Compartir a través de


Debug only user code with Just My Code

Applies to: yesVisual Studio noVisual Studio for Mac

Note

This article applies to Visual Studio 2017. If you're looking for the latest Visual Studio documentation, see Visual Studio documentation. We recommend upgrading to the latest version of Visual Studio. Download it here

Just My Code is a Visual Studio debugging feature that automatically steps over calls to system, framework, and other non-user code. In the Call Stack window, Just My Code collapses these calls into [External Code] frames.

Just My Code works differently in .NET, C++, and JavaScript projects.

Enable or disable Just My Code

For most programming languages, Just My Code is enabled by default.

  • To enable or disable Just My Code in Visual Studio, under Tools > Options (or Debug > Options) > Debugging > General, select or deselect Enable Just My Code.

Note

Enable Just My Code is a global setting that applies to all Visual Studio projects in all languages.

Just My Code debugging

During a debugging session, the Modules window shows which code modules the debugger is treating as My Code (user code), along with their symbol loading status. For more information, see Get more familiar with how the debugger attaches to your app.

In the Call Stack or Tasks window, Just My Code collapses non-user code into a grayed-out annotated code frame labeled [External Code].

Tip

To open the Modules, Call Stack, Tasks, or most other debugging windows, you must be in a debugging session. While debugging, under Debug > Windows, select the windows you want to open.

To view the code in a collapsed [External Code] frame, right-click in the Call Stack or Task window, and select Show External Code from the context menu. The expanded external code lines replace the [External Code] frame.

Note

Show External Code is a current user profiler setting that applies to all projects in all languages that are opened by the user.

Double-clicking an expanded external code line in the Call Stack window highlights the calling code line in green in the source code. For DLLs or other modules not found or loaded, a symbol or source not found page may open.

.NET Just My Code

In .NET projects, Just My Code uses symbol (.pdb) files and program optimizations to classify user and non-user code. The .NET debugger considers optimized binaries and non-loaded .pdb files to be non-user code.

Three compiler attributes also affect what the .NET debugger considers to be user code:

The .NET debugger considers all other code to be user code.

During .NET debugging:

  • Debug > Step Into (or F11) on non-user code steps over the code to the next line of user code.
  • Debug > Step Out (or Shift+F11) on non-user code runs to the next line of user code.

If there's no more user code, debugging continues until it ends, hits another breakpoint, or throws an error.

If the debugger breaks in non-user code (for example, you use Debug > Break All and pause in non-user code), the No Source window appears. You can then use a Debug > Step command to go to the next line of user code.

If an unhandled exception occurs in non-user code, the debugger breaks at the user code line where the exception was generated.

If first chance exceptions are enabled for the exception, the calling user-code line is highlighted in green in source code. The Call Stack window displays the annotated frame labeled [External Code].

C++ Just My Code

Starting in Visual Studio 2017 version 15.8, Just My Code for code stepping is also supported. This feature also requires use of the /JMC (Just my code debugging) compiler switch. The switch is enabled by default in C++ projects. For Call Stack window and call stack support in Just My Code, the /JMC switch is not required.

To be classified as user code, the PDB for the binary containing the user code must be loaded by the debugger (use the Modules window to check this).

For call stack behavior, such as in the Call Stack window, Just My Code in C++ considers only these functions to be non-user code:

  • Functions with stripped source information in their symbols file.
  • Functions where the symbol files indicate that there is no source file corresponding to the stack frame.
  • Functions specified in *.natjmc files in the %VsInstallDirectory%\Common7\Packages\Debugger\Visualizers folder.

For code stepping behavior, Just My Code in C++ considers only these functions to be non-user code:

  • Functions for which the corresponding PDB file has not been loaded in the debugger.
  • Functions specified in *.natjmc files in the %VsInstallDirectory%\Common7\Packages\Debugger\Visualizers folder.

Note

For code stepping support in Just My Code, C++ code must be compiled using the MSVC compilers in Visual Studio 15.8 Preview 3 or later, and the /JMC compiler switch must be enabled (it is enabled by default). For additional details, see Customize C++ call stack and code stepping behavior and this blog post. For code compiled using an older compiler, .natstepfilter files are the only way to customize code stepping, which is independent of Just My Code. See Customize C++ stepping behavior.

During C++ debugging:

  • Debug > Step Into (or F11) on non-user code steps over the code to the next line of user code.
  • Debug > Step Out (or Shift+F11) on non-user code runs to the next line of user code.

