New SAML 2024.05.30 - Not to merge but just for SAML feature branch testing #2908

Co-authored-by: Peter Chen <peter-h.chen@broadcom.com> Co-authored-by: Bruce Ricard <bruce.ricard@broadcom.com> Co-authored-by: Danny Faught <danny.faught@broadcom.com>

@ignore

* Instead of calling fail(). We have a suspicion that there is a bug in the way the tests are running (most of them are somehow not running with "./gradlew test" and we have a theory that a combination of mixing junit4 imports and the junit5 fail() might be contributing. * I was careful to use @ignore for tests importing the junit4 @test, and @disabled for tests using the junit5 @test. * These annotations were added, with the idea that you can search for '@ignore("SAML' and '@disabled("SAML' to find the tests that need attention before we finish the SAML library conversion. @ignore("SAML test fails") @ignore("SAML test doesn't compile") @ignore("SAML test setup doesn't compile") @disabled("SAML test fails") @disabled("SAML test doesn't compile") * A few tests are set to ignore because they're failing for the right reasons, but more work is needed to finish that and get back to green. The goal is to start tracking these annotations instead of failing tests, so we can stay green. * Tests now running: server module: 3,435 (in IntelliJ) (98 total ignored) uaa module: 67 (command line run of "./gradlew test" for all tests - still needs troubleshooting) Co-authored-by: Danny Faught <danny.faught@broadcom.com>

Co-authored-by: Hongchol Sinn <hongchol.sinn@broadcom.com>

* Removed commented-out references to the outdated SAML extension library Co-authored-by: Duane May <duane.may@broadcom.com>

- Adds back endpoint and incorporates forwarding for new pattern saml2 endpoints, Still has some wip elements WithHttpsNotRequired > samlMetadataReturnsOk still red RelyingPartyRegistration is hardcoded in xml, /saml/metadata/ with trailing slash not working missing parity with develop [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

…NotRequired -> samlMetadataReturnsOk is green - fixed one test but still WithHttpsRequired > samlMetadataReturnsOk is red after fixing this test - HealthzShouldNotBeProtectedMockMvcTests > WithHttpsRequired > samlMetadataRedirects() FAILED java.lang.AssertionError: Range for response status value 200 expected:<REDIRECTION> but was:<SUCCESSFUL> [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

Co-authored-by: Peter Chen <peter-h.chen@broadcom.com> Co-authored-by: Bruce Ricard <bruce.ricard@broadcom.com> Co-authored-by: Danny Faught <danny.faught@broadcom.com>

@ignore

* Instead of calling fail(). We have a suspicion that there is a bug in the way the tests are running (most of them are somehow not running with "./gradlew test" and we have a theory that a combination of mixing junit4 imports and the junit5 fail() might be contributing. * I was careful to use @ignore for tests importing the junit4 @test, and @disabled for tests using the junit5 @test. * These annotations were added, with the idea that you can search for '@ignore("SAML' and '@disabled("SAML' to find the tests that need attention before we finish the SAML library conversion. @ignore("SAML test fails") @ignore("SAML test doesn't compile") @ignore("SAML test setup doesn't compile") @disabled("SAML test fails") @disabled("SAML test doesn't compile") * A few tests are set to ignore because they're failing for the right reasons, but more work is needed to finish that and get back to green. The goal is to start tracking these annotations instead of failing tests, so we can stay green. * Tests now running: server module: 3,435 (in IntelliJ) (98 total ignored) uaa module: 67 (command line run of "./gradlew test" for all tests - still needs troubleshooting) Co-authored-by: Danny Faught <danny.faught@broadcom.com>

- Adds back endpoint and incorporates forwarding for new pattern saml2 endpoints, Still has some wip elements WithHttpsNotRequired > samlMetadataReturnsOk still red RelyingPartyRegistration is hardcoded in xml, /saml/metadata/ with trailing slash not working missing parity with develop [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

