Add test automation #44

This change causes com.sun.tools to be included as a dependency only if the JDK version is 1.8. Otherwise, without this change, the build fails when run under any Java version later than Java 8.

This change causes the Maven AntRun plugin to invoke the javac and java commands with fork=true when run in a Java 8 environment. Otherwise, without this change, the build fails when run under Java 8.

This change updates the TokenizerTester code to expect its input test data to have the string "Data state" to identify PCDATA tests — rather than the string "DATA state". The test data in the html5lib-tests suite uses "Data state", so without this change, running TokenizerTester against html5lib-tests causes TokenizerTester to fail with a “Broken test data” harness failure.

This change makes TokenizerTester, TreeTester, and EncodingTester exit with status code 1 if any test cases fail. Otherwise, without this change, we won’t catch the test failures when running tests under CI.

This change makes TokenizerTester, TreeTester, and EncodingTester stop emitting the word “Success” to standard error every time a test passes. Otherwise, without this change, in test output, we end up with so many “Success” lines that the actual test failures are effectively obscured (you have to scroll back through the output/log to hunt for failures).

This change makes the sniffing limit in HtmlInputStreamReader settable. Without this change, the HtmlInputStreamReader sniffing limit is hardcoded to 1024 — and in the context of testing, that has the effect of limiting HtmlInputStreamReader to only being useful for testing expected output of the meta prescan. So this change makes it possible for HtmlInputStreamReader to also be used for testing the results for the state where the expected character encoding is not limited to what can be determined by checking the first 1024 bytes of the input stream.

This change updates EncodingTester to make it test the result for cases when the expected character encoding is not limited to what can be determined by checking only the first 1024 bytes of the input stream. Otherwise, without this change, EncodingTester is limited to only being useful for testing the output of the meta prescan.

This change makes the TokenizerTester(InputStream stream) constructor public — as the corresponding constructors for TreeTester and EncodingTester already are.

This change refines how Html5libTest handles filenames; it adds a mechanism to allow a required/expected file extension to be specified for each test type, and uses the mechanism to specify that ".test" is the required/expected extension for tokenizer tests, and that ".dat" is the required/expected extension for tree-construction test files and for encoding test files. Without this change, Html5libTest only deals correctly with the ".test" extension for tokenizer test files — but not with the ".dat" extension for tree-construction test files and encoding test files.

This change makes Html5libTest correctly handle tests in the html5lib-tests suite which have cases with so-called “double-escaped” “input” and “output” values — for example, values that contain the literals “\\u0000” and “\\uFFFD" rather than “\u0000” and “\uFFFD”.

This change replaces Path.of() calls in Html5libTest with Paths.get(). Per https://docs.oracle.com/en/java/javase/11/docs/api/java.base/java/nio/file/Path.html#of(java.net.URI) Path.of() was introduced in Java 11. So Java 8 has no Path.of(); see also https://docs.oracle.com/javase/8/docs/api/java/nio/file/Path.html We need to continue to support Java 8 for the time being. It seems Paths.get() will eventually end up being formally deprecated; by the time it finally is, we may also be able to quit supporting Java 8 — and so we can just switch to Path.of() then.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add test automation #44

Add test automation #44

Commits on Sep 4, 2020