If there's no more user code, debugging continues until it ends, hits another breakpoint, or throws an error.

If the debugger breaks in non-user code (for example, you use Debug > Break All and pause in non-user code), stepping continues in the non-user code.

If the debugger hits an exception, it stops on the exception, whether it is in user or non-user code. User-unhandled options in the Exception Settings dialog box are ignored.

Customize C++ call stack and code stepping behavior

For C++ projects, you can specify the modules, source files, and functions the Call Stack window treats as non-user code by specifying them in *.natjmc files. This customization also applies to code stepping if you are using the latest compiler (see C++ Just My Code).

  • To specify non-user code for all users of the Visual Studio machine, add the .natjmc file to the %VsInstallDirectory%\Common7\Packages\Debugger\Visualizers folder.
  • To specify non-user code for an individual user, add the .natjmc file to the %USERPROFILE%\My Documents\<Visual Studio version>\Visualizers folder.

A .natjmc file is an XML file with this syntax:

<?xml version="1.0" encoding="utf-8"?>
<NonUserCode xmlns="http://schemas.microsoft.com/vstudio/debugger/jmc/2015">

  <!-- Modules -->
  <Module Name="ModuleSpec" />
  <Module Name="ModuleSpec" Company="CompanyName" />

  <!-- Files -->
  <File Name="FileSpec"/>

  <!-- Functions -->
  <Function Name="FunctionSpec" />
  <Function Name="FunctionSpec" Module ="ModuleSpec" />
  <Function Name="FunctionSpec" Module ="ModuleSpec" ExceptionImplementation="true" />

</NonUserCode>

Module element attributes

Attribute Description
Name Required. The full path of the module or modules. You can use the Windows wildcard characters ? (zero or one character) and * (zero or more characters). For example,

<Module Name="?:\3rdParty\UtilLibs\*" />

tells the debugger to treat all modules in \3rdParty\UtilLibs on any drive as external code.
Company Optional. The name of the company that publishes the module that is embedded in the executable file. You can use this attribute to disambiguate the modules.

File element attributes

Attribute Description
Name Required. The full path of the source file or files to treat as external code. You can use the Windows wildcard characters ? and * when specifying the path.

Function element attributes

Attribute Description
Name Required. The fully qualified name of the function to treat as external code.
Module Optional. The name or full path to the module that contains the function. You can use this attribute to disambiguate functions with the same name.
ExceptionImplementation When set to true, the call stack displays the function that threw the exception rather than this function.

Customize C++ stepping behavior independent of Just My Code settings

In C++ projects, you can specify functions to step over by listing them as non-user code in *.natstepfilter files. Functions listed in *.natstepfilter files are not dependent on Just My Code settings.

  • To specify non-user code for all local Visual Studio users, add the .natstepfilter file to the %VsInstallDirectory%\Common7\Packages\Debugger\Visualizers folder.
  • To specify non-user code for an individual user, add the .natstepfilter file to the %USERPROFILE%\My Documents\<Visual Studio version>\Visualizers folder.

A .natstepfilter file is an XML file with this syntax:

<?xml version="1.0" encoding="utf-8"?>
<StepFilter xmlns="http://schemas.microsoft.com/vstudio/debugger/natstepfilter/2010">
    <Function>
        <Name>FunctionSpec</Name>
        <Action>StepAction</Action>
    </Function>
    <Function>
        <Name>FunctionSpec</Name>
        <Module>ModuleSpec</Module>
        <Action>StepAction</Action>
    </Function>
</StepFilter>

Element Description
Function Required. Specifies one or more functions as non-user functions.
Name Required. An ECMA-262 formatted regular expression specifying the full function name to match. For example:

<Name>MyNS::MyClass.*</Name>

tells the debugger that all methods in MyNS::MyClass are to be considered non-user code. The match is case-sensitive.
Module Optional. An ECMA-262 formatted regular expression specifying the full path to the module containing the function. The match is case-insensitive.
Action Required. One of these case-sensitive values:

NoStepInto - tells the debugger to step over the function.
StepInto - tells the debugger to step into the function, overriding any other NoStepInto for the matched function.

JavaScript Just My Code

JavaScript Just My Code controls stepping and call stack display by categorizing code in one of these classifications:

Classification Description
MyCode User code that you own and control.
LibraryCode Non-user code from libraries that you use regularly and your app relies on to function correctly (for example WinJS or jQuery).
UnrelatedCode Non-user code in your app that you don't own and your app doesn't rely on to function correctly. For example, an advertising SDK that displays ads could be UnrelatedCode. In UWP projects, any code that is loaded into your app from an HTTP or HTTPS URI is also considered UnrelatedCode.

