JavaScript/TypeScript/CSS

Requirements and recommendations

This section describes requirements or recommendations regarding the machine running the scanner that are specific to the analysis of JavaScript/TypeScript/CSS. For general requirements, see General requirements on scanner environment.

Memory

A minimum of 4GB memory is recommended.

To allow the analysis to use more memory, see Slow or unresponsive analysis in the Troubleshooting section below.

Node.js

The scanner performs the analysis using the Node.js runtime environment. No Node.js installation is required if your architecture is Linux x64, Windows x64, or Apple ARM64. 

Otherwise, you will need to provide the Node.JS runtime yourself on the scanner machine. The scanner will retrieve the Node.js runtime according to the following options and in the following order:

1. The Node.js defined through the parameter sonar.nodejs.executable (absolute path to Node.js) if the runtime version is compatible.
2. The Node.js downloaded by the scanner from SonarQube Server during analysis if the detected architecture is one of the supported ones: Linux x64, Windows x64, and Apple ARM64.
3. The Node.js defined with node in the PATH if the runtime version is compatible.

If your architecture is neither Linux x64, Windows x64, nor Apple ARM64 then you must set up option 1 or 3 (to know how to set up option 1, see Analysis parameters). In that case, we recommend using the active LTS of Node.js, currently v20. Otherwise, Node.js v18.17.0 is the minimum supported version.

The scanner temporarily deploys the Node.js runtime specific to the architecture on the host machine. It considers the following possible deployment locations in order:

1. The scanner property sonar.userHome.
2. The environment variable SONAR_USER_HOME.
3. Otherwise, it creates a .sonar directory in the user's home directory.

You can also choose to skip the deployment of the embedded Node.js runtime altogether by using either:

• The scanner property sonar.scanner.skipNodeProvisioning set to true, or
• The scanner property sonar.nodejs.executable set to a Node.js runtime path.

The deployment may fail due to insufficient permissions on the location directory. If this occurs, the scanner property or environment variable must refer to a folder with adequate permissions.

File encoding

During analysis, the scanner defaults to the host file encoding. However, analyzing JavaScript and TypeScript source files requires always using the UTF-8 file encoding. If this is not the case, set the scanner property sonar.sourceEncoding to UTF-8.

TypeScript configuration

The scanner analyzes JavaScript and TypeScript using the TypeScript compiler. It leverages TypeScript's semantic model and features like type-checking to improve analysis accuracy. The scanner will use a TypeScript configuration (tsconfig.json) if it is already present in your project or transparently create one in the background if it is not available.

When the analysis starts, the scanner follows these strategies to resolve all the TSConfig files of the project:

1. It considers only TSConfig files based on the scanner property sonar.typescript.tsconfigPaths.
   1. The property expects a comma-separated list of TSConfig path patterns.
2. If not specified, it traverses the filesystem from the project root to collect all the existing TSConfig files.
   1. This operation is time-consuming and can impact the analysis. If that’s the case, and as a workaround, users can explicitly define which TSConfig files the scanner should use.
3. If none are found, it creates a single temporary TSConfig file.

Either way, the TypeScript compiler will resolve all the files that belong to a TSConfig file. However, the scanner will only analyze the files specified through the scanner property sonar.sources. Therefore, the value of this property needs to be consistent with your TypeScript configuration.

Other

If you have a community plugin for CSS analysis installed on your SonarQube Server instance, it will conflict with CSS analysis, so it should be removed.

Language-specific properties

Discover and update the JavaScript/TypeScript properties in Administration > Configuration > General Settings > Languages > JavaScript/TypeScript

Discover and update the CSS properties in Administration > Configuration > General Settings > Languages > CSS

Troubleshooting

Slow or unresponsive analysis

On a big project, more memory may need to be allocated to analyze the project. This would be manifested by analysis getting stuck and the following stack trace might appear in the logs

ERROR: Failed to get response while analyzing [file].ts
java.io.InterruptedIOException: timeout

You can use sonar.javascript.node.maxspace property to allow the analysis to use more memory. Set this property to 4096 or 8192 for big projects. This property should be set in sonar-project.properties file or on command line for scanner (with -Dsonar.javascript.node.maxspace=4096).

