The Nu Html Checker (v.Nu) helps you catch unintended mistakes in your HTML, CSS, and SVG. It enables you to batch-check documents from the command line and from other scripts/apps, and to deploy your own instance of the checker as a service (like validator.w3.org/nu). Its source code is available, as are instructions on how to build, test, and run the code.
A Dockerfile (see Pulling the Docker image below) and npm, pip, and brew packages are also available.
It is released upstream in these formats:
vnu.jar — a portable version you can use on any system that has Java 11 or above installed
vnu.war — for deploying the checker service through a servlet container such as Tomcat
Note: The vnu.jar and vnu.war files require you to have Java 11 or above installed. The pre-compiled Linux, Windows, and macOS binaries don’t require you to have any version of Java already installed at all.
You can get the latest release or run docker run -it --rm -p 8888:8888 ghcr.io/validator/validator:latest, npm install vnu-jar, npm install --registry=https://npm.pkg.github.com @validator/vnu-jar, brew install vnu, or pip install html5validator and see the Usage and Web-based checking sections below. Or automate your document checking with a frontend such as:
pip package (for integration in Travis CI, CircleCI, CodeShip, Jekyll, Pelican, etc.)
Usage · Options · Web-based checking
vnu-runtime-image/bin/vnu OPTIONS FILES (Linux or macOS)vnu-runtime-image\bin\vnu.bat OPTIONS FILES (Windows)java -jar ~/vnu.jar OPTIONS FILES (any system with Java8+ installed)
…where FILES are the documents to check, and OPTIONS are zero or more of the following options:
--errors-only --Werror --exit-zero-always --stdout --asciiquotes --user-agent USER_AGENT --no-langdetect --no-stream --filterfile FILENAME --filterpattern PATTERN --css --skip-non-css --also-check-css --svg --skip-non-svg --also-check-svg --xml --html --skip-non-html --format gnu|xml|json|text --help --verbose --version
The Options section below provides details on each option, and the rest of this section provides some specific examples.
Note:
Throughout these examples, replace ~/vnu.jar with the actual path to that jar file on your system, and replace vnu-runtime-image/bin/vnu and vnu-runtime-image\bin\vnu.bat with the actual path to the vnu or vnu.bat program on your system — or if you add the vnu-runtime-image/bin or vnu-runtime-image\bin directory your system PATH environment variable, you can invoke the checker with just vnu.
To check one or more documents from the command line:
vnu-runtime-image/bin/vnu FILE.html FILE2.html FILE3.html...
vnu-runtime-image\bin\vnu.bat FILE.html FILE2.html FILE3.html...
java -jar ~/vnu.jar FILE.html FILE2.html FILE3.html...
Note: If you get a StackOverflowError error when invoking the checker, try adjusting the thread stack size by providing the -Xss option to java:
java -Xss512k -jar ~/vnu.jar ...
vnu-runtime-image/bin/java -Xss512k \
-m vnu/nu.validator.client.SimpleCommandLineValidator ...
To check all documents in a particular directory DIRECTORY_PATH as HTML:
java -jar ~/vnu.jar DIRECTORY_PATH
vnu-runtime-image/bin/vnu DIRECTORY_PATH
vnu-runtime-image\bin\vnu.bat DIRECTORY_PATH
Note:
The examples in this section assume you have the vnu-runtime-image/bin or vnu-runtime-image\bin directory in your system PATH environment variable. If you’re using the jar file instead, replace vnu in the examples with java -jar ~/vnu.jar.
