AuthoringSink Class

Definition

public ref class AuthoringSink

[Windows::Foundation::Metadata::WebHostHidden]
class AuthoringSink

[System.CLSCompliant(false)]
public class AuthoringSink

public class AuthoringSink

[<System.CLSCompliant(false)>]
type AuthoringSink = class

type AuthoringSink = class

Public Class AuthoringSink

Inheritance

Object

AuthoringSink

Attributes

CLSCompliantAttribute

Remarks

An instantiation of this class is used extensively by the parser to remember various language elements that help with the following features:

• Error reporting.

• Brace matching.

• IntelliSense Member Selection, Complete Word, Quick Info, and Method Tips.

• Hidden regions.

• Management of contents of the Autos window in the debugger.

• Breakpoint validation.

The base class has support for all but the last two features.

Errors

The base class uses an internal ArrayList called errors to contain the errors found during parsing. This list is displayed at the end of a full parsing operation by converting each error message into a task list item by calling CreateErrorTaskItem and adding it the Error List task window. These errors are also shown in the source file itself as squiggly lines under the code that is in error. This is handled automatically by the MPF classes. The AuthoringSink class constructor takes a parameter that specifies the maximum number of errors that can be retained by the class.

Brace Matching

The base class uses an internal ArrayList called Braces to contain matching pairs of language elements (such as "{" and "}") as well as matching triples (such as "for()", "{", and "}"). A base class called BraceMatch is used for matching pairs while a derived class called TripleMatch adds information for a third language element. Calling the MatchPair and MatchTriple methods is done only when the reason for parsing is to match braces (see the BraceMatching property for details on the exact reasons). See the Example section of MatchPairfor a description of the BraceMatch class and see the Example section of MatchTriplefor a description of the TripleMatch class.

IntelliSense

IntelliSense has at least four distinct modes that can be supported by the language service. These are:

1. Member Selection: This option provides a list of members for the current scope, typically supplied after the user types a particular character. For example, if the user has entered a variable name followed by a "." then a member list for the type of that variable is displayed for selection.

2. Complete Word: This option presents a list of possible completions for a word being entered by the user.

3. Quick Info: This option presents information about an identifier. This is triggered either by the user holding the mouse cursor over the identifier or by positioning the edit caret on an identifier and selecting Quick Info from the IntelliSense menu.

4. Method Tip: This option presents parameter information to help the user while entering a method and its parameters. A tool tip is displayed showing one of possible several overloaded forms of the method along with the parameters. As the user enters each parameter, the tool tip is updated to show the parameter being typed.

All of these modes are supported by the methods StartName and QualifyName and the internal lists MethodCalls, Names, and SourceLocations. MethodCalls is used to track method signatures as they are parsed. Names and SourceLocations are used to track the method parameters while parsing a method parameter list. Note that for each name in the Names list, there is a corresponding TextSpan object in the SourceLocations list.

Hidden Regions

Hidden regions are sections of code that can optionally be hidden from view by the user. These regions are typically used as part of support for outlining where each method and class can be collapsed into a single line, making the overall class structure clearer. Some languages support specific hidden regions through special keywords. For example, C# uses #region/#endregion to bracket a region that the user normally wants hidden. A hidden region is indicated by a TextSpan object that is wrapped in a NewHiddenRegion object that in turn is stored in an internal array called hiddenRegions. Hidden regions are added by calling AddHiddenRegion.

Autos

While debugging, the Autos window can show all local variables and parameters available in a particular scope of a stack frame. The language service can support these variables by locating them during parsing. This allows the Autos window to be dynamically updated while editing during debugging. The base class AuthoringSink does not support autos so you have to derive a class from the AuthoringSink class and implement the AutoExpression method yourself.

Breakpoint Validation

When a breakpoint is placed, the location of the breakpoint is not validated until debugging actually starts and the debug engine has been loaded. If your language service supports identifying valid sections of code where a breakpoint can be placed, then this information can be used to validate the placement of breakpoints without loading the debug engine. Note that the debug engine is always the final judge when validating breakpoint locations but the language service can provide fast feedback to the user. The base class AuthoringSink does not support breakpoint validation so you have to derive a class from the AuthoringSink class and implement the CodeSpan method yourself.

Notes to Inheritors

If your language service supports expressions in the Autos window or validating breakpoints in a span of code, you need derive a class from the AuthoringSink class and override the appropriate methods (AutoExpression(TextSpan) and CodeSpan). Then override the CreateAuthoringSink(ParseReason, Int32, Int32) method in the Source class to instantiate your version of the AuthoringSink class. The AuthoringSink class is instantiated every time a parse operation is called for.

Notes to Callers

This class is accessed in several places in the Source class in order to handle brace matching, IntelliSense, and error reporting.

Tip: When implementing your class derived from AuthoringScope, add a field to your class to store the AuthoringSink object created for the ParseRequest object. This way, your AuthoringScope object can access all the information stored in the AuthoringSink object.

Constructors

Properties

Methods

Applies to