File encoding errors

If you encounter file encoding errors, use the sonar.sourceEncoding=UTF-8 configuration. For instructions on performing this configuration, see Analysis parameters.

Default exclusions for JS/TS

By default, analysis will exclude files from dependencies in usual directories, such as node_modules, bower_components, dist, vendor, and external. It will also ignore .d.ts files. If for some reason analysis of files in these directories is desired, it can be configured by setting sonar.javascript.exclusions property to empty value, i.e. sonar.javascript.exclusions="", or to comma separated list of paths to be excluded. This property will exclude only JavaScript/TypeScript files, while sonar.exclusions property will exclude all files. sonar.exclusions property should be preferred to configure general exclusions for the project.

By default, analysis will exclude all files whose size is greater than 1000 KB. To change this limit, set the sonar.javascript.maxFileSize property on scanner side, or change the limit in the UI (Maximum size of analyzed files field under Project Settings > General Settings > Languages > JavaScript / TypeScript (at project level) or Administration > Configuration > General Settings > Languages > JavaScript / TypeScript (at global level)).

Detection of code bundles

The analyzer will attempt to detect bundled code or generated code. This means code that was automatically transformed and optimized with tools such as Webpack and similar. We consider generated code out of scope of the analysis since developers are not able to act upon the findings in such code. Whenever generated code is detected, the analysis will print a log message: once per the whole project on INFO level, and for each file on the DEBUG level. If you want to opt-in for analyzing the generated code or in case the detection is incorrect, you can disable it by setting sonar.javascript.detectBundles=false.

Custom rules for JS/TS

Custom rules are not supported by the analyzer. As an alternative we suggest you to have a look at ESLint. It provides custom rules that you can then import thanks to the External issues feature.

Running out of memory

While analyzing a large project or file, the scanner may run out of memory. If this occurs, you will be notified with the following analysis logs:

The analysis will stop due to the Node.js process running out of memory
You can see how Node.js heap usage evolves during analysis with "sonar.javascript.node.debugMemory=true"
Try setting "sonar.javascript.node.maxspace" to a higher value to increase Node.js heap size limit
If the problem persists, please report the issue at https://community.sonarsource.com

Consider the property setting sonar.javascript.node.maxspace to a higher value depending on the host's available memory.

Large projects and monorepos

When analyzing a large project, you may encounter memory issues, such as with monorepo projects. In these cases, a possible workaround is to divide the analysis into subfolders. Given this project structure:

my-app/
├─ app1/
│  ├─ tsconfig.sonar.json
├─ app2/
│  ├─ tsconfig.sonar.json
├─ ...
├─ tsconfig.json

The default analysis will use the root tsconfig.json which may include too many files if the project is very big and creates memory issues. Splitting the project into several TSConfig files should help in that case. To do so, create intermediate tsconfig.sonar.json for each of the subfolders and use:

sonar.typescript.tsconfigPaths=my-app/app1/tsconfig.sonar.json,my-app/app2/tsconfig.sonar.json

Unavailable dependencies

In certain situations, analysis may be conducted in environments where dependencies are not available, such as with Autoscan. If possible, it is recommended to install these dependencies (e.g. npm ci) to enhance TypeScript type inference precision. If a tsconfig.json file extends external TSConfigs and cannot locate them, unexpected analysis results may occur due to potential differences in compilerOptions. In these cases, it’s advised to directly copy the essential contents of the extended TSConfigs into a custom tsconfig.sonar.json file and use it for analysis.

Unsupported compiler options

The scanner includes a recent version of the TypeScript compiler. Sometimes, a project might use new TSConfig options that are not supported by the embedded scanner version. We suggest holding off on using these options until the scanner is updated to the new version. If that's not possible, you can create a custom tsconfig.sonar.json for the analysis without using those options.

Related Pages

• Importing external issues (ESLint, TSLint, StyleLint)
• Test coverage and execution (LCOV format)
• SonarJS plugin for ESLint
• Adding coding rules