}) {
    init {
        smokeBuildTypes.forEach(this::buildType)
    }

    companion object {
        /**
         * Specific builds shown under the "Smoke Test" subproject in Pull Request Feedback.
         */
        val pullRequestFeedbackSmokeBuilds: Set<SpecificBuild> =
            setOf(
                SpecificBuild.SmokeTestsMaxJavaVersion,

    /**
     * Notification message displayed on the login page.
     * This message is shown to users on the authentication page.
     */
    @Size(max = 3000)
    public String notificationLogin;

    /**
     * Notification message displayed on the search top page.
     * This message is shown to users on the main search page.
     */
    @Size(max = 3000)
    public String notificationSearchTop;

* If the source example(s) file can have different syntax for different Python versions, there should be different versions of the file, and they should be shown in tabs in the docs.
* There should be tests testing the source example.
* Before the PR is applied, the new tests should fail.
* After applying the PR, the new tests should pass.
* Coverage should stay at 100%.

        assertTrue(details.contains("Total: 1 event(s)"));
        assertTrue(details.contains("Something went wrong"));
        // Throwable is no longer included in summary-oriented format
        // Only log message is shown per event line
    }

    @Test
    public void test_formatDetails_emptyList() {
        TestableLogNotificationJob testableJob = new TestableLogNotificationJob();

                * `@app.get("/users/", tags=["users"])`
                * `@app.get("/items/", tags=["items"])`

                The order of the tags can be used to specify the order shown in
                tools like Swagger UI, used in the automatic path `/docs`.

                It's not required to specify all the tags used.

    *   Add `tf.keras.layers.AbstractRNNCell` as the preferred implementation
        for RNN cells in TF v2. User can use it to implement RNN cells with
        custom behavior.
    *   Keras training and validation curves are shown on the same plot when
        using the TensorBoard callback.
    *   Switched Keras `fit/evaluate/predict` execution to use only a single
        unified path by default unless eager execution has been explicitly

* 🐛 Fix edge case with repeated aliases names not shown in OpenAPI. PR [#2351](https://github.com/tiangolo/fastapi/pull/2351) by [@klaa97](https://github.com/klaa97).