test: Add Cucumber BDD coverage for the Articles API

[devin-ai-integration]

Summary

Adds a Cucumber layer on top of the existing suite: 51 scenarios across create-article, article-feed and favorite-article, driving the real HTTP API (security filters, MyBatis, SQLite) rather than mocks. ./gradlew test now runs 119 tests (68 before); ./gradlew cucumberTest runs the scenarios alone.

Infrastructure, all under src/test/java/io/spring/cucumber:

• RunCucumberTest — @Suite + @IncludeEngines("cucumber"), glue io.spring.cucumber, features at src/test/resources/features.
• CucumberSpringConfiguration — @CucumberContextConfiguration + @SpringBootTest(RANDOM_PORT) on a new cucumber profile (build/cucumber.db, Flyway stops at V1 so scenarios own their data). Its beans live in a nested @TestConfiguration on purpose: the app component-scans io.spring, and a @Component here would leak into every other Spring test context.
• ApiClient resolves the port lazily (environment.getRequiredProperty("local.server.port")) — injecting it as a constructor @Value fails, the property does not exist yet when the bean is created.
• Hooks wipes every table before each scenario, so scenarios are order-independent and reruns are clean.

Steps build the request from the columns actually present in the table, which is what makes the missing-field outline data-driven:

When I create an article with:
  | title       | [empty]     |
  | description | Ever wonder |
Then the response status should be 422
And the response should report "can't be empty" for the field "title"

if (article.containsKey(field)) payload.put(field, EMPTY.equals(v) ? "" : v);
payload.put("tagList", tagList(article.get("tags")));  // always sent: a null tagList NPEs in Article

Two API behaviors constrained the scenarios and are worth knowing:

• Sending "title": null (or omitting it) returns 500, not 422: DuplicatedArticleValidator calls Article.toSlug(value) on a null value before @NotBlank can reject it. The feature therefore sends [empty] for the fields under test; a blank Gherkin cell is null, which is why only description/body appear in the second Examples table. Not fixed here — this PR only adds tests.
• /articles/feed applies LIMIT to the tag-joined rows (findArticlesOfAuthors joins article_tags without a distinct id subquery), so a tagged article consumes several rows of the page. The global feed does not have this problem (selectArticleIds). The user-feed pagination outline uses untagged articles to stay meaningful.

Not addressed: the pre-existing JaCoCo gate still fails (0.34 vs 0.80) as AGENTS.md instructs — run with -x jacocoTestCoverageVerification.

Link to Devin session: https://partner-workshops.devinenterprise.com/sessions/ced6e245b02a49059d54e07f1b07ca25

---

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment, CI, and merge conflict monitoring

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no bugs or issues to report.