🛑 Read carefully before you jump to conclusions on this page!

There are easy ways to configure TypeScript to ensure faster compilations and editing experiences.The earlier that these practices can be adopted, the better.Beyond best-practices, there are some common techniques for investigating slow compilations/editing experiences, some common fixes, and some common ways of helping the TypeScript team investigate the issues as a last resort.

• Writing Easy-to-Compile Code
  • Preferring Interfaces Over Intersections
  • Using Type Annotations
  • Preferring Base Types Over Unions
  • Naming Complex Types
• Using Project References
  • New Code
  • Existing Code
  • Performance Considerations
• Configuringtsconfig.json orjsconfig.json
  • Specifying Files
  • Controlling@types Inclusion
  • Incremental Project Emit
  • Skipping.d.ts Checking
  • Using Faster Variance Checks
• Configuring Other Build Tools
  • Concurrent Type-Checking
  • Isolated File Emit
• Investigating Issues
  • Disabling Editor Plugins
  • extendedDiagnostics
  • showConfig
  • listFilesOnly
  • explainFiles
  • traceResolution
  • Runningtsc Alone
  • Upgrading Dependencies
  • Performance Tracing
• Common Issues
  • Misconfiguredinclude andexclude
• Filing an Issue
  • Reporting Compiler Perf Issues
    • Providing Performance Traces
    • Profiling the Compiler
    • Profiling the Compiler with pprof
  • Reporting Editing Perf Issues
    • Taking a TSServer Log
      • Collecting a TSServer Log in Visual Studio Code

Note that the following is not a bullet-proof set of rules.There may be exceptions to each rule depending on your codebase.

Much of the time, a simple type alias to an object type acts very similarly to an interface.

interfaceFoo{prop:string}typeBar={prop:string};

However, and as soon as you need to compose two or more types, you have the option of extending those types with an interface, or intersecting them in a type alias, and that's when the differences start to matter.

Interfaces create a single flat object type that detects property conflicts, which are usually important to resolve!Intersections on the other hand just recursively merge properties, and in some cases producenever.Interfaces also display consistently better, whereas type aliases to intersections can't be displayed in part of other intersections.Type relationships between interfaces are also cached, as opposed to intersection types as a whole.A final noteworthy difference is that when checking against a target intersection type, every constituent is checked before checking against the "effective"/"flattened" type.

For this reason, extending types withinterfaces/extends is suggested over creating intersection types.

- type Foo = Bar & Baz & {-     someProp: string;- }+ interface Foo extends Bar, Baz {+     someProp: string;+ }

Adding type annotations, especially return types, can save the compiler a lot of work.In part, this is because named types tend to be more compact than anonymous types (which the compiler might infer), which reduces the amount of time spent reading and writing declaration files (e.g. for incremental builds).Type inference is very convenient, so there's no need to do this universally - however, it can be a useful thing to try if you've identified a slow section of your code.

- import { otherFunc } from "other";+ import { otherFunc, OtherType } from "other";- export function func() {+ export function func(): OtherType {      return otherFunc();  }

// foo.tsexportinterfaceResult{headers:any;body:string;}exportasyncfunctionmakeRequest():Promise<Result>{thrownewError("unimplemented");}// bar.tsimport{makeRequest}from"./foo";exportfunctiondoStuff(){returnmakeRequest();}

// foo.d.tsexportinterfaceResult{headers:any;body:string;}exportdeclarefunctionmakeRequest():Promise<Result>;// bar.d.tsexportdeclarefunctiondoStuff():Promise<import("./foo").Result>;

Union types are great - they let you express the range of possible values for a type.

interfaceWeekdaySchedule{day:"Monday"|"Tuesday"|"Wednesday"|"Thursday"|"Friday";wake:Time;startWork:Time;endWork:Time;sleep:Time;}interfaceWeekendSchedule{day:"Saturday"|"Sunday";wake:Time;familyMeal:Time;sleep:Time;}declarefunctionprintSchedule(schedule:WeekdaySchedule|WeekendSchedule);

However, they also come with a cost.Every time an argument is passed toprintSchedule, it has to be compared to each element of the union.For a two-element union, this is trivial and inexpensive.However, if your union has more than a dozen elements, it can cause real problems in compilation speed.For instance, to eliminate redundant members from a union, the elements have to be compared pairwise, which is quadratic.This sort of check might occur when intersecting large unions, where intersecting over each union member can result in enormous types that then need to be reduced.One way to avoid this is to use subtypes, rather than unions.

