docs: document that navigating to the shown view does nothing

[mcollovati]

Part of #25580

DX/docs · flow-server · everyone who navigates to the view that is
already shown

Navigating to the view that is already shown does nothing, and the
javadoc promised the opposite: it said the return value tells you
whether the navigation happened, when it never does. Anyone who wanted
the view rebuilt had no way to find the method that rebuilds it. Both
are documented now; behaviour is unchanged.

Risks:

• ✅ Javadoc only, plus one test that pins the documented behaviour: no
  behaviour change, no API change, nothing deprecated.

Context. The skip prevents navigation loops, because a BeforeEnter
reroute and the client history sync go through the same call, so the
behaviour stays as it is and only the javadoc was wrong.
refreshCurrentRoute(boolean) already does the rebuild that was asked
for.

• Documented on all nine UI.navigate overloads that navigating to the
  location that is already shown does nothing: the view is not
  re-instantiated, no navigation lifecycle event fires, and
  refreshCurrentRoute(boolean) is the way to rebuild it.
• Corrected @return on the seven overloads that return an Optional.
  It said "the view instance, if navigation actually happened"; the
  Optional holds the view that was already shown when the navigation
  was skipped, and is empty when a reroute or a forward went elsewhere.
• Added the reciprocal cross-reference on refreshCurrentRoute(boolean),
  which now says it is the way to rebuild the view that is currently
  shown and links back to navigate.
• Added UITest coverage: navigating twice to the same target returns
  the same instance, and refreshCurrentRoute(false) replaces it.

[mcollovati]

Type of change

• Documentation

Javadoc only, so there is nothing to exercise by hand: the review is
reading the wording. mvn javadoc:javadoc -pl flow-server builds with
no reference errors, and UITest passes (77 tests, 1 pre-existing
@Disabled).

Note

This covers option 1 of item 1 in #25580 only. Whether navigate
should return Optional.empty() when it skipped the navigation, and
whether it should gain a force-rebuild overload, were left out as a
separate API ticket, as agreed in the issue. Item 4 of the same issue,
the BeanValidationBinder note about where a constraint message comes
from, is still open and not part of this PR.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

Test Results

 1 440 files  ±0   1 524 suites  ±0   1h 41m 25s ⏱️ -54s
12 046 tests +1  11 978 ✅ +1  68 💤 ±0  0 ❌ ±0 
12 364 runs  +1  12 296 ✅ +1  68 💤 ±0  0 ❌ ±0 

Results for commit 63af331. ± Comparison against base commit 3080cc0.