glusk / caesar Goto Github PK

After closing #25, this dependency is no longer needed!

Lets add a simple xor function that operates on Bytes.

Lets add this feature to the library!

1. We want to have a thread-safe MessageDigest object
2. We want to have an simple api for computing embedded hashes. For example: H(H(b1), b2, b3)

The Java 9 modules source path is not properly detected by the coveralls maven plugin.

Coveralls plugin further complains:

Execution default-cli of goal org.eluder.coveralls:coveralls-maven-plugin:4.3.0:report failed: A required class was missing while executing org.eluder.coveralls:coveralls-maven-plugin:4.3.0:report: javax/xml/bind/DatatypeConverter

A workaround was proposed here.

Refer to this SO Question!

Added the repo to codacy:
https://www.codacy.com/manual/Glusk/caesar

Setup Travis to deploy the project to Maven Central every time a new git tag is pushed.

Resources:

• https://bpodgursky.com/2019/07/31/using-travisci-to-deploy-to-maven-central-via-git-tagging-aka-death-to-commit-clutter/
• https://www.phillip-kruger.com/post/continuous_integration_to_maven_central/

Enable Checkstyle checks for tests but skip Javadoc checks.

Reasons against enforcing Javadoc comments in tests:

• documentation is hard to do
• test Javadocs don't get published to Maven Central
• in my experience, most of the test Javadocs are useless anyway (for example: /** The test for class Foo. */)
• if you change the test you have to change the doc
• this SO answer also suggests not to enforce test documentation

Some code snippets seem to lack focus. The accompanying text could be made shorter.

Javadocs feel vague and imprecise.

Add an introduction paragraph with maven dependency library usage example.

Document the release process - how to bump the version and triger a release/deployment.

Implement HMAC as specified in RFC 2104.

After implementing #25, the following API changes should be made:

• remove constructor DigestArgument(String, int)
• remove constructor DigestArgument(byte[])
• DigestArgument should be renamed to PlainText and moved to package com.github.glusk.caesar
• remove class BinaryString
• update examples in README.md

Update push condition to:

if [ "$success"] && [ "$dryRun" = false ]

It contains only one abstract method.

And WrappedBytes class is not really needed. Wherever there is:

new WrappedBytes(...)

in the code base, one can replace it with:

() -> ...

We want to specify the contract of some of the Object's methods at the Bytes interface level. As these can't be implemented as default methods, we can use a pattern similar to the one used in java.util (List and AbstractList).

Add an abstract class AbstractBytes that extends implements the Bytes interface.

Remove the @FunctionalInterface annotation in Bytes interface definition since the use of Bytes in a lambda expression is discouraged. Doing so breaks its equals() contract, for example.

AbstractBytes should implement equals(), hashCode() and toString() (YAGNI, at the moment).

Classes that used to implement Bytes should now extend AbstractBytes.

A release version should not be committed. The flow should be as follows:

1. checkout to a new release branch; bump version to a new release tag and push it
2. checkout to a master branch; bump version to a new snapshot version and push the change

There should be a "dry run" option that doesn't push anything to the remote repository.

Test coverage should be 100% given the currently (version 0.1.0) still relatively small codebase size.

Compare Bytes instead of arrays.

1. add -B (batch mode) flag to all Maven commands
2. make sure you're not running any redundant maven phases and goals

Show the core concept of wrapping a MessageDigest instance into an ImmutableMessageDigestObject.

At the moment, we're wired to an external dependency jakarta.xml.bind:jakarta.xml.bind-api just so that we can use DatatypeConverter. This dependency could be removed.

Java 8 introduced Base64 class and the hex encoding/decoding looks easy enough to implement.

The library currently only parses Base64 and Base16 strings. We should add default methods to Bytes that would encode byte sequences to Base64 and Base16 strings.

This fails:

@Test
public void parsesANegativeByteValue()  {
    assertArrayEquals(
        new byte[] { (byte) 0b10000000 },
        new Hex("80").asArray()
    );
}