dev.md

Writing an analysis.

Lya provides the following hooks.

• sourceTransform(src): Apply a source transformation to the loaded library. Example analyses: uncomment.

  • src: String representation of the library
  • Expected return: a script representation of the library

• onImport(caller, callee, name): Each time we load a library this hook is called. Usefull for dependency graphs. Analysis example: print-require

  • caller: the path of the module that loads the module
  • calle: the path of the loaded library
  • name: the name of the library

• onRead(target, name, ...): Each time we read a value from an imported module this function is called. Example Analysis: read-write-execute

  • target: The target object on which we get the property.
  • name: The name of the property
  • nameToStore: The recommended name to use by Lya.
  • currentModule: The name of the module that calls the function
  • typeClass: Indicates which category the analysis belongs to (user-global, es-globals, node-globals etc)

• onCallPre(target, thisArg, ...): Before the execution of a function this hook is called. Example analysis: profiling-relative

  • target: The target function to call.
  • thisArg: The value of this provided for the call to target.
  • argumentsList: An array-like object specifying the arguments with which target will be called.
  • name: The name of the function
  • nameToStore: Recommended name to use for storing the result of the analysis
  • currentModule: The name of the module that calls the function
  • declareModule: The name of the module where we declared the target function
  • typeClass: Indicates which category the analysis belongs to (user-global, es-globals, node-globals etc)

• onCallPost(target, thisArg, ...): After the execution of a function this hooked is called. Extremly usefull for type analysis or time analysis. Example Analysis: simple-types

  • target: The target function to call.
  • thisArg: The value of this provided for the call to target.
  • argumentsList: An array-like object specifying the arguments with which target will be called.
  • name: The name of the function
  • nameToStore: Recommended name to use for storing the result of the analysis
  • currentModule: The name of the module that calls the function
  • declareModule: The name of the module where we declared the target function
  • typeClass: Indicates which category the analysis belongs to (user-global, es-globals, node-globals etc)
  • result: It contains the result of the target function execution.

• onWrite(target, name, value, ...): This hook is called is called before we set a property on a object. Example Analysis: read-write-execute

  • target: The target object on which to set the property
  • name: The name of the property to set
  • value: The value to set
  • declareModule: The name of the module where we declared the target function
  • parentName: The name of the parent object
  • nameToStore: Recommended name to use for storing the result of the analysis

• onConstruct(target, args, currentName, nameToStore): We call this hook before the execution of a constructor.

  • target: The target function to call
  • args: An array-like object specifying the arguments with which target should be called
  • currentName: The name of the module where the constructor is called
  • nameToStore: Recommended name to use for storing the result of the analysis

• onHas(target, prop, currentName, nameToStore): This method is called before you check if a property is in an object.

  • target: The target object in which to look for the property
  • prop: The property we look for
  • currentName: The name of the module where the constructor is called
  • nameToStore: Recommended name to use for storing the result of the analysis

• onExit(data): Last call before program exit -- commonly used for flushing results. Example analyses: coarse-types.

  • data:
  • Expected return: None.

Lya provides the following utility function:

const getObjectInfo = (obj) => {
      name: objName,
      path: objPath,
}