Fix debug message that does not properly handle single quotes
according to java.text.MessageFormat, which in turn causes the
replacement to not be handled.
Signed-off-by: Phil Brown <brownp@stellarscience.com>
I use Keycloak Spring Adapter (KSA) to secure existing application. Today I realized that some functions didn't work anymore because of security checking like this:
```
GrantedAuthority adminRole = new MySpecialGrantedAuthority( "superadmin" );
for ( GrantedAuthority role : userRoles ) {
if ( role.equals( adminRole ) ) {
return true;
}
}
```
In this example, when I use KSA authorization fails.
I believe, that more preferable in `KeycloakRole` use this implementation of `equals` method.
I use Keycloak Spring Adapter (KSA) to secure existing application. Today I realized that some functions didn't work anymore because of security checking like this:
```
GrantedAuthority adminRole = new MySpecialGrantedAuthority( "superadmin" );
for ( GrantedAuthority role : userRoles ) {
if ( role.equals( adminRole ) ) {
return true;
}
}
```
In this example, when I use KSA authorization fails.
I believe, that more preferable in `KeycloakRole` use this implementation of `equals` method.
Users who want to use PKCE support with the KeycloakInstalled adapter need to set the property
``"enable-pkce": true` in the adapter configuration / `keycloak.json`.
This adds support for the "S256" code_challenge_method to the JS Adapter.
Note that the method "plain" was deliberately left out as is not recommended
to be used in new applications.
Note that this PR includes two libraries:
- [base64-js]{@link https://github.com/beatgammit/base64-js}
- [js-sha256]{@link https://github.com/emn178/js-sha256}
`base64-js` is needed for cross-browser support for decoding the
Uint8ArrayBuffer returned by `crypto.getRandomValues` to a PKCE
compatible base64 string.
`js-sha256` library is required because the `crypto.subtle.digest`
support is not available for all browsers.
The PKCE codeVerifier is stored in the callbackStore of the JS Adapter.
Note: This PR is based on #5255 which got messed up during a rebase.