interfaceSchedule{day:"Monday"|"Tuesday"|"Wednesday"|"Thursday"|"Friday"|"Saturday"|"Sunday";wake:Time;sleep:Time;}interfaceWeekdayScheduleextendsSchedule{day:"Monday"|"Tuesday"|"Wednesday"|"Thursday"|"Friday";startWork:Time;endWork:Time;}interfaceWeekendScheduleextendsSchedule{day:"Saturday"|"Sunday";familyMeal:Time;}declarefunctionprintSchedule(schedule:Schedule);

A more realistic example of this might come up when trying to model every built-in DOM element type.In this case, it would be preferable to create a baseHtmlElement type with common members whichDivElement,ImgElement, etc. all extend from, rather than to create an exhaustive union likeDivElement | /*...*/ | ImgElement | /*...*/.

Complex types can be written anywhere a type annotation is allowed.

interfaceSomeType<T>{foo<U>(x:U):UextendsTypeA<T> ?ProcessTypeA<U,T> :UextendsTypeB<T> ?ProcessTypeB<U,T> :UextendsTypeC<T> ?ProcessTypeC<U,T> :U;}

This is convenient, but today, every timefoo is called, TypeScript has to re-run the conditional type.What's more, relating any two instances ofSomeType requires re-relating the structure of the return type offoo.

If the return type in this example was extracted out to a type alias, more information can be cached by the compiler:

typeFooResult<U,T>=UextendsTypeA<T> ?ProcessTypeA<U,T> :UextendsTypeB<T> ?ProcessTypeB<U,T> :UextendsTypeC<T> ?ProcessTypeC<U,T> :U;interfaceSomeType<T>{foo<U>(x:U):FooResult<U,T>;}

When building up any codebase of a non-trivial size with TypeScript, it is helpful to organize the codebase into several independentprojects.Each project has its owntsconfig.json that has dependencies on other projects.This can be helpful to avoid loading too many files in a single compilation, and also makes certain codebase layout strategies easier to put together.

There are some very basic ways oforganizing a codebase into projects.As an example, one might be a program with a project for the client, a project for the server, and a project that's shared between the two.

              ------------              |          |              |  Shared  |              ^----------^             /            \            /              \------------                ------------|          |                |          ||  Client  |                |  Server  |-----^------                ------^-----

Tests can also be broken into their own project.

              ------------              |          |              |  Shared  |              ^-----^----^             /      |     \            /       |      \------------  ------------  ------------|          |  |  Shared  |  |          ||  Client  |  |  Tests   |  |  Server  |-----^------  ------------  ------^-----     |                            |     |                            |------------                ------------|  Client  |                |  Server  ||  Tests   |                |  Tests   |------------                ------------

You canread up more about project references here.

When a workspace becomes so large that it's hard for the editor to handle (and you've usedperformance tracing to confirm that there are no hotspots, making scale the most likely culprit), it can be helpful to break it down into a collection of projects that reference each other.If you're working in a monorepo, this can be as simple as creating a project for each package and mirroring the package dependency graph in project references.Otherwise the process is more ad hoc - you may be able to follow the directory structure or you may have to use carefully choseninclude andexclude globs.Some things to keep in mind:

• Aim for evenly-sized projects - avoid having a single humongous project with lots of tiny satellites
• Try to group together files that will be edited together - this will limit the number of projects the editor needs to load
• Separating out test code can help prevent product code from accidentally depending on it

As with any encapsulation mechanism, projects come with a cost.For example, if all projects depend on the same packages (e.g. a popular UI framework), some parts of that package's types will be checked repeatedly - once for each project consuming them.Empirically, it seems that (for a workspace with more than one project) 5-20 projects is an appropriate range - fewer may result in editor slowdowns and more may result in excessive overhead.Some good reasons to split out a project:

• It has a different output location (e.g. because it's a package in a monorepo)
• It requires different settings (e.g.lib ormoduleResolution)
• It contains global declarations that you want to scope (either for encapsulation or to limit expensive global rebuilds)
• The editor's language service runs out of memory when trying to process the code as a single project
  • In this case, you will want to set"disableReferencedProjectLoad": true and"disableSolutionSearching": true to limit project loading while editing

TypeScript and JavaScript users can always configure their compilations with atsconfig.json file.jsconfig.json files can also be used to configure the editing experience for JavaScript users.

