Fix Http2ClientReconnectTest timeout issue in vertx4 module

[velo]

Summary

Fixes the unstable timeout issue in Http2ClientReconnectTest.allRequestsShouldBeAnswered() that was causing test failures after 30 seconds.

Root Cause

The test was timing out because the asynchronous Future chain was not properly handled by the VertxTestContext. The specific issues were:

1. Incorrect Future composition: Used CompositeFuture.all(List<Future>) instead of Future.all(List<Future<?>>)
2. Missing test context completion: The Future chain was created but not properly attached to the test context
3. No error handling: Failed async operations weren't being reported to the test context

Changes Made

• Changed CompositeFuture.all(List<Future>) to Future.all(List<Future<?>>) in the sendRequests method
• Added proper .onComplete() handlers to ensure test context receives completion signals
• Added error handling to call testContext.failNow() when async operations fail
• Applied the same fix to both test methods: the main test and the nested "After server is available again" test

Test Results

✅ All vertx4 tests now pass consistently (26 tests, 0 failures, 0 errors)
✅ The specific failing test Http2ClientReconnectTest.allRequestsShouldBeAnswered() now completes in ~0.2s instead of timing out after 30s

Test Plan

• Run the specific failing test: mvn test -pl vertx/feign-vertx4-test -Dtest=Http2ClientReconnectTest
• Run all vertx4 tests to ensure no regressions: mvn test -pl vertx/feign-vertx4-test
• Verify test execution time is reasonable (no 30s timeouts)

🤖 Generated with Claude Code