Context

A shorthand to create an expr.

Unlike the conventional expr factory, which requires a scala.reflect.api.TreeCreator, this one accepts a regular tree, but the resulting exprs are unable of being migrated to other universes/mirrors (the functionality normally not needed for macros, since there is only one compile-time universe and only one compile-time mirror).

Indicates that an argument to c.typecheck should be typechecked as a pattern.

Indicates that an argument to c.typecheck should be typechecked as a term. This is the default typechecking mode in Scala 2.11 and the only one supported in Scala 2.10.

Indicates that an argument to c.typecheck should be typechecked as a type.

A shorthand to create a type tag.

Unlike the conventional type tag factory, which requires a scala.reflect.api.TypeCreator, this one accepts a regular type, but the resulting type tags are unable of being migrated to other universes/mirrors (the functionality normally not needed for macros, since there is only one compile-time universe and only one compile-time mirror).

A shorthand to create a weak type tag.

Unlike the conventional type tag factory, which requires a scala.reflect.api.TypeCreator, this one accepts a regular type, but the resulting type tags are unable of being migrated to other universes/mirrors (the functionality normally not needed for macros, since there is only one compile-time universe and only one compile-time mirror).

Abruptly terminates current macro expansion leaving a note about what happened. Use enclosingPosition if you're in doubt what position to pass to pos.

Exposes current classpath.

Exposes current compiler settings as a list of options. Use scalac -help, scalac -X and scalac -Y to learn about currently supported options.

For sending a message which should not be labelled as a warning/error, but also shouldn't require -verbose to be visible. Use enclosingPosition if you're in doubt what position to pass to pos.

Information about one of the currently considered implicit candidates. Candidates are used in plural form, because implicit parameters may themselves have implicit parameters, hence implicit searches can recursively trigger other implicit searches.

Can be useful to get information about an application with an implicit parameter that is materialized during current macro expansion. If we're in an implicit macro being expanded, it's included in this list.

Unlike openImplicits, this is a val, which means that it gets initialized when the context is created and always stays the same regardless of whatever happens during macro expansion.

Contexts that represent macros in-flight, including the current one. Very much like a stack trace, but for macros only. Can be useful for interoperating with other macros and for imposing compiler-friendly limits on macro expansion.

Is also priceless for emitting sane error messages for macros that are called by other macros on synthetic (i.e. position-less) trees. In that dire case navigate the enclosingMacros stack, and it will most likely contain at least one macro with a position-ful macro application. See enclosingPosition for a default implementation of this logic.

Unlike openMacros, this is a val, which means that it gets initialized when the context is created and always stays the same regardless of whatever happens during macro expansion.

Tries to guess a position for the enclosing application. But that is simple, right? Just dereference pos of macroApplication? Not really. If we're in a synthetic macro expansion (no positions), we must do our best to infer the position of something that triggered this expansion. Surprisingly, quite often we can do this by navigation the enclosingMacros stack.

Emits a compilation error. Use enclosingPosition if you're in doubt what position to pass to pos.

Takes a typed wrapper for a tree of type T and evaluates it to a value of type T.

Can be used to perform compile-time computations on macro arguments to the extent permitted by the shape of the arguments.

Known issues: because of https://issues.scala-lang.org/browse/SI-5748 trees being evaluated first need to undergo untypecheck. Resetting symbols and types mutates the tree in place, therefore the conventional approach is to duplicate the tree first.

scala> def impl(c: Context)(x: c.Expr[String]) = {
     | val x1 = c.Expr[String](c.untypecheck(x.tree.duplicate))
     | println(s"compile-time value is: ${c.eval(x1)}")
     | x
     | }
impl: (c: Context)(x: c.Expr[String])c.Expr[String]

scala> def test(x: String) = macro impl
test: (x: String)String

scala> test("x")
compile-time value is: x
res0: String = x

scala> test("x" + "y")
compile-time value is: xy
res1: String = xy

scala> val x = "x"
x: String = x

scala> test(x + "y")
compile-time value is: xy
res2: String = xy

scala> { val x = "x"; test(x + "y") }
error: exception during macro expansion:
scala.tools.reflect.ToolBoxError: reflective compilation failed

Note that in the last case evaluation has failed, because the argument of a macro refers to a runtime value x, which is unknown at compile time.

Creates a more or less unique name having a given name as a prefix and having the same flavor (term name or type name) as the given name. Consult scala.reflect.macros.Names for more information on uniqueness of such names.

Creates a string that represents a more or less unique name having a given prefix. Consult scala.reflect.macros.Names for more information on uniqueness of such names.