You should always make sure that your configuration files aren't including too many files at once.

Within atsconfig.json, there are two ways to specify files in a project.

• thefiles list
• theinclude andexclude lists

The primary difference between the two is thatfiles expects a list of file paths to source files, andinclude/exclude use globbing patterns to match against files.

While specifyingfiles will allow TypeScript to quickly load up files up directly, it can be cumbersome if you have many files in your project without just a few top-level entry-points.Additionally, it's easy to forget to add new files to yourtsconfig.json, which means that you might end up with strange editor behavior where those new files are incorrectly analyzed.All this can be cumbersome.

include/exclude help avoid needing to specify these files, but at a cost: files must be discovered by walking through included directories.When running through alot of folders, this can slow compilations down.Additionally, sometimes a compilation will include lots of unnecessary.d.ts files and test files, which can increase compilation time and memory overhead.Finally, whileexclude has some reasonable defaults, certain configurations like mono-repos mean that a "heavy" folders likenode_modules can still end up being included.

For best practices, we recommend the following:

• Specify only input folders in your project (i.e. folders whose source code you want to include for compilation/analysis).
• Don't mix source files from other projects in the same folder.
• If keeping tests in the same folder as other source files, give them a distinct name so they can easily be excluded.
• Avoid large build artifacts and dependency folders likenode_modules in source directories.

Note: without anexclude list,node_modules is excluded by default;as soon as one is added, it's important to explicitly addnode_modules to the list.

Here is a reasonabletsconfig.json that demonstrates this in action.

{"compilerOptions":{// ...},"include":["src"],"exclude":["**/node_modules","**/.*/"],}

By default, TypeScript automatically includes every@types package that it finds in yournode_modules folder, regardless of whether you import it.This is meant to make certain things "just work" when using Node.js, Jasmine, Mocha, Chai, etc. since these tools/packages aren't imported - they're just loaded into the global environment.

Sometimes this logic can slow down program construction time in both compilation and editing scenarios, and it can even cause issues with multiple global packages with conflicting declarations, causing errors like

Duplicate identifier 'IteratorResult'.Duplicate identifier 'it'.Duplicate identifier 'define'.Duplicate identifier 'require'.

In cases where no global package is required, the fix is as easy as specifying an empty field forthe"types" option in atsconfig.json/jsconfig.json

// src/tsconfig.json{"compilerOptions":{// ...// Don't automatically include anything.// Only include `@types` packages that we need to import."types" :[]},"files":["foo.ts"]}

If you still need a few global packages, add them to thetypes field.

// tests/tsconfig.json{"compilerOptions":{// ...// Only include `@types/node` and `@types/mocha`."types" :["node","mocha"]},"files":["foo.test.ts"]}

The--incremental flag allows TypeScript to save state from the last compilation to a.tsbuildinfo file.This file is used to figure out the smallest set of files that might to be re-checked/re-emitted since it last ran, much like how TypeScript's--watch mode works.

Incremental compiles are enabled by default when using thecomposite flag for project references, but can bring the same speed-ups for any project that opts in.

By default, TypeScript performs a full re-check of all.d.ts files in a project to find issues and inconsistencies; however, this is typically unnecessary.Most of the time, the.d.ts files are known to already work - the way that types extend each other was already verified once, and declarations that matter will be checked anyway.

TypeScript provides the option to skip type-checking of the.d.ts files that it ships with (e.g.lib.d.ts) using theskipDefaultLibCheck flag.

Alternatively, you can also enable theskipLibCheck flag to skip checkingall.d.ts files in a compilation.

These two options can often hide misconfiguration and conflicts in.d.ts files, so we suggest using themonly for faster builds.

Is a list of dogs a list of animals?That is, isList<Dog> assignable toList<Animals>?The straightforward way to find out is to do a structural comparison of the types, member by member.Unfortunately, this can be very expensive.However, if we know enough aboutList<T>, we can reduce this assignability check to determining whetherDog is assignable toAnimal (i.e. without considering each member ofList<T>).(In particular, we need to know thevariance of the type parameterT.)The compiler can only take full advantage of this potential speedup if thestrictFunctionTypes flag is enabled (otherwise, it uses the slower, but more lenient, structural check).For this reason, we recommend building with--strictFunctionTypes (which is enabled by default under--strict).