To check all documents in a particular directory DIRECTORY_PATH as HTML, but skip any documents whose names don’t end with the extensions .html, .htm, .xhtml, or .xht:
vnu --skip-non-html DIRECTORY_PATH
To check all documents in a particular directory as CSS:
vnu --css DIRECTORY_PATH
To check all documents in a particular directory as CSS, but skip any documents whose names don’t end with the extension .css:
vnu --skip-non-css DIRECTORY_PATH
To check all documents in a particular directory, with documents whose names end in the extension .css being checked as CSS, and all other documents being checked as HTML:
vnu --also-check-css DIRECTORY_PATH
To check all documents in a particular directory as SVG:
vnu --svg DIRECTORY_PATH
To check all documents in a particular directory as SVG, but skip any documents whose names don’t end with the extension .svg:
vnu --skip-non-svg DIRECTORY_PATH
To check all documents in a particular directory, with documents whose names end in the extension .svg being checked as SVG, and all other documents being checked as HTML:
vnu --also-check-svg DIRECTORY_PATH
To check a Web document:
vnu URL example: vnu http://example.com/foo
To check standard input:
vnu - example:
echo '<!doctype html><title>...' | vnu -
echo '<!doctype html><title>...' | java -jar ~/vnu.jar -
When used from the command line as described in this section, the checker provides the following options:
Specifies whether ASCII quotation marks are substituted for Unicode smart quotation marks in messages. default: [unset; Unicode smart quotation marks are used in messages]
Specifies that only error-level messages and non-document-error messages are reported (so that warnings and info messages are not reported). default: [unset; all messages reported, including warnings & info messages]
Makes the checker exit non-zero if any warnings are encountered (even if there are no errors). default: [unset; checker exits zero if only warnings are encountered]
Makes the checker exit zero even if errors are reported for any documents. default: [unset; checker exits 1 if errors are reported for any documents]
Makes the checker report errors and warnings to stdout rather than stderr. default: [unset; checker reports errors and warnings to stderr]
Specifies a filename. Each line of the file contains either a regular expression or starts with "#" to indicate the line is a comment. Any error message or warning message that matches a regular expression in the file is filtered out (dropped/suppressed). default: [unset; checker does no message filtering]
Specifies a regular expression. Any error message or warning message that
matches the regular expression is filtered out (dropped/suppressed).
As with all other checker options, this option may only be
specified once. So to filter multiple error messages or warning messages,
you must provide a single regular expression that will match all the
messages. The typical way to do that for regular expressions is to OR
multiple patterns together using the "|" character.
default: [unset; checker does no message filtering]
Specifies the output format for reporting the results. default: "gnu" possible values: "gnu", "xml", "json", "text" [see information at URL below] https://github.com/validator/validator/wiki/Service-%C2%BB-Common-params#out
Shows detailed usage information.
Check documents as CSS but skip documents that don’t have *.css extensions. default: [unset; all documents found are checked]
Force all documents to be checked as CSS, regardless of extension. default: [unset]
Check documents as SVG but skip documents that don’t have *.svg extensions. default: [unset; all documents found are checked]
Force all documents to be checked as SVG, regardless of extension. default: [unset]
Skip documents that don’t have *.html, *.htm, *.xhtml, or *.xht extensions. default: [unset; all documents found are checked, regardless of extension]
Forces any *.xhtml or *.xht documents to be parsed using the HTML parser. default: [unset; XML parser is used for *.xhtml and *.xht documents]
Forces any *.html documents to be parsed using the XML parser. default: [unset; HTML parser is used for *.html documents]
Check CSS documents (in addition to checking HTML documents). default: [unset; no documents are checked as CSS]
Check SVG documents (in addition to checking HTML documents). default: [unset; no documents are checked as SVG]
Specifies the value of the User-Agent request header to send when
checking HTTPS/HTTP URLs.
default: "Validator.nu/LV"
Disables language detection, so that documents are not checked for missing or mislabeledhtml[lang]attributes. default: [unset; language detection &html[lang]checking are performed]
Forces all documents to be be parsed in buffered mode instead of streaming mode (causes some parse errors to be treated as non-fatal document errors instead of as fatal document errors). default: [unset; non-streamable parse errors cause fatal document errors]
Specifies "verbose" output. (Currently this just means that the names of files being checked are written to stdout.) default: [unset; output is not verbose]
Shows the checker version number.
The Nu Html Checker — along with being usable as a standalone command-line client — can be run as an HTTP service, similar to validator.w3.org/nu, for browser-based checking of HTML documents, CSS stylesheets, and SVG images over the Web. To that end, the checker is released as several separate packages:
vnu.jar for
deploying the checker as a simple self-contained service on a system with Java installed
vnu.war for
deploying the checker to a servlet container such as Tomcat
All deployments expose a REST API that enables checking of HTML documents, CSS stylesheets, and SVG images from other clients, not just web browsers. And the Linux, Windows, and macOS binaries and vnu.jar package also include a simple HTTP client that enables you to either send documents to a locally-running instance of the checker HTTP service — for fast command-line checking — or to any remote instance of the checker HTTP service running anywhere on the Web.
The
latest
releases of the Linux, Windows, and macOS binaries and vnu.jar and
vnu.war packages
are available from the validator project at github.
The following are detailed instructions on using them.
Note:
Throughout these instructions, replace ~/vnu.jar with the actual path to that jar file on your system, and replace vnu-runtime-image/bin/java and vnu-runtime-image\bin\java.exe with the actual path to the checker java or java.exe program on your system — or if you add the vnu-runtime-image/bin or vnu-runtime-image\bin directory your system PATH environment variable, you can invoke the checker with just java nu.validator.servlet.Main 8888.