Creates a string that represents a more or less unique name. Consult scala.reflect.macros.Names for more information on uniqueness of such names.

Does the compilation session have any errors?

Does the compilation session have any warnings?

Infers an implicit value of the expected type pt in the macro callsite context. Optional pos parameter provides a position that will be associated with the implicit search.

If silent is false, TypecheckException will be thrown in case of an inference error. If silent is true, the typecheck is silent and will return EmptyTree if an error occurs. Such errors don't vanish and can be inspected by turning on -Xlog-implicits. Unlike in typecheck, silent is true by default.

Infers an implicit view from the provided tree tree of the type from to the type to in the macro callsite context. Optional pos parameter provides a position that will be associated with the implicit search.

If silent is false, TypecheckException will be thrown in case of an inference error. If silent is true, the typecheck is silent and will return EmptyTree if an error occurs. Such errors don't vanish and can be inspected by turning on -Xlog-implicits. Unlike in typecheck, silent is true by default.

Emits an informational message, suppressed unless -verbose or force=true. Use enclosingPosition if you're in doubt what position to pass to pos.

The tree that undergoes macro expansion. Can be useful to get an offset or a range position of the entire tree being processed.

The mirror of the compile-time universe.

Information about one of the currently considered implicit candidates. Candidates are used in plural form, because implicit parameters may themselves have implicit parameters, hence implicit searches can recursively trigger other implicit searches.

Can be useful to get information about an application with an implicit parameter that is materialized during current macro expansion. If we're in an implicit macro being expanded, it's included in this list.

Unlike enclosingImplicits, this is a def, which means that it gets recalculated on every invocation, so it might change depending on what is going on during macro expansion.

Contexts that represent macros in-flight, including the current one. Very much like a stack trace, but for macros only. Can be useful for interoperating with other macros and for imposing compiler-friendly limits on macro expansion.

Is also priceless for emitting sane error messages for macros that are called by other macros on synthetic (i.e. position-less) trees. In that dire case navigate the openMacros stack, and it will most likely contain at least one macro with a position-ful macro application. See enclosingPosition for a default implementation of this logic.

Unlike enclosingMacros, this is a def, which means that it gets recalculated on every invocation, so it might change depending on what is going on during macro expansion.

Parses a string with a Scala expression into an abstract syntax tree. Only works for expressions, i.e. parsing a package declaration will fail.

The prefix tree from which the macro is selected.

For example, for a macro filter defined as an instance method on a collection Coll, prefix represents an equivalent of this for normal instance methods:

scala> class Coll[T] {
     | def filter(p: T => Boolean): Coll[T] = macro M.filter[T]
     | }; object M {
     | def filter[T](c: Context { type PrefixType = Coll[T] })
     |              (p: c.Expr[T => Boolean]): c.Expr[Coll[T]] =
     |   {
     |     println(c.prefix.tree)
     |     c.prefix
     |   }
     | }
defined class Coll
defined module Macros

scala> new Coll[Int]().filter(_ % 2 == 0)
new Coll[Int]()
res0: Coll[Int] = ...

scala> val x = new Coll[String]()
x: Coll[String] = ...

scala> x.filter(_ != "")
$line11.$read.$iw.$iw.$iw.$iw.$iw.$iw.$iw.$iw.$iw.$iw.$iw.$iw.x
res1 @ 35563b4b: x.type = ...

Note how the value of prefix changes depending on the qualifier of the macro call (i.e. the expression that is at the left-hand side of the dot).

Another noteworthy thing about the snippet above is the Context { type PrefixType = Coll[T] } type that is used to stress that the macro implementation works with prefixes of type Coll[T].

Given a type, generate a tree that when compiled and executed produces the runtime class of the enclosing class or module. Returns EmptyTree if there does not exist an enclosing class or module.

Given a type, generate a tree that when compiled and executed produces the runtime class of the original type. If concrete is true, then this function will bail on types, who refer to abstract types (like ClassTag does).

Given a tree, generate a tree that when compiled and executed produces the original tree. For more information and examples see the documentation for Universe.reify.

The produced tree will be bound to the specified universe and mirror. Possible values for universe include universe.internal.gen.mkRuntimeUniverseRef. Possible values for mirror include EmptyTree (in that case the reifier will automatically pick an appropriate mirror).

This function is deeply connected to Universe.reify, a macro that reifies arbitrary expressions into runtime trees. They do very similar things (Universe.reify calls Context.reifyTree to implement itself), but they operate on different metalevels (see below).