…NotRequired -> samlMetadataReturnsOk is green - fixed one test but still WithHttpsRequired > samlMetadataReturnsOk is red after fixing this test - HealthzShouldNotBeProtectedMockMvcTests > WithHttpsRequired > samlMetadataRedirects() FAILED java.lang.AssertionError: Range for response status value 200 expected:<REDIRECTION> but was:<SUCCESSFUL> [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- With the new SAML lib, SAML SP metadata generation relies on a relyingPartyRegistration, which requires a valid SAML IDP metadata. In the context of UAA external SAML IDP login, UAA does not know what the SAML IDP metadata is, until the operator adds it via the /identity-providers endpoint. Also, some SAML IDPs might require you to supply the SAML SP metadata first before you can obtain the SAML IDP metadata. See relevant issue: spring-projects/spring-security#11369 - Previously, to solve this problem, the SAML SP metadata generation relies on relyingPartyRegistration values in saml-providers.xml, which hardcodes a SAML IDP metadata URL (point to some example Okta SAML instance); this means that UAA's SP metadata generation relies on the example Okta SAML instance to be running. - This commit, instead, supplies a hardcoded dummy SAML IDP metadata here to unblock the SAML SP metadata generation, at the advice of Spring Security team, so that UAA's functioning does not rely on some external running Okta instance. - code reference: https://github.com/spring-projects/spring-security-samples/blob/1b28351693d60f01a511cbcc18b64590452a3851/servlet/java-configuration/saml2/login/src/main/java/example/SecurityConfiguration.java#L62 [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- A continuation of 65d1f0f - This test is failing as early as e7beec7 due to the removal of SAML code, as this test is related the SAML feature [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

@disabled

* Has to be commented out of the erb file even when the test method used @disabled. Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- A continuation of 65d1f0f - This is a test recently added to develop branch, so ignoring this here because the SAML feature is still being built. [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- to reflect the fact that this IDP metadata just needs to exist in its bare minimal form, where the specific fields in it do not affect the SP metadata generation [#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- previously some tests error with: ``` net.shibboleth.utilities.java.support.xml.XMLParserException: Unable to parse inputstream, it contained invalid XML ``` - this issue is fixed once we switch to loading the idp saml metadata via a file (instead of an InputStream) [186822654] Co-authored-by: Danny Faught <danny.faught@broadcom.com>

Co-authored-by: Danny Faught <danny.faught@broadcom.com>

* We're reprioritizing the test to get this test to pass. Co-authored-by: Bruce Ricard <bruce.ricard@broadcom.com>

Co-authored-by: Duane May <duane.may@broadcom.com>

…ect request - Tests are failing but they are behaving as expected with curl and browser for /saml/metadata /saml/metadata/example and /saml/metadata/example/ - /saml/metadata/ is not returning xml - The dispatcher ordering along with position in the filter-mapping must be set properly. [#186986697] Co-authored-by: Bruce Ricard <bruce.ricard@broadcom.com>

Co-authored-by: Duane May <duane.may@broadcom.com>

…VC Tests - /saml/metadata/ is not returning xml [#186986697] Co-authored-by: Filip Hanik <fhanik@vmware.com>

Co-authored-by: Alicia Yingling <alicia.yingling@broadcom.com> Co-authored-by: Duane May <duane.may@broadcom.com>

* Must be changed back to /saml/metadata later, removing "example". Co-authored-by: Alicia Yingling <alicia.yingling@broadcom.com> Co-authored-by: Duane May <duane.may@broadcom.com>

Co-authored-by: Alicia Yingling <alicia.yingling@broadcom.com> Co-authored-by: Duane May <duane.may@broadcom.com>

…ault - Updated to use direct GetMapping [#186986697] Co-authored-by: Filip Hanik <fhanik@vmware.com>

…pulated in SPSSODescriptor - Building out EntityDescriptor in the RelyingPartyRegistration which contains the SPSSODescriptor picked up by the resolve method [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

- Need to fix credential type being empty Caused by: java.lang.IllegalArgumentException: credentials types cannot be empty ....(SamlRelyingPartyRegistrationRepository.java:84) [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

- Signature is not positioned correctly. It should be a child of EntityDescriptor, but the singingX509Credential.signing call positions it in SPSODescriptor [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

…RequestsSigned - correctly reads off UAA configs to populate these fields, instead of using hardcoded values - refactor to directly reading `login.saml.NameID` config (a more modern approach) instead of constructing a bean in xml (a more legacy approach) - side note: update the UAA config used in mock mvc tests (/uaa/src/test/resources/integration_test_properties.yml) to use a non-default option of `login.saml.nameID` so that we can test that the correct value is being piped through Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- there are many commented out codes from prior wip commits (which at this point, I decided, are too hard to fix or tidy up). Hence, in this commit, clean them up [186822654] Co-authored-by: Duane May <duane.may@broadcom.com>

