Backport SessionScheduleTest coverage for previous-day start-of-week handling

[Copilot]

This PR copies over the missing SessionScheduleTest case covering a schedule that starts in one timezone/day context and ends in another. It preserves regression coverage for StartDay resolution when locale week boundaries can shift how the session window is interpreted.

• Added regression coverage

  • Backports testSessionStartInPreviousDayOfWeek() into quickfixj-core/src/test/java/quickfix/SessionScheduleTest.java.
  • Covers a session configured with:
    • StartTime=13:00:00 Australia/Victoria
    • StartDay=Sunday
    • EndTime=17:00:00 America/New_York
    • EndDay=Saturday

• Edge case under test

  • Exercises locale-sensitive week-start behavior using Locale("en", "AU").
  • Verifies DefaultSessionSchedule.isSessionTime() remains correct when the effective session start can resolve into the previous day-of-week.

• Why this matters

  • Keeps this PR aligned with the upstream regression coverage for cross-timezone session scheduling.
  • Protects against behavior differences across JDK/locale combinations that affect weekday interpretation.

Locale.setDefault(new Locale("en", "AU"));

settings.setString(Session.SETTING_START_TIME, "13:00:00 Australia/Victoria");
settings.setString(Session.SETTING_START_DAY, "Sunday");
settings.setString(Session.SETTING_END_TIME, "17:00:00 America/New_York");
settings.setString(Session.SETTING_END_DAY, "Saturday");

assertEquals(true, schedule.isSessionTime());

[chrjohn]

@copilot LogUtilTest.testLogThrowable:51 fails

[Copilot]

@copilot LogUtilTest.testLogThrowable:51 fails

Fixed in 3ea3d92 by updating LogUtilTest to throw from getCreationTime(), which is the MessageStore method the current session path uses. I also re-ran LogUtilTest and SessionScheduleTest locally and both pass.