Let's study the differences between Context.reifyTree and Universe.reify on an example of using them inside a fooMacro macro:

* Since reify itself is a macro, it will be executed when fooMacro is being compiled (metalevel -1) and will produce a tree that when evaluated during macro expansion of fooMacro (metalevel 0) will recreate the input tree.

This provides a facility analogous to quasi-quoting. Writing "reify{ expr }" will generate an AST that represents expr. Afterwards this AST (or its parts) can be used to construct the return value of fooMacro.

* reifyTree is evaluated during macro expansion (metalevel 0) and will produce a tree that when evaluated during the runtime of the program (metalevel 1) will recreate the input tree.

This provides a way to retain certain trees from macro expansion time to be inspected later, in the runtime. For example, DSL authors may find it useful to capture DSL snippets into ASTs that are then processed at runtime in a domain-specific way.

Also note the difference between universes of the runtime trees produced by two reifies:

* The result of compiling and running the result of reify will be bound to the Universe that called reify. This is possible because it's a macro, so it can generate whatever code it wishes.

* The result of compiling and running the result of reifyTree will be the prefix that needs to be passed explicitly. This happens because the Universe of the evaluated result is from a different metalevel than the Context the called reify.

Typical usage of this function is to retain some of the trees received/created by a macro into the form that can be inspected (via pattern matching) or compiled/run (by a reflective ToolBox) during the runtime.

Given a type, generate a tree that when compiled and executed produces the original type. The produced tree will be bound to the specified universe and mirror. For more information and examples see the documentation for Context.reifyTree and Universe.reify.

Exposes macro-specific settings as a list of strings. These settings are passed to the compiler via the "-Xmacro-settings:setting1,setting2...,settingN" command-line option.

Typechecks the provided tree against the expected type pt in the macro callsite context under typechecking mode specified in mode with EXPRmode being default. This populates symbols and types of the tree and possibly transforms it to reflect certain desugarings.

If silent is false, TypecheckException will be thrown in case of a typecheck error. If silent is true, the typecheck is silent and will return EmptyTree if an error occurs. Such errors don't vanish and can be inspected by turning on -Ymacro-debug-verbose. Unlike in inferImplicitValue and inferImplicitView, silent is false by default.

Typechecking can be steered with the following optional parameters: withImplicitViewsDisabled recursively prohibits implicit views (though, implicit vals will still be looked up and filled in), default value is false withMacrosDisabled recursively prohibits macro expansions and macro-based implicits, default value is false

The compile-time universe.

Undoes reification of a tree.

This reversion doesn't simply restore the original tree (that would lose the context of reification), but does something more involved that conforms to the following laws:

1) unreifyTree(reifyTree(tree)) != tree // unreified tree is tree + saved context // in current implementation, the result of unreify is opaque // i.e. there's no possibility to inspect underlying tree/context

2) reifyTree(unreifyTree(reifyTree(tree))) == reifyTree(tree) // the result of reifying a tree in its original context equals to // the result of reifying a tree along with its saved context

3) compileAndEval(unreifyTree(reifyTree(tree))) ~ compileAndEval(tree) // at runtime original and unreified trees are behaviorally equivalent

In the current implementation of Scala's reflection API, untyped trees (also known as parser trees or unattributed trees) are observationally different from typed trees (also known as typer trees, typechecked trees or attributed trees),

Usually, if some compiler API takes a tree, then both untyped and typed trees will do. However in some cases, only untyped or only typed trees are appropriate. For example, eval only accepts untyped trees and one can only splice typed trees inside typed trees. Therefore in the current reflection API, there is a need in functions that go back and forth between untyped and typed trees. For this we have typecheck and untypecheck.

Note that untypecheck is currently afflicted by https://issues.scala-lang.org/browse/SI-5464, which makes it sometimes corrupt trees so that they don't make sense anymore. Unfortunately, there's no workaround for that. We plan to fix this issue soon, but for now please keep it in mind.

Emits a warning. Use enclosingPosition if you're in doubt what position to pass to pos.

Tree that corresponds to the enclosing class, or EmptyTree if not applicable.

Tree that corresponds to the enclosing DefDef tree. Throws EnclosureException if there's no such enclosing tree.

Tree that corresponds to the enclosing ImplDef tree (i.e. either ClassDef or ModuleDef). Throws EnclosureException if there's no such enclosing tree.

Tree that corresponds to the enclosing method, or EmptyTree if not applicable.

Tree that corresponds to the enclosing PackageDef tree. Throws EnclosureException if there's no such enclosing tree.

Compilation run that contains this macro application.