- the SAML SP metadata is still WIP, so this IT will fail. Ignoring it for now so that "CI" is green along with all other SAML tests currently failing / non-functional due to the WIP state of the SAML feature. - see defails of this approach in 73520d9 [186822654] Co-authored-by: Duane May <duane.may@broadcom.com>

https: //docs.spring.io/spring-security/reference/5.8/migration/servlet/saml2.html Co-authored-by: Duane May <duane.may@broadcom.com>

Use RestController, Slf4j, Getter Use textblocks Co-authored-by: Duane May <duane.may@broadcom.com>

Use assertThat Use textblocks Co-authored-by: Duane May <duane.may@broadcom.com>

Co-authored-by: Duane May <duane.may@broadcom.com>

[#186986697] Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- also: refactor the UAA config used in mock mvc tests (/uaa/src/test/resources/integration_test_properties.yml) from the deprecated saml key fields (eg: login.serviceProviderKey) to the new ones (eg: login.saml.keys), so that we test for the new fields. - also fix the api docs test so that it now correctly marks the retrieve id zones response's `config.samlConfig.certificate` as optional (this field is only returned if you use the deprecated saml key config fields) [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

- populate saml sp metadata field for use='encryption' cert - might be counter-intuitive that the setting on rp registration that controls this is "decryptionX509Credentials", but the resulting sp metadata indeed includes use='encryption' which matches develop branch [186822654] Co-authored-by: Duane May <duane.may@broadcom.com>

@ConfigurationProperties

- to be processed by a single class "SamlConfiguration" where the @ConfigurationProperties(prefix="login.saml") annotation has the ability to process all fields under the login.saml section of UAA.yml - this is helpful because we can now centrally read, process, even validate all saml config fields under "login.saml" - pay attention to @ConfigurationProperties annotation's various requirements though: such as the private field names need to match the actually UAA.yml field name (e.g.: login.saml.fooBar -> private String fooBar); and that there need to be public setters and getters for each field - see: https://docs.spring.io/spring-boot/docs/current/reference/html/features.html#features.external-config.typesafe-configuration-properties.using-annotated-types - the exception of the saml entity id, which in UAA.yml is somehow outside of the login.saml context (set by login.entityID) so that field stays under class SamlEntityIdConfiguration Co-authored-by: Duane May <duane.may@broadcom.com>

@ConfigurationProperties

- these getters and setters are required for @ConfigurationProperties annotation to work; use lombok so that we don't need to explicitly define them [186822654] Co-authored-by: Duane May <duane.may@broadcom.com>

@DaTa

- as @DaTa covers the getters and setters Co-authored-by: Duane May <duane.may@broadcom.com>

- configure the file name of the saml sp metadata (the downloaded xml file name when accessing the metadata endpoint: http://localhost:8080/uaa/saml/metadata) to match the status quo on develop branch: "saml-sp.xml" - This file name likely do not matter, but out of caution, we should maintain the same file name as before [186822654] Co-authored-by: Duane May <duane.may@broadcom.com>

- now that the metadata is being provided at the correct location: /saml/metadata, we can correct the test expectation to reflect that (hence matching the develop branch) [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

- Removed forwarding of `/saml/metadata` endpoint to `/saml/metadata/example`. It is not necessary because `/saml/metadata` endpoint method already calls `/saml/metadata/{registrationId}` with `example` as the default registrationId. (See class `SamlMetadataEndpoint`.) - Made `HttpsEnforcementFilter` to be added to the top of the `SecurityFilterChainPostProcessor`'s `SecurityFilterChain`. - Added `secFilterOpen06SAMLMetadata` to `SecurityFilterChainPostProcessor`'s `redirectToHttps` list. [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com> Co-authored-by: Peter Chen <peter-h.chen@broadcom.com>

- Removed SamlExtensionUrlForwardingFilter. Just commented out for now in case we need it later. - Removed unneeded comments in test code. [#186986697] Co-authored-by: Duane May <duane.may@broadcom.com>