The JavaScript debugger classifies code as user or non-user in this order:

  1. The default classifications.

    • Script executed by passing a string to the host-provided eval function is MyCode.
    • Script executed by passing a string to the Function constructor is LibraryCode.
    • Script in a framework reference, such as WinJS or the Azure SDK, is LibraryCode.
    • Script executed by passing a string to the setTimeout, setImmediate, or setInterval functions is UnrelatedCode.
  2. Classifications specified for all Visual Studio JavaScript projects in the %VSInstallDirectory%\JavaScript\JustMyCode\mycode.default.wwa.json file.

  3. Classifications in the mycode.json file of the current project.

Each classification step overrides the previous steps.

All other code is classified as MyCode.

You can modify the default classifications, and classify specific files and URLs as user or non-user code, by adding a .json file named mycode.json to the root folder of a JavaScript project. See Customize JavaScript Just My Code.

During JavaScript debugging:

  • If a function is non-user code, Debug > Step Into (or F11) behaves the same as Debug > Step Over (or F10).
  • If a step begins in non-user (LibraryCode or UnrelatedCode) code, stepping temporarily behaves as if Just My Code isn't enabled. When you step back to user code, Just My Code stepping is re-enabled.
  • When a user code step results in leaving the current execution context, the debugger stops at the next executed user code line. For example, if a callback executes in LibraryCode code, the debugger continues until the next line of user code executes.
  • Step Out (or Shift+F11) stops on the next line of user code.

If there's no more user code, debugging continues until it ends, hits another breakpoint, or throws an error.

Breakpoints set in code are always hit, but the code is classified.

  • If the debugger keyword occurs in LibraryCode, the debugger always breaks.
  • If the debugger keyword occurs in UnrelatedCode, the debugger doesn't stop.

If an unhandled exception occurs in MyCode or LibraryCode code, the debugger always breaks.

If an unhandled exception occurs in UnrelatedCode, and MyCode or LibraryCode is on the call stack, the debugger breaks.

If first-chance exceptions are enabled for the exception, and the exception occurs in LibraryCode or UnrelatedCode:

  • If the exception is handled, the debugger doesn't break.
  • If the exception is not handled, the debugger breaks.

Customize JavaScript Just My Code

To categorize user and non-user code for a single JavaScript project, you can add a .json file named mycode.json to the root folder of the project.

Specifications in this file override the default classifications and the mycode.default.wwa.json file. The mycode.json file does not need to list all key value pairs. The MyCode, Libraries, and Unrelated values can be empty arrays.

Mycode.json files use this syntax:

{
    "Eval" : "Classification",
    "Function" : "Classification",
    "ScriptBlock" : "Classification",
    "MyCode" : [
        "UrlOrFileSpec",
        . . .
        "UrlOrFileSpec"
    ],
    "Libraries" : [
        "UrlOrFileSpec",
        . .
        "UrlOrFileSpec"
    ],
    "Unrelated" : [
        "UrlOrFileSpec",
        . . .
        "UrlOrFileSpec"
    ]
}

Eval, Function, and ScriptBlock

The Eval, Function, and ScriptBlock key value pairs determine how dynamically generated code is classified:

Name Description
Eval Script that is executed by passing a string to the host-provided eval function. By default, Eval script is classified as MyCode.
Function Script that is executed by passing a string to the Function constructor. By default, Function script is classified as LibraryCode.
ScriptBlock Script that is executed by passing a string to the setTimeout, setImmediate, or setInterval functions. By default, ScriptBlock script is classified as UnrelatedCode.

You can change the value to one of these keywords:

  • MyCode classifies the script as MyCode.
  • Library classifies the script as LibraryCode.
  • Unrelated classifies the script as UnrelatedCode.

MyCode, Libraries, and Unrelated

The MyCode, Libraries, and Unrelated key value pairs specify the URLs or files that you want to include in a classification:

Name Description
MyCode An array of URLs or files that are classified as MyCode.
Libraries An array of URLs or files that are classified as LibraryCode.
Unrelated An array of URLs or files that are classified as UnrelatedCode.

The URL or file string can have one or more * characters, which match zero or more characters. * is the same as the regular expression .*.