Tree that corresponds to the enclosing Template tree. Throws EnclosureException if there's no such enclosing tree.

Compilation unit that contains this macro application.

Creates a more or less unique name having a given name as a prefix and having the same flavor (term name or type name) as the given name. Consult scala.reflect.macros.Names for more information on uniqueness of such names.

Creates a string that represents a more or less unique name having a given prefix. Consult scala.reflect.macros.Names for more information on uniqueness of such names.

Creates a string that represents a more or less unique name. Consult scala.reflect.macros.Names for more information on uniqueness of such names.

Shorthand for Literal(Constant(x: Char)) in the underlying universe.

Shorthand for Literal(Constant(x: String)) in the underlying universe.

Shorthand for Literal(Constant(x: Double)) in the underlying universe.

Shorthand for Literal(Constant(x: Float)) in the underlying universe.

Shorthand for Literal(Constant(x: Long)) in the underlying universe.

Shorthand for Literal(Constant(x: Int)) in the underlying universe.

Shorthand for Literal(Constant(x: Short)) in the underlying universe.

Shorthand for Literal(Constant(x: Byte)) in the underlying universe.

Shorthand for Literal(Constant(x: Boolean)) in the underlying universe.

Shorthand for Literal(Constant(false)) in the underlying universe.

Shorthand for Literal(Constant(null)) in the underlying universe.

Shorthand for Literal(Constant(true)) in the underlying universe.

Shorthand for Literal(Constant(())) in the underlying universe.

Recursively resets locally defined symbols and types in a given tree. WARNING: Don't use this API, go for untypecheck instead.

Test two objects for inequality.

Equivalent to x.hashCode except for boxed numeric types and null. For numerics, it returns a hash value which is consistent with value equality: if two value type instances compare as true, then ## will produce the same hash value for each of them. For null returns a hashcode where null.hashCode throws a NullPointerException.

The expression x == that is equivalent to if (x eq null) that eq null else x.equals(that).

Constructor/Extractor for Expr.

Constructor/Extractor for TypeTag.

Constructor/Extractor for WeakTypeTag.

Cast the receiver object to be of type T0.

Note that the success of a cast at runtime is modulo Scala's erasure semantics. Therefore the expression 1.asInstanceOf[String] will throw a ClassCastException at runtime, while the expression List(1).asInstanceOf[List[String]] will not. In the latter example, because the type argument is erased as part of compilation it is not possible to check whether the contents of the list are of the requested type.

Create a copy of the receiver object.

The default implementation of the clone method is platform dependent.

Tests whether the argument (that) is a reference to the receiver object (this).

The eq method implements an equivalence relation on non-null instances of AnyRef, and has three additional properties:

• It is consistent: for any non-null instances x and y of type AnyRef, multiple invocations of x.eq(y) consistently returns true or consistently returns false.
• For any non-null instance x of type AnyRef, x.eq(null) and null.eq(x) returns false.
• null.eq(null) returns true.

When overriding the equals or hashCode methods, it is important to ensure that their behavior is consistent with reference equality. Therefore, if two objects are references to each other (o1 eq o2), they should be equal to each other (o1 == o2) and they should hash to the same value (o1.hashCode == o2.hashCode).

The equality method for reference types. Default implementation delegates to eq.

See also equals in scala.Any.

Called by the garbage collector on the receiver object when there are no more references to the object.

The details of when and if the finalize method is invoked, as well as the interaction between finalize and non-local returns and exceptions, are all platform dependent.

A representation that corresponds to the dynamic class of the receiver object.

The nature of the representation is platform dependent.

The hashCode method for reference types. See hashCode in scala.Any.

Test whether the dynamic type of the receiver object is T0.

Note that the result of the test is modulo Scala's erasure semantics. Therefore the expression 1.isInstanceOf[String] will return false, while the expression List(1).isInstanceOf[List[String]] will return true. In the latter example, because the type argument is erased as part of compilation it is not possible to check whether the contents of the list are of the specified type.

Equivalent to !(this eq that).

Wakes up a single thread that is waiting on the receiver object's monitor.

Wakes up all threads that are waiting on the receiver object's monitor.

Type symbol of x as derived from a type tag.

Creates a String representation of this object. The default representation is platform dependent. On the java platform it is the concatenation of the class name, "@", and the object's hashcode in hexadecimal.

Shortcut for implicitly[TypeTag[T]].tpe

Shortcut for implicitly[TypeTag[T]]

Shortcut for implicitly[WeakTypeTag[T]].tpe

Shortcut for implicitly[WeakTypeTag[T]]