TypeScript compilation is often performed with other build tools in mind - especially when writing web apps that might involve a bundler.While we can only make suggestions for a few build tools, ideally these techniques can be generalized.

Make sure that in addition to reading this section, you read up about performance in your choice of build tool - for example:

• ts-loader's section onFaster Builds
• awesome-typescript-loader's section onPerformance Issues

Type-checking typically requires information from other files, and can be relatively expensive compared to other steps like transforming/emitting code.Because type-checking can take a little bit longer, it can impact the inner development loop - in other words, you might experience a longer edit/compile/run cycle, and this might be frustrating.

For this reason, some build tools can run type-checking in a separate process without blocking emit.While this means that invalid code can run before TypeScript reports an error in your build tool, you'll often see errors in your editor first, and you won't be blocked for as long from running working code.

An example of this in action is thefork-ts-checker-webpack-plugin plugin for Webpack, orawesome-typescript-loader which also sometimes does this.

By default, TypeScript's emit requires semantic information that might not be local to a file.This is to understand how to emit features likeconst enums andnamespaces.But needing to checkother files to generate the output for an arbitrary file can make emit slower.

The need for features that need non-local information is somewhat rare - regularenums can be used in place ofconst enums, and modules can be used instead ofnamespaces.For that reason, TypeScript provides theisolatedModules flag to error on features powered by non-local information.EnablingisolatedModules means that your codebase is safe for tools that use TypeScript APIs liketranspileModule or alternative compilers like Babel.

As an example, the following code won't properly work at runtime with isolated file transforms becauseconst enum values are expected to be inlined; but luckily,isolatedModules will tell us that early on.

// ./src/fileA.tsexportdeclareconstenumE{A=0,B=1,}// ./src/fileB.tsimport{E}from"./fileA";console.log(E.A);//          ~// error: Cannot access ambient const enums when the '--isolatedModules' flag is provided.

Remember:isolatedModules doesn't automatically make code generation faster - it just tells you when you're about to use a feature that might not be supported.The thing you're looking for is isolated module emit in different build tools and APIs.

Isolated file emit can be leveraged by using the following tools:

• ts-loader providesatranspileOnly flag which performs isolated file emit by usingtranspileModule.
• awesome-typescript-loader providesatranspileOnly flag which performs isolated file emit by usingtranspileModule.
• TypeScript'stranspileModule API can be used directly.
• awesome-typescript-loader provides theuseBabel flag.
• babel-loader compiles files in an isolated manner (but does not provide type-checking on its own).
• gulp-typescript enables isolated file emit whenisolatedModules is enabled.
• rollup-plugin-typescriptonly performs isolated file compilation.
• ts-jest can use be configured with the [isolatedModules flag set totrue]isolatedModules: true(.
• ts-node can detectthe"transpileOnly" option in the"ts-node" field of atsconfig.json, and also has a--transpile-only flag.

In-editor diagnostics are typically fetched a few seconds after typing stops.ts-server's performance characteristics will always be related to the performance of type-checking your entire project usingtsc, so the other performance optimization guidance here also applies to improving the editing experience.As you type, the checker is completely starting from scratch, but it only requests information about what you're typing.This means that the editing experience can vary based on how much work TypeScript needs to do to check the type of what you are actively editing.In most editors, like VS Code, diagnostics are requested for all open files, not the entire project.Accordingly, diagnostics will appear faster compared to checking the entire project withtsc, but slower than viewing a type with hover, since viewing a type with hoveronly asks TypeScript to compute and check that specific type.

There are certain ways to get hints of what might be going wrong.

Editor experiences can be impacted by plugins.Try disabling plugins (especially JavaScript/TypeScript-related plugins) to see if that fixes any issues in performance and responsiveness.

Certain editors also have their own troubleshooting guides for performance, so consider reading up on them.For example, Visual Studio Code has its own page forPerformance Issues as well.

You can run TypeScript with--extendedDiagnostics to get a printout of where the compiler is spending its time.

Files:                         6Lines:                     24906Nodes:                    112200Identifiers:               41097Symbols:                   27972Types:                      8298Memory used:              77984KAssignability cache size:  33123Identity cache size:           2Subtype cache size:            0I/O Read time:             0.01sParse time:                0.44sProgram time:              0.45sBind time:                 0.21sCheck time:                1.07stransformTime time:        0.01scommentTime time:          0.00sI/O Write time:            0.00sprintTime time:            0.01sEmit time:                 0.01sTotal time:                1.75s

Note thatTotal time won't be the sum of all times preceding it, since there is some overlap and some work is not instrumented.

