If you use the WPT test runner you are getting a bunch of JSON test results which you can't do much that's immediately, humanly useful with. There are quite a few of those in the test results repository.
The purpose of wptreport is to generate human-oriented reports for such data.
The typical installation is:
npm install wptreport
If you're hacking on the code, or want to be on the bleeding edge (which shouldn't be needed: I ship the code pretty much anytime I make a change), you can alternatively
git clone https://github.com/darobin/wptreport.git
cd wptreport
npm link .
This is a command-line tool, it takes some basic options.
wptreport [--input /path/to/dir] [--output /path/to/dir] [--spec SpecName] [-f] [-m] [-d description]
--input <directory>,-i <directory>: Path to the directory that contains all the JSON data. JSON files must match the pattern\w{2}\d{d}\.jsonwhere the two first letters are an identifier for the browser engine (use whatever makes sense to your audience) and the number is the version number. Defaults to the current directory. This is also where thefilter.jsis found, if any (see the section on filtering).--output <directory>,-o <directory>: The directory where the generated reports are stored. Defaults to the current directory.--spec SpecName,-s SpecName: The specification name to use in titling the report. Optional, but certainly looks nicer.--description DescFile,-d DescFile: Include a description of report at the top.--failures,-f: Include messages about failures in report.--markdown,-m: Interpret subtest names as markdown.--help,-h: Get some help (about the tool, you're on your own for your other issues).--version,-v: Get the version number.
If there is a wptreport.json file in the input directory, options will be read from that file
(but command line options will override those settings).
The format of the JSON files is defined by the test runner.
If the input directory contains a filter.js file, it will be loaded as a Node module (and can
therefore completely operate like one). Some test suites have experimental tests, tests that can
provide useful information in some cases that you don't want in others, etc. As such, it can be
useful to filter out some results before reporting.
The module can optionally export two methods:
excludeFile(fileName). Receives the name of the test file being looked at (as listed in the JSON data). Returntrueto skip it (and therefore all of its content),falseto keep it.excludeCase(fileName, caseName). Receives both the name of the test file and that of the test case being looked at. Skip withtrue, keep withfalse. This is the same as the previous one, only more granular.