DocLinkChecker 1.21.0

Documentation link checker

This tool can be used to check references in markdown files.

Usage

DocLinkChecker -d <docs folder> [-vac] [-f <config file>]

One of these options is required. The other can be added to overwrite settings.
-d, --docfolder       Folder containing the documents.
-f, --config             Configuration file.

Other options:
-v, --verbose          Show verbose messages.
-a, --attachments   Check the .attachments folder in the root of the docfolder for unreferenced files.
-c, --cleanup          Remove all unreferenced files from the .attachments folder in the root of the docfolder. Must be used in combination with -a flag.
-t, --table               Check that tables are well formed.
--help                     Display this help screen.
--version                Display version information.

If normal return code of the tool is 0, but on error it returns 1.

Warnings, errors and verbose

The tool will first read all markdown files in the configured documentation root and parse them to extract links, headings and tables. Then it will validate tables and links. Once that is done output is written in the console. When there are errors or warnings, they are written to the output sorted by file path, line number and then column number.

The tool always outputs the version of the used tool and a summary of the result, including the exit code.

The exit codes of the tool are defined in the table below. The exit code can be used to determine if other possible actions are taken or not.

If you want to trace what the tool is doing in detail, use the -v or verbose flag to output all details of processing the files and folders.

Configuration file

A configuration file can be used for (more) settings used by the tool. Command line parameters will overwrite these settings if provided.

An initialized configuration file called docfx-companion-tools.json can be generated in the working directory by using the command:

DocLinkChecker INIT

If a docfx-companion-tools.json file already exists in the working directory, an error is given that it will not be overwritten. The main structure looks like this:

{
  "DocumentationFiles": {
    "src": "",
    "Files": [
      "**/*.md"
    ],
    "Exclude": []
  },
  "ResourceFolderNames": [
    ".attachments"
  ],
  "DocLinkChecker": {
    "RelativeLinkStrategy": "All",
    "CheckForOrphanedResources": false,
    "CleanupOrphanedResources": false,
    "ValidatePipeTableFormatting": false,
    "ValidateExternalLinks": false,
    "ConcurrencyLevel": 5,
    "MaxHttpRedirects": 20,
    "ExternalLinkDurationWarning": 3000,
    "WhitelistUrls": [
      "http://localhost"
    ]
  }
}

Files to include or exclude

The configuration file has an entry DocumentationFiles that can be partially filled in or can be used in it's full form. The full form looks like this:

  "DocumentationFiles": {
    "src": "",
    "Files": [
      "**/*.md"
    ],
    "Exclude": []
  }

These properties are used in combination with the File Globbing in .NET pattern.

The src property indicates the folder where the documentation hierarchy is located. This value can be overwritten by the -d <folder> command line parameter.

The Files property is a list of patterns of files to include in the check. If nothing is provided, **/*.md is automatically added to include all markdown files.

The Exclude property is also a list of patterns of files to exclude in the check.

Resource folder names

Resources are files that are referenced in markdown files that are not markdown files themselves. Think of images and such.

The ResourceFolderNames property is a list of names that can be used as resource folder. When resources are validated, they should be stored in a folder (or subfolder of a folder) with a name defined in this list. By default ".attachments" is defined here, which means that a resource must be something like "..../.attachments/my-image.png" or "..../.attachments/my-topic/my-image.png".

These folders are also checked when orphaned resources are detected or removed. See further down for more details.

DocLinkChecker settings

Various settings can be provided in the configuration file, specific for the DocLinkChecker. We list them in the table below.

What is validated

The tool uses markdig to extract links, tables and headings from markdown files.

Local references

Links to other markdown files in the same hierarchy are validated to exist. When a file is referenced outside of the documents root an error or warning is reported, depending on the AllowLinksOutsideDocumentsRoot setting. When local files are referenced with a full rooted path (like d:\\git\\project\\docs\\some-file.md) an error is reported that this is not allowed. Such a reference will be different on other machines.

Links to headings (like #some-heading or ./another-file.md#another-heading) are validated by the tool. If they can't be found, but the file can be found a warning is reported. The link will work to open the document (or page), but it won't locate you somewhere in that document.

Ignored links

Links like mailto: (email address) or xref: (reference link by id) are ignored and not validated at all.

Web Links

Markdown links can be used to reference web links. The tool validates links as web link when the start with:

• http or https
• ftp or ftps

When ValidateExternalLinks is set to true and the link is not excluded by the WhitelistUrls, the URL is validated to exist. The tool will report an error for a URL when:

• The URL is not found (404)
• The URL is reported as gone (410)
• The Request URI is too long (414)

Redirects (300-399) are taken as an existing URL. Other status codes (like 401 Unauthorized, 403 Forbidden, 500 Internal Server Error or 503 Service Unavailable and more) are taken as an existing resource, but URLs with those status codes will be reported as a warning.

Whitelisting Web Links

It is possible to exclude certain web links from validation. This can be done with the WhitelistUrls list. Wildcards (* and ?) can be used to defined a pattern of URLs to skip.

Examples of valid patterns are:

If no wildcards are used, the * wildcard is automatically appended to the URL. But if one or more wildcards are used this is not done.

If the link is a web URL, an internal reference (starting with a '#') an e-mail address or a reference to a folder, it's not checked. Other links are checked if they exist in the existing docs hierarchy or on local disc (for code references). Errors are written to the output mentioning the filename, the line number and position in the line. In the check we also decode the references to make sure we properly check HTML encoded strings as well (using %20 for instance).

All references are stored in a table to use in the check of the .attachments folder (with the -a flag). All files in this folder that are not referenced are marked as 'unreferenced'. If the -c flag is provided as well, the files are removed from the .attachments folder.

Table Validation

Tables (also known as Pipe Tables) can be used in markdown. Not all renderers act the same with the definitions though. To make sure that all renderers properly render the tables, table validation will check for these things:

1. Do all rows have the same amount of columns? The first row defines the width.
2. Do all rows start and end with a pipe character '|'?
3. Do all columns in the second row (separator line) have at least three dashes per column?

The tool will report each of these errors for tables when ValidatePipeTableFormatting is set to true.

Resources

Resources are files that are referenced in markdown files that are not markdown files themselves. Think of images and such. The tool only allows resources to be stored in a folder that is listed in the ResourceFolderNames setting. It can also be stored in a sub folder of that folder.

By default ".attachments" is defined in the configuration, which means that a resource must be something like "..../.attachments/my-image.png" or "..../.attachments/my-topic/my-image.png". The .... indicate that it can be anywhere in the documentation hierarchy.

When checking for orphaned resources, we check resource files in all folders (and it's sub folders) to be referenced by one of the markdown files in the documentation hierarchy. If that's not the case, the tool will report that file as orphaned. This is only done when CheckForOrphanedResources is set to true. When CleanupOrphanedResources is set to true as well, the tool will also delete the orphaned resources when no other errors are reported by the scan.