The most relevant information for most users is:

Things that you might want to ask given this input:

• Does the number of files/number of lines of code roughly correspond to the number of files in your project? Try running--listFiles if not.
• DoesProgram time orI/O Read time seem fairly high?Ensure yourinclude/exclude settings are configured correctly.
• Do other times seem off?You might want to file an issue! Things you can do to help diagnose it might be
  • Running withemitDeclarationOnly ifprintTime is high.
  • Read up instructions onReporting Compiler Performance Issues.

It's not always obvious what settings a compilation is being run with when runningtsc, especially given thattsconfig.jsons can extend other configuration files.showConfig can explain whattsc will calculate for an invocation.

tsc --showConfig# or to select a specific config file...tsc --showConfig -p tsconfig.json

Sometimes you might be surprised to find out TypeScript is reading more files than it should be - but which files is it actually reading?listFilesOnly can tell you.

tsc --listFilesOnly

Note:--listFiles is a somewhat-deprecated version of this flag. It is usually less desirable because--listFiles will still perform a full compilation, whereas--listFilesOnly will quit as soon as it manages to find every file that a compilation would need.

Running withexplainFiles can help explainwhy a file was included in a compilation.The emit is somewhat verbose, so you might want to redirect output to a file.

tsc --explainFiles> explanations.txt

If you find a file that shouldn't be present, you may need to look intofixing up yourinclude/exclude lists in yourtsconfig.json, or alternatively, you might need to adjust other settings liketypes,typeRoots, orpaths.

WhileexplainFiles can point out how a file made its way into your program,traceResolution can help diagnose the precise steps that were taken in resolving an import path.The emit is somewhat verbose, so you might want to redirect output to a file.

tsc --traceResolution> resolutions.txt

You might find that there are issues with yourmodule/moduleResolution settings, or even that your dependencies'package.json files are not configured correctly.

Much of the time, users run into slow performance using 3rd party build tools like Gulp, Rollup, Webpack, etc.Running withtsc --extendedDiagnostics to find major discrepancies between using TypeScript and the tool can indicate external misconfiguration or inefficiencies.

Some questions to keep in mind:

• Is there a major difference in build times betweentsc and the build tool with TypeScript integration?
• If the build tool provides diagnostics, is there a difference between TypeScript's resolution and the build tool's?
• Does the build tool haveits own configuration that could be the cause?
• Does the build tool have configurationfor its TypeScript integration that could be the cause? (e.g. options for ts-loader?)

Sometimes TypeScript's type-checking can be impacted by computationally intensive.d.ts files.This is rare, but can happen.Upgrading to a newer version of TypeScript (which can be more efficient) or to a newer version of an@types package (which may have reverted a regression) can often solve the issue.

In some cases, the approaches above might not give you enough insight to understand why TypeScript feels slow.TypeScript 4.1 and higher provides a--generateTrace option that can give you a sense of the work the compiler is spending time on.--generateTrace will create an output file that can be analyzed by the@typescript/analyze-trace utility, or within Edge or Chrome.

Ideally, TypeScript will be able to compile your project without any errors, though it's not a strict requirement for tracing.

Once you're ready to get a trace, you can run TypeScript with the--generateTrace flag.

tsc -p ./some/project/src/tsconfig.json --generateTrace tracing_output_folder

In some cases, you can also take a trace from your editor. In Visual Studio Code, that can be toggled by settingTypeScript > Tsserver: Enable Tracing in the UI or adding the following JSON setting:

"typescript.tsserver.enableTracing":true,

To quickly list performance hot-spots, you can install and run@typescript/analyze-trace from npm.

Alternatively, you can review the details manually:

1. Visitabout://tracing on Edge/Chrome,
2. Click on theLoad button at the top left,
3. Open the generated JSON file (trace.*.json) in your output directory.

Note that, even if your build doesn't directly invoketsc (e.g. because you use a bundler) or the slowdown you're seeing is in the editor, collecting and interpreting a trace fromtsc will generally be much easier than trying to investigate your slowdown directly and will still produce representative results.

You canread more about performance tracing in more detail here.