Standalone · Servlet · Client · Options
To run the checker as a standalone service (using a built-in Jetty server), open a new terminal window and invoke the checker like this:
java -cp ~/vnu.jar nu.validator.servlet.Main 8888
vnu-runtime-image/bin/java nu.validator.servlet.Main 8888
vnu-runtime-image\bin\java.exe nu.validator.servlet.Main 8888
Then open
http://0.0.0.0:8888
in a browser. (To listen on a different port, replace 8888 with the port number.)
Warning: Future checker releases will bind by default to the address 127.0.0.1. Your checker deployment might become unreachable unless you use the nu.validator.servlet.bind-address system property to bind the checker to a different address:
java -cp ~/vnu.jar \
-Dnu.validator.servlet.bind-address=128.30.52.73 \
nu.validator.servlet.Main 8888
vnu-runtime-image/bin/java \
-Dnu.validator.servlet.bind-address=128.30.52.73 \
nu.validator.servlet.Main 8888
vnu-runtime-image\bin\java.exe \
-Dnu.validator.servlet.bind-address=128.30.52.73 \
nu.validator.servlet.Main 8888
When you open http://0.0.0.0:8888 (or whatever URL corresponds to the nu.validator.servlet.bind-address value you’re using), you’ll see a form similar to validator.w3.org/nu that allows you to enter the URL of an HTML document, CSS stylesheet, or SVG image, and have the results of checking that resource displayed in the browser.
Note: If you get a StackOverflowError error when using the checker, try adjusting the thread stack size by providing the -Xss option to java:
java -Xss512k -cp ~/vnu.jar nu.validator.servlet.Main 8888
vnu-runtime-image/bin/java -Xss512k -m vnu/nu.validator.servlet.Main 8888
To run the checker inside of an existing servlet container such as Apache Tomcat you will need to deploy the vnu.war file to that server following its documentation. For example, on Apache Tomcat you could do this using the Manager application or simply by copying the file to the webapps directory (since that is the default appBase setting). Typically you would see a message similar to the following in the catalina.out log file.
May 7, 2014 4:42:04 PM org.apache.catalina.startup.HostConfig deployWAR INFO: Deploying web application archive /var/lib/tomcat7/webapps/vnu.war
Assuming your servlet container is configured to receive HTTP requests sent to localhost on port 80 and the context root of this application is vnu (often the default behavior is to use the WAR file's filename as the context root unless one is explicitly specified) you should be able to access the application by connecting to http://localhost/vnu/.
Note: You may want to customize the /WEB-INF/web.xml file inside the WAR file (you can use any ZIP-handling program) to modify the servlet filter configuration. For example, if you wanted to disable the inbound-size-limit filter, you could comment out that filter like this:
<!--
<filter>
<filter-name>inbound-size-limit-filter</filter-name>
<filter-class>nu.validator.servlet.InboundSizeLimitFilter</filter-class>
</filter>
<filter-mapping>
<filter-name>inbound-size-limit-filter</filter-name>
<url-pattern>/*</url-pattern>
</filter-mapping>
-->
The checker is packaged with an HTTP client you can use from the command line to either send documents to a locally-running instance of the checker HTTP service — for fast command-line checking — or to a remote instance anywhere on the Web.
To check documents locally using the packaged HTTP client, do this:
java -cp ~/vnu.jar nu.validator.client.HttpClient FILE.html...
vnu-runtime-image/bin/java nu.validator.client.HttpClient FILE.html...
java -cp ~/vnu.jar -Dnu.validator.client.port=80 \ -Dnu.validator.client.host=html5.validator.nu \ nu.validator.client.HttpClient FILE.html...…or like this:
vnu-runtime-image/bin/java -Dnu.validator.client.port=80 \ -Dnu.validator.client.host=html5.validator.nu \ nu.validator.client.HttpClient FILE.html...
Other options are documented below.
When using the packaged HTTP client for sending documents to an instance of the checker HTTP service for checking, you can set Java system properties to control configuration options for the checker behavior.
For example, you can suppress warning-level messages and only show error-level ones by setting the value of the nu.validator.client.level system property to error, like this:
java -Dnu.validator.client.level=error \
-cp ~/vnu.jar nu.validator.client.HttpClient FILE.html...
…or like this:
vnu-runtime-image/bin/java -Dnu.validator.client.level=error \
-cp ~/vnu.jar nu.validator.client.HttpClient FILE.html...
