Fix Paginator dropping results on empty token pages

The Paginator stopped iterating at the first empty page even when the
response carried a non-empty next_page_token, silently dropping results
on later pages. Token-paginated list methods (e.g. tables().list()) can
return empty intermediate pages with a valid token, so results could be
under-returned.

Add two explicit pagination factories:
  - Paginator.newTokenPagination: iterate until nextPageFn returns null;
    empty pages are skipped, not treated as the end of results.
  - Paginator.newOffsetPagination: stop on the first empty page (offset,
    SCIM and legacy SQL APIs have no token and rely on this).
The existing Paginator constructor is kept and deprecated (it retains its
previous offset/limit behavior), so this is not a breaking change.

Regenerate all *API.java list methods to the appropriate factory. Update
hand-written APIs: SharesExtAPI -> newTokenPagination; SCIM
Users/Groups/ServicePrincipals (and Account* variants) -> newOffsetPagination.

Signed-off-by: Hector Castejon Diaz <hector.castejon@databricks.com>

Author: hectorcast-db

NEXT_CHANGELOG.md

@@ -4,10 +4,14 @@
 
 ### New Features and Improvements
 
+* Added `Paginator.newTokenPagination(...)` and `Paginator.newOffsetPagination(...)` factory methods in `com.databricks.sdk.support`, which make the pagination strategy explicit. The `Paginator` constructor is now deprecated in favor of these; it keeps its previous (offset/limit) behavior.
+
 ### Breaking Changes
 
 ### Bug Fixes
 
+* Fixed `Paginator` silently dropping results when a token-paginated response returned an empty page with a non-empty `next_page_token`. List methods (e.g. `tables().list()`) now keep paging until the page token is absent instead of stopping at the first empty page.
+
 ### Security Vulnerabilities
 
 ### Documentation