Once you've trouble-shooted, you might want to explore some fixes to common issues.If the following solutions don't work, it may be worthfiling an issue.

As mentioned above, theinclude/exclude options can be misused in several ways.

If your project is already properly and optimally configured, you may want tofile an issue.

The best reports of performance issues containeasily obtainable andminimal reproductions of the problem.In other words, a codebase that can easily be cloned over git that contains only a few files.They require either no external integration with build tools - they can either be invoked viatsc or use isolated code which consumes the TypeScript API.Codebases that require complex invocations and setups cannot be prioritized.

We understand that this is not always easy to achieve - specifically, because it is hard to isolate the source of a problem within a codebase, and because sharing intellectual property may be an issue.In some cases, the team will be willing to send a non-disclosure agreement (NDA) if we believe the issue is highly impactful.

Regardless of whether a reproduction is possible, following these directions when filing issues will help us provide you with performance fixes.

Sometimes you'll witness performance issues in both build times as well as editing scenarios.In these cases, it's best to focus on the TypeScript compiler.

First, a nightly version of TypeScript should be used to ensure you're not hitting a resolved issue:

npm install --save-dev typescript@next# oryarn add typescript@next --dev

A compiler perf issue should include

• The version of TypeScript that was installed (i.e.npx tsc -v oryarn tsc -v)
• The version of Node on which TypeScript ran (i.e.node -v)
• The output of running withextendedDiagnostics (tsc --extendedDiagnostics -p tsconfig.json)
• Ideally, a project that demonstrates the issues being encountered.
• Output logs from profiling the compiler (isolate-*-*-*.log and*.cpuprofile files)

Performance traces are meant to help teams figure out build performance issues in their own codebases;however, they can also be useful for the TypeScript team in diagnosing and fixing issues.See the above section onperformance traces and continue reading more on our dedicatedperformance tracing page.

You can provide the team with diagnostic traces by runningdexnode alongside TypeScript with the--generateCpuProfile flag:

npmexec dexnode -- ./node_modules/typescript/lib/tsc.js --generateCpuProfile profile.cpuprofile -p tsconfig.json

Here./node_modules/typescript/lib/tsc.js can be replaced with any path to where your version of the TypeScript compiler is installed, andtsconfig.json can be any TypeScript configuration file.profile.cpuprofile is an output file of your choice.

This will generate two files:

• dexnode will emit to a file of theisolate-*-*-*.log (e.g.isolate-00000176DB2DF130-17676-v8.log).
• --generateCpuProfile will emit to a file with the name of your choice. In the above example, it will be a file namedprofile.cpuprofile.

pprof is a helpful utility for visualizing CPU and memory profiles.pprof has different visualization modes that may make problem areas more obvious, and its profiles tend to be smaller than those produced from--generateCpuProfile.

The easiest way to generate a profile for pprof is to usepprof-it.There aredifferent ways to use pprof-it, but a quick way is to use npx or a similar tool:

npx pprof-it ./node_modules/typescript/lib/tsc.js ...

You can also install it locally:

npm install --no-save pprof-it

and run certain build scripts via npm, npx, and similar tools with the--node-option flag:

npm --node-option="--require pprof-it" run<your-script-name>

To actually view the generated profile withpprof, the Go toolset is required at minimum, and Graphviz is required for certain visualization capabilities.See more here.

Alternatively, you can useSpeedScope directly from your browser.

Perceived editing performance is frequently impacted by a number of things, and the only thing within the TypeScript team's control is the performance of the JavaScript/TypeScript language service, as well as the integration between that language service and certain editors (i.e. Visual Studio, Visual Studio Code, Visual Studio for Mac, and Sublime Text).Ensure that all 3rd-party plugins are turned off in your editor to determine whether there is an issue with TypeScript itself.

Editing performance issues are slightly more involved, but the same ideas apply: clone-able minimal repro codebases are ideal, and though in some cases the team will be able to sign an NDA to investigate and isolate issues.

Including the output fromtsc --extendedDiagnostics is always good context, but taking a TSServer trace is the most helpful.

1. Open up your command palette and either
   • open your global settings by enteringPreferences: Open User Settings
   • open your local project by enteringPreferences: Open Workspace Settings
2. Set the option"typescript.tsserver.log": "verbose",
3. Restart VS Code and reproduce the problem
4. In VS Code, run theTypeScript: Open TS Server log command
5. This should open thetsserver.log file.