Most of the properties listed below map to the common input parameters for the checker service, as documented at github.com/validator/validator/wiki/Service-»-Common-params.
Specifies the hostname of the checker for the client to connect to.
default: "127.0.0.1"
Specifies the hostname of the checker for the client to connect to.
default: "8888"
example: java -Dnu.validator.client.port=8080 -jar ~/vnu.jar FILE.html
Specifies the severity level of messages to report; to
suppress warning-level messages, and only show error-level ones, set
this property to "error".
default: [unset]
possible values: "error"
example: java -Dnu.validator.client.level=error -jar ~/vnu.jar FILE.html
Specifies which parser to use. default: "html"; or, for *.xhtml input files, "xml" possible values: [see information at URL below] https://github.com/validator/validator/wiki/Service-%C2%BB-Common-params#parser
Specifies the encoding of the input document. default: [unset]
Specifies the content-type of the input document. default: "text/html"; or, for *.xhtml files, "application/xhtml+xml"
Specifies the output format for messages. default: "gnu" possible values: [see information at URL below] https://github.com/validator/validator/wiki/Service-%C2%BB-Common-params#out
Specifies whether ASCII quotation marks are substituted for Unicode smart quotation marks in messages. default: "yes" possible values: "yes" or "no"
Binds the validator service to the specified IP address. default: 0.0.0.0 [causes the checker to listen on all interfaces] possible values: The IP address of any network interface example: -Dnu.validator.servlet.bind-address=127.0.0.1
Specifies the connection timeout. default: 5000 possible values: number of milliseconds example: -Dnu.validator.servlet.connection-timeout=5000
Specifies the socket timeout. default: 5000 possible values: number of milliseconds example: -Dnu.validator.servlet.socket-timeout=5000
You can pull the checker Docker image from https://ghcr.io/validator/validator in the GitHub container registry.
To pull and run the latest version of the checker:
docker run -it --rm -p 8888:8888 ghcr.io/validator/validator:latest
To pull and run a specific tag/version of the checker from the container registry — for example, the 17.11.1 version:
docker run -it --rm -p 8888:8888 ghcr.io/validator/validator:17.11.1
To bind the checker to a specific address (rather than have it listening on all interfaces):
docker run -it --rm -p 128.30.52.73:8888:8888 ghcr.io/validator/validator:latest
To make the checker run with a connection timeout and socket timeout different than the default 5 seconds, use the CONNECTION_TIMEOUT_SECONDS and SOCKET_TIMEOUT_SECONDS environment variables:
docker run -it --rm \
-e CONNECTION_TIMEOUT_SECONDS=15 \
-e SOCKET_TIMEOUT_SECONDS=15 \
-p 8888:8888 \
validator/validator
To make the checker run with particular Java system properties set, use the JAVA_TOOL_OPTIONS environment variable:
docker run -it --rm \
-e JAVA_TOOL_OPTIONS=-Dnu.validator.client.asciiquotes=yes \
-p 8888:8888 \
validator/validator
To define a service named vnu for use with docker compose, create a Compose file named docker-compose.yml (for example), with contents such as the following:
version: '2'
services:
vnu:
image: validator/validator
ports:
- "8888:8888"
network_mode: "host" #so "localhost" refers to the host machine.
Follow the steps below to build, test, and run the checker such that you can open http://0.0.0.0:8888/ in a Web browser to use the checker Web UI.
JAVA_HOME environment variable:
export JAVA_HOME=/usr/lib/jvm/java-8-openjdk-amd64 <-- Ubuntu, etc.
export JAVA_HOME=$(/usr/libexec/java_home) <-- MacOS
git clone https://github.com/validator/validator.git
cd validator
python ./checker.py all
The first time you run the checker Python script, you’ll need to be online and the build will need time to download several megabytes of dependencies.
The steps above will build, test, and run the checker such that you can open http://0.0.0.0:8888/ in a Web browser to use the checker Web UI.
Warning: Future checker releases will bind by default to the address 127.0.0.1. Your checker deployment might become unreachable unless you use the --bind-address option to bind the checker to a different address:
python ./checker.py --bind-address=128.30.52.73 all
Use python ./checker.py --help to see command-line options for controlling the behavior of the script, as well as build-target names you can call separately; e.g.:
python ./checker.py build # to build only
python ./checker.py build test # to build and test
python ./checker.py run # to run only
python ./checker.py jar # to compile vnu.jar
python ./checker.py update-shallow && \
python ./checker.py dldeps && \
python ./checker.py jar # to compile vnu.jar faster