We were already passing `ParsingContext` parameters for most of them, so it's not clear they were benefitting from the isolation that `Source` offered. Effectively this means that all parsing has access to the context and diagnostics. - Converting error throwing statements into calls to the parser's `error` method (or `unreachable` method for unreachables). This commit also updates the parser tests to be able to be able to match against multiple diagnostics.
Scan to the closing delimiter of an invalid identifier, and better diagnose an invalid text segment option.
We now always run validation, which is fine because the resulting AST can still be returned.
|
@swift-ci please test |
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.
No description provided.