Documentation ¶
Overview ¶
Package checker defines the implementation of the checker commands. The same code drives the multi-analysis driver, the single-analysis driver that is conventionally provided for convenience along with each analysis package, and the test driver.
Index ¶
Constants ¶
This section is empty.
Variables ¶
var ( // Debug is a set of single-letter flags: // // f show [f]acts as they are created // p disable [p]arallel execution of analyzers // s do additional [s]anity checks on fact types and serialization // t show [t]iming info (NB: use 'p' flag to avoid GC/scheduler noise) // v show [v]erbose logging // Debug = "" // Log files for optional performance tracing. CPUProfile, MemProfile, Trace string // IncludeTests indicates whether test files should be analyzed too. IncludeTests = true // Fix determines whether to apply all suggested fixes. Fix bool )
Functions ¶
func RegisterFlags ¶
func RegisterFlags()
RegisterFlags registers command-line flags used by the analysis driver.
func Run ¶
Run loads the packages specified by args using go/packages, then applies the specified analyzers to them. Analysis flags must already have been set. Analyzers must be valid according to analysis.Validate. It provides most of the logic for the main functions of both the singlechecker and the multi-analysis commands. It returns the appropriate exit code.
Types ¶
type TestAnalyzerResult ¶
type TestAnalyzerResult struct { Pass *analysis.Pass Diagnostics []analysis.Diagnostic Facts map[types.Object][]analysis.Fact Result interface{} Err error }
func TestAnalyzer ¶
func TestAnalyzer(a *analysis.Analyzer, pkgs []*packages.Package) []*TestAnalyzerResult
TestAnalyzer applies an analyzer to a set of packages (and their dependencies if necessary) and returns the results. The analyzer must be valid according to analysis.Validate.
Facts about pkg are returned in a map keyed by object; package facts have a nil key.
This entry point is used only by analysistest.