Files
sdk/pkg/analyzer_cli
Sam Rawlins 958ae6ab1e analyzer: Simplify AnalysisContextCollectionImpl with new updateAnalysisOptions4 parameter
The new updateAnalysisOptions4 parameter is a callback function,
similar to updateAnalysisOptions3, but with one removed parameter
itself, the `DartSdk sdk` parameter.

The one place where this parameter was used in the callback passed by
the caller, was in DAS's ContextManager, and it was just used in order
to compute a FeatureSet from (a) the computed DartSdk and (b) the
command-line `--enable-experiment` options. We can instead provide a
`enabledExperiments` parameter on AnalysisContextCollectionImpl,
leading to a simpler API and still a simple implementation in
ContextBuilder.

The reason for the incremental migration is that dartdoc, and possibly
some internal clients, use updateAnalysisOptions3.

Change-Id: Ie990ab8f7b850aed1ce3ef39ab71c01835c080da
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/447400
Commit-Queue: Samuel Rawlins <srawlins@google.com>
Reviewed-by: Konstantin Shcheglov <scheglov@google.com>
2025-08-27 11:05:29 -07:00
..
2023-11-17 20:17:23 +00:00

SDK development code analysis

dartanalyzer used to be the tool for statically analyzing dart code at the command line. However, this tool has been replaced with dart analyze for this purpose in current SDKs and will no longer be published on pub.

Do not depend on the command line interface or other semantics in this directory as it is now an internal tool for SDK development, used as the dart2analyzer "compiler" for tools/test.py in the SDK. It is configured as part of the test runner, here.

SDK development usage

For SDK development, run analysis from the test tool to validate analysis conclusions on language samples in the testing directory. From the root of the SDK:

tools/test.py --build --use-sdk -c dart2analyzer co19 language

This will build the Dart VM and compile dartanalyzer into a snapshot, then use that snapshot while analyzing those directories under testing/. Without --use-sdk, test.py will use the source code version of the analyzer instead of the compiled one, which can be useful for debugging.