JP Simard
JP Simard
@walterg2 looks like the identifiers are always null in the reporters?
This is a generally useful feature for all rules, not just the function body length rule. Is there a way all rules that point to an identifier could benefit from...
If we do this it needs to work with all rules where this applies. Not all violations point to a specific identifier though, and not all identifiers are methods.
> I'm having trouble finding the entire inheritance chain through sourcekitten, any suggestions would be helpful! Normal SwiftLint linter rules only operate on the pre-typechecked AST. SwiftLint has "analyzer" rules...
I've audited the OSSCheck results and the only changes are fixed false positives.
Let me know when you'd like this reviewed @VldmrVstsk !
Sadly OSSCheck reports that the new Swift parser is 31-69% slower than the previous C++ one 😢 There are also some violations that were introduced and others that were removed...
There are new TSan races with this which I’ve reported in https://github.com/apple/swift-syntax/pull/807#issuecomment-1250700072
> Sadly OSSCheck reports that the new Swift parser is 31-69% slower than the previous C++ one 😢 ~Then again, linting Lyft's iOS monorepo locally is now faster (459 files/s...
Turns out that when compiling SwiftLint with `bazel build -c opt swiftlint`, this change has a 30%-70% performance degradation. However, when compiling with `swift build -c release`, this change has...