fix(build): allow kalimba board to configure, drop dead duplicate dispatch

[HoneyTyagii]

Summary

Selecting -DEBLDR_BOARD=kalimba always failed at configure time with Unknown board: kalimba, even though boards/kalimba/ exists and the branch had already registered the port. The cause was a stray message(FATAL_ERROR ...) sitting inside the kalimba branch of the board dispatch in CMakeLists.txt. The same region also contained a full second copy of the board dispatch (cortex_m3 through kalimba) that was unreachable dead code.

This PR fixes the dispatch and adds a configure-only regression check.

Type of Change

• feat: New feature
• fix: Bug fix
• docs: Documentation only
• style: Formatting, no code change
• refactor: Code restructuring without behavior change
• test: Add or fix tests
• build: Build system or dependency changes
• ci: CI/CD pipeline changes
• perf: Performance improvement

Changes

• CMakeLists.txt: remove the misplaced message(FATAL_ERROR "Unknown board: ...") from inside the kalimba branch, and delete the duplicate, unreachable dispatch block that followed it. The result is a single dispatch chain in which each of the 83 board ports is matched exactly once, terminated by the intended else() fallback that rejects unknown board names.
• scripts/check_board_config.sh: new configure-only regression check. It asserts that every board directory under boards/ configures successfully with -DEBLDR_BOARD=<name>, and that an unknown board name is still rejected with a fatal error. No cross toolchains are required (configure only, no compilation).

Testing

• Unit tests pass (ctest --test-dir build --output-on-failure)
• Integration tests pass
• Manual testing performed
• New tests added for new functionality

Verification performed (CMake 4.4.2, GCC 16.2.1, Linux x86_64 host build):

• Before the fix: cmake -DEBLDR_BOARD=kalimba aborts with CMake Error ... Unknown board: kalimba.
• After the fix: cmake -DEBLDR_BOARD=kalimba configures cleanly, and cmake --build <dir> --target board_kalimba compiles the port.
• All 83 boards under boards/ configure successfully (checked one by one).
• An unknown board (-DEBLDR_BOARD=does_not_exist) is still rejected with the fatal Unknown board error.
• scripts/check_board_config.sh reports RESULT: FAIL on unmodified master (it catches the kalimba regression) and RESULT: PASS with this PR applied.
• Native host build with -DEBLDR_BUILD_TESTS=ON and ctest give the same result as master: all tests pass except the pre-existing test_recovery link failure (undefined eos_boot_log_* symbols), which exists on master independently of this change and is addressed by separate open PRs.

Pre-Submission Checklist

• Code compiles without warnings (-Wall -Wextra -Werror for C)
• All existing tests pass
• New tests added for new functionality
• Documentation updated if API changed
• Commit messages follow the type(scope): description convention
• Branch is rebased on latest master

Related Issues

N/A. No open issue tracks this; found while exploring the codebase.

Screenshots / Logs

Reproduction on master:

$ cmake -B build -DEBLDR_BOARD=kalimba
--   Board port: kalimba
CMake Error at CMakeLists.txt:320 (message):
  Unknown board: kalimba.  See boards/ directory for available ports.
-- Configuring incomplete, errors occurred!

After the fix:

$ cmake -B build -DEBLDR_BOARD=kalimba
-- Configuring done
-- Generating done
$ scripts/check_board_config.sh
  ...
  kalimba              OK
  ...
  unknown board        OK (rejected as expected)
RESULT: PASS

Additional Notes

• No board behavior or public API changes; this only repairs the configure-time dispatch and removes dead code.
• The removed duplicate block was a strict subset of the first dispatch chain (verified: every branch it contained already appears earlier), so deleting it changes no reachable behavior.
• scripts/check_board_config.sh is ready to be wired into CI (for example the lint/host job in .github/workflows/build.yml) if maintainers want it to run automatically; I left that out to keep this change minimal.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.