refactor(startup): rotate ascii banners, drop the 1s startup sleep (#36887)

[wezell]

Fixes #36887

Summary

Cosmetic startup change with one real win: doArt() no longer sleeps for a second on every boot.

Before	After
private static boolean artDone + check-then-set	AtomicBoolean.compareAndSet — two racing threads can't both print
One hardcoded banner	7 banners, one picked per boot
Thread.sleep(1000)	gone — a second back on every startup

Note for reviewers: text blocks and backslashes

Java text blocks still process escape sequences, so ASCII art containing \ does not compile as-is:

error: illegal escape character
 \     .'_  .-'),-----. /     '._  .-----.
  ^

Every backslash in the art is escaped as \\, and \s is used where a trailing space must survive (text blocks strip incidental trailing whitespace). If you add a banner, run it through javac before pushing.

Verification

Compiled the class standalone against a stub Logger and rendered all 7 banners — output is byte-identical to the source art, including the trailing spaces preserved by \s.

Checklist

• Docs — n/a (no config or API surface)
• Translations — n/a
• Security considered — log output only, no input
• Tests — none added; it's a startup banner with no logic worth pinning

🤖 Generated with Claude Code

[github-actions]

❌ Linked Issue Needs Team Label

This PR is linked to issue #36887, but that issue has no Team : * label. Every linked issue must be owned by a team for tracking and triage.

How to fix this:

Apply a Team : * label to the linked issue (e.g., Team : Scout, Team : Platform, Team : Falcon, Team : Maintenance). Then push a new commit or edit the PR description to re-run this check.

---

This comment was automatically generated by the issue linking workflow

[claude]

Claude finished @wezell's task in 2m 29s —— View job

---

Code Review

Reviewed the diff against origin/main. Verified: ascii-art.txt has 6 %%% separators → 7 banners (matches the PR description), the only caller is ContextLifecycleListener.java:23, and the resource lives under src/main/resources so it ships on the classpath.

New Issues

No blocking issues found. Two minor notes below, neither blocks merge.

• 🟡 Medium: AsciiArt.java:23-30 — compareAndSet(false, true) runs before load(). If the resource is ever missing or unreadable on the first call, artDone is already latched to true, so the banner is silently skipped forever (no retry on later calls). This is harmless for a cosmetic banner and arguably intentional, but worth a one-line comment noting the ordering is deliberate so a future reader doesn't "fix" it.

• 🟡 Medium: AsciiArt.java:18,46 — SEPARATOR = "\n%%%\n" is passed to String.split(...), which interprets it as a regex. It works today only because % and \n have no special regex meaning. If the separator token ever changes to something with regex metacharacters (., |, *, etc.) this breaks silently. Consider Pattern.quote(SEPARATOR) or a literal-split helper to make the intent explicit. Fix this →

Notes (non-issues, verified OK)

• art[(int)(System.currentTimeMillis() % art.length)] — long % int promotes to long, result is always in [0, art.length-1], and currentTimeMillis() is non-negative, so the cast to int and the index are safe.
• AtomicBoolean.compareAndSet correctly closes the check-then-set race the old boolean artDone had — two racing threads can no longer both print.
• load() handles both the null-stream and IOException cases by returning an empty array, and doArt() early-returns on empty. Failure path is contained.
• Dropping Thread.sleep(1000) is a straight win with no dependents (only caller is ContextLifecycleListener).
• Config/Logger used correctly; no System.out/getProperty/getenv. Log-output-only, no input — no security surface.

Unrelated to the code: the CI comment flags that linked issue #36887 needs a Team : * label — that's a triage/label action, not a code change.

Overall this does what the description says and is safe to merge. The two medium notes are optional polish.
· issue-asciiart-startup-banners