Feature/sliding freq (#3280)

* Add sliding-frequency audio receiver tuning
* Move ADS-B receiver to external app
* Move AIS receiver to external app
* Move APRS receiver to external app
* Move APRS transmitter to external app
* Isolate filtered spectrum collection from WFM
* Fixed the issues #3280
* fpga_bridge.c: init registers per reference fpga_init, add quarter-shift mode setter
* radio.hpp: expose cached FPGA quarter-rate shift
* radio.cpp: program FPGA quarter-rate shift together with tuning offset
* tuning.hpp: pass AFE rate and direction to tuning config, add quarter_shift field
* tuning.cpp: import full PRALINE RX/TX tuning tables with quarter-shift offsets
* clock_manager: stop writing bogus RX digital gain, keep quarter shift across rate changes
* receiver_model: port LPF bandwidth from reference auto_bandwidth, use cached quarter shift
* adsb_rx: enable RF amp by default on first run
* ui_geomap.hpp: add hemisphere fields, degrees become magnitude
* ui_geomap: use hemisphere selector for lat/lon sign, fix minute/second wrap carry
* ui_menu: guard select against empty menu
* waterfall_designer: header updates for profile file handling
* waterfall_designer: exception-free profile parsing, CRLF handling, deferred nav callbacks, backup cleanup
* external.ld: scope app section globs to their own object directories
* CMakeLists: relink when external.ld changes
* tools: add external app symbol placement checker
* tools: add guru meditation address lookup script
* tools: add per-function stack usage report script
Co-authored-by: gullradriel <gullradriel@users.noreply.github.com>

firmware/baseband/dsp_decimate.hpp

@@ -170,6 +170,24 @@ class FIRC16xR16x16Decim2 {
     int32_t output_scale = 0;
 };
 
+class FIRC16xR16x63HalfbandDecim2 {
+   public:
+    static constexpr size_t taps_count = 63;
+    static constexpr size_t decimation_factor = 2;
+
+    void configure(const std::array<int16_t, taps_count>& taps);
+    void reset();
+
+    buffer_c16_t execute(
+        const buffer_c16_t& src,
+        const buffer_c16_t& dst);
+
+   private:
+    alignas(4) std::array<complex16_t, taps_count * 2> samples_{};
+    alignas(4) std::array<int16_t, taps_count> taps_{};
+    size_t samples_head_{0};
+};
+
 class FIRC16xR16x32Decim8 {
    public:
     static constexpr size_t taps_count = 32;
@@ -210,6 +228,23 @@ class FIRAndDecimateComplex {
         configure(taps.data(), taps.size(), decimation_factor);
     }
 
+    template <size_t N>
+    void configure(
+        const std::array<int16_t, N>& taps,
+        const size_t decimation_factor) {
+        configure_common(N, decimation_factor);
+        for (size_t i = 0; i < N; ++i) {
+            taps_reversed_[i] = {taps[N - 1 - i], 0};
+        }
+    }
+
+    template <size_t N>
+    void set_taps(const std::array<complex16_t, N>& taps) {
+        if (N == taps_count_) {
+            std::reverse_copy(taps.begin(), taps.end(), &taps_reversed_[0]);
+        }
+    }
+
     buffer_c16_t execute(
         const buffer_c16_t& src,
         const buffer_c16_t& dst);
@@ -221,6 +256,7 @@ class FIRAndDecimateComplex {
     std::unique_ptr<taps_t> taps_reversed_{};
     size_t taps_count_{0};
     size_t decimation_factor_{1};
+    size_t samples_head_{0};
 
     template <typename T>
     void configure(