[#187084275] Co-authored-by: Duane May <duane.may@broadcom.com>

- Change SamlRelyingPartyRegistrationRepository to Configuration - Use constructor args instead of Autowired Co-authored-by: Duane May <duane.may@broadcom.com>

still had opensaml 3.4.6 Co-authored-by: Duane May <duane.may@broadcom.com>

- when SAML IDP is configured via uaa.yml, when the user goes to "/uaa/saml2/authenticate/{saml-idp-alias}", they will get sent to the configured SAML IDP with a SAML authn request. Specifically, spring-security will do the following: - when the IDP's Binding mode is "HTTP-Redirect", the user is redirected to the IDP - when the IDP's Binding mode is "HTTP-POST", the user's browser is triggered to POST to the IDP. For this to work, the ContentSecurityPolicyFilter needs to updated to exempt "/saml2" from policy enforcement, such that the script that initiates the POST can be executed in the browser. Similar to how this filter exempts /saml (the existing saml-related path on develop branch). - refactor: update the dummy IDP metadata file dummy-saml-idp-metadata.xml to not point to example.com, but to https://www.cloudfoundry.org (which is more of a known destination) - refactor: use constant DEFAULT_REGISTRATION_ID [#187084275] Co-authored-by: Duane May <duane.may@broadcom.com>

prefix="login.saml" was in 2 ConfigProps classes before merged into 1

Reads provider info from database Passes the registrationId as relayState Signed-off-by: Prateek Gangwal <prateek.gangwal@broadcom.com>

when running multiple IT tests, the simplesamlphp2 link was also listed, and causing a conflict with url matcher Signed-off-by: Duane May <duane.may@broadcom.com>

Signed-off-by: Duane May <duane.may@broadcom.com> Signed-off-by: Ivan Protsiuk <ivan.protsiuk@broadcom.com> #187106956

Signed-off-by: Ivan Protsiuk <ivan.protsiuk@broadcom.com>

- Improve Testing of SAML Request/Response with Saml2TestUtils - Configure assertionConsumerServiceLocation in one location. - Attempted move to OpenSaml4AuthenticationProvider requires a shadow dependency on opensaml to remove the need for non-FIPS compliant security provider. Not yet in place Signed-off-by: Duane May <duane.may@broadcom.com> Signed-off-by: Alicia Yingling <alicia.yingling@broadcom.com>

Move user shadowing, attribute processing, and authorities processing to their own classes. Enable Authorities Signed-off-by: Ivan Protsiuk <ivan.protsiuk@broadcom.com>

This provides general response validation. Signed-off-by: Prateek Gangwal <prateek.gangwal@broadcom.com>

Signed-off-by: Prateek Gangwal <prateek.gangwal@broadcom.com> #187809240

Signed-off-by: Duane May <duane.may@broadcom.com>

Signed-off-by: Duane May <duane.may@broadcom.com> Signed-off-by: Hongchol Sinn <hongchol.sinn@broadcom.com>

Signed-off-by: Hongchol Sinn <hongchol.sinn@broadcom.com>

- Main logout flows are working - IDP Initiated logout is working - Handle metadata XML passed in instead of metadata location for both bootstrap and SamlIdentityProviderConfigurator Signed-off-by: Duane May <duane.may@broadcom.com>

- clean up the rest of the pageObjects package Signed-off-by: Duane May <duane.may@broadcom.com>

- now attempts to retrieve the non-existent url https://simplesamlphp.uaa.com/saml2/idp/metadata.php Signed-off-by: Duane May <duane.may@broadcom.com>

- Implemented to the same level as the default IdenityZone's SP metadata generation. - Minus `NameIDFormat` value populaition and registration-ID specific implementation. [#187846376]

…not valid anymore.

- No longer have Ignored tests only Disabled Signed-off-by: Duane May <duane.may@broadcom.com>

Signed-off-by: Duane May <duane.may@broadcom.com>

- correctly populates the basic fields of non-default zone SAML SP metadata (such as WantAssertionsSigned and AuthnRequestsSigned), so that for default vs. non-default zones, the SP metadatas have feature parity. [#187846376] Signed-off-by: Duane May <duane.may@broadcom.com> Signed-off-by: Peter Chen <peter-h.chen@broadcom.com>

Signed-off-by: Peter Chen <peter-h.chen@broadcom.com>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

New SAML 2024.05.30 - Not to merge but just for SAML feature branch testing #2908

New SAML 2024.05.30 - Not to merge but just for SAML feature branch testing #2908

Commits on Jul 12, 2024

New SAML 2024.05.30 - Not to merge but just for SAML feature branch testing #2908

Are you sure you want to change the base?

New SAML 2024.05.30 - Not to merge but just for SAML feature branch testing #2908

Commits on Jul 12, 2024