RFR: 8349532: Refactor ./util/Pem/encoding.sh to run in java

[Fernando Guallini]

On Thu, 6 Feb 2025 12:12:35 GMT, Mikhail Yankelevich <duke at openjdk.org> wrote:

> * ./util/Pem/encoding.sh is now a java test
> * also added the validations for the test results, not only an error check

test/jdk/sun/security/util/Pem/PemEncoding.java line 47:

> 45:                 File.separator,
> 46:                 File.separator,
> 47:                 File.separator);

nit: the following may be easier to read

Suggestion:

        final var certPath = Path.of(System.getProperty("test.src", "."))
            .getParent()
            .resolve("HostnameChecker")
            .resolve("cert5.crt");

test/jdk/sun/security/util/Pem/PemEncoding.java line 58:

> 56:         final var subProcessOutput = result.getStdout();
> 57:         final var subProcessErr = result.getStderr();
> 58:         System.out.printf("Output:\n%s%n", subProcessOutput);

No need to print the output here, JTreg automatically creates a .log file in the scratch directory with the subprocess stderr and stdout for debugging

test/jdk/sun/security/util/Pem/PemEncoding.java line 72:

> 70:         // Some systems may automatically convert non-ASCII to ?
> 71:         // so this will detect the signature in all possible scenarios
> 72:         final var isUTF16Used = matcher.find() || subProcessOutput.contains("??: ?");

IMO, this is confusing. **isUTF16Used** seems misleading, as it actually checks for non-ASCII characters in the output, and it does not confirm that UTF-16 was used.

If the goal is to double-check the default file encoding in the subprocess, instead I would simply add a verification in PemEncodingTest:


if (!"UTF-16".equals(Charset.defaultCharset().displayName())) {
    throw new RuntimeException("File encoding is not UTF-16");
}

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/23492#discussion_r1944946184
PR Review Comment: https://git.openjdk.org/jdk/pull/23492#discussion_r1944964050
PR Review Comment: https://git.openjdk.org/jdk/pull/23492#discussion_r1945089066