Merge branch 'enum_mask_helper' into integration

esphome/components/climate/climate.cpp

@@ -524,13 +524,23 @@ ClimateCall ClimateDeviceRestoreState::to_call(Climate *climate) {
   if (traits.has_feature_flags(climate::CLIMATE_SUPPORTS_TARGET_HUMIDITY)) {
     call.set_target_humidity(this->target_humidity);
   }
-  if (traits.get_supports_fan_modes() || !traits.get_supported_custom_fan_modes().empty()) {
+  if (this->uses_custom_fan_mode) {
+    if (this->custom_fan_mode < traits.get_supported_custom_fan_modes().size()) {
+      call.fan_mode_.reset();
+      call.custom_fan_mode_ = *std::next(traits.get_supported_custom_fan_modes().cbegin(), this->custom_fan_mode);
+    }
+  } else if (traits.supports_fan_mode(this->fan_mode)) {
     call.set_fan_mode(this->fan_mode);
   }
-  if (traits.get_supports_presets() || !traits.get_supported_custom_presets().empty()) {
+  if (this->uses_custom_preset) {
+    if (this->custom_preset < traits.get_supported_custom_presets().size()) {
+      call.preset_.reset();
+      call.custom_preset_ = *std::next(traits.get_supported_custom_presets().cbegin(), this->custom_preset);
+    }
+  } else if (traits.supports_preset(this->preset)) {
     call.set_preset(this->preset);
   }
-  if (traits.get_supports_swing_modes()) {
+  if (traits.supports_swing_mode(this->swing_mode)) {
     call.set_swing_mode(this->swing_mode);
   }
   return call;
@@ -549,41 +559,25 @@ void ClimateDeviceRestoreState::apply(Climate *climate) {
   if (traits.has_feature_flags(climate::CLIMATE_SUPPORTS_TARGET_HUMIDITY)) {
     climate->target_humidity = this->target_humidity;
   }
-  if (traits.get_supports_fan_modes() && !this->uses_custom_fan_mode) {
+  if (this->uses_custom_fan_mode) {
+    if (this->custom_fan_mode < traits.get_supported_custom_fan_modes().size()) {
+      climate->fan_mode.reset();
+      climate->custom_fan_mode = *std::next(traits.get_supported_custom_fan_modes().cbegin(), this->custom_fan_mode);
+    }
+  } else if (traits.supports_fan_mode(this->fan_mode)) {
     climate->fan_mode = this->fan_mode;
+    climate->custom_fan_mode.reset();
   }
-  if (!traits.get_supported_custom_fan_modes().empty() && this->uses_custom_fan_mode) {
+  if (this->uses_custom_preset) {
-    // std::set has consistent order (lexicographic for strings)
+    if (this->custom_preset < traits.get_supported_custom_presets().size()) {
-    const auto &modes = traits.get_supported_custom_fan_modes();
+      climate->preset.reset();
-    if (custom_fan_mode < modes.size()) {
+      climate->custom_preset = *std::next(traits.get_supported_custom_presets().cbegin(), this->custom_preset);
-      size_t i = 0;
-      for (const auto &mode : modes) {
-        if (i == this->custom_fan_mode) {
-          climate->custom_fan_mode = mode;
-          break;
-        }
-        i++;
-      }
     }
-  }
+  } else if (traits.supports_preset(this->preset)) {
-  if (traits.get_supports_presets() && !this->uses_custom_preset) {
     climate->preset = this->preset;
+    climate->custom_preset.reset();
   }
-  if (!traits.get_supported_custom_presets().empty() && uses_custom_preset) {
+  if (traits.supports_swing_mode(this->swing_mode)) {
-    // std::set has consistent order (lexicographic for strings)
-    const auto &presets = traits.get_supported_custom_presets();
-    if (custom_preset < presets.size()) {
-      size_t i = 0;
-      for (const auto &preset : presets) {
-        if (i == this->custom_preset) {
-          climate->custom_preset = preset;
-          break;
-        }
-        i++;
-      }
-    }
-  }
-  if (traits.get_supports_swing_modes()) {
     climate->swing_mode = this->swing_mode;
   }
   climate->publish_state();

esphome/components/climate/climate.h

@@ -33,6 +33,7 @@ class Climate;
 class ClimateCall {
  public:
   explicit ClimateCall(Climate *parent) : parent_(parent) {}
+  friend struct ClimateDeviceRestoreState;
 
   /// Set the mode of the climate device.
   ClimateCall &set_mode(ClimateMode mode);

esphome/components/esp32/__init__.py

@@ -877,6 +877,11 @@ async def to_code(config):
     for clean_var in ("IDF_PATH", "IDF_TOOLS_PATH"):
         os.environ.pop(clean_var, None)
 
+    # Set the location of the IDF component manager cache
+    os.environ["IDF_COMPONENT_CACHE_PATH"] = str(
+        CORE.relative_internal_path(".espressif")
+    )
+
     add_extra_script(
         "post",
         "post_build.py",

esphome/components/light/addressable_light.cpp

@@ -61,6 +61,10 @@ void AddressableLightTransformer::start() {
   this->target_color_ *= to_uint8_scale(end_values.get_brightness() * end_values.get_state());
 }
 
+inline constexpr uint8_t subtract_scaled_difference(uint8_t a, uint8_t b, int32_t scale) {
+  return uint8_t(int32_t(a) - (((int32_t(a) - int32_t(b)) * scale) / 256));
+}
+
 optional<LightColorValues> AddressableLightTransformer::apply() {
   float smoothed_progress = LightTransformer::smoothed_progress(this->get_progress_());
 
@@ -74,38 +78,37 @@ optional<LightColorValues> AddressableLightTransformer::apply() {
   // all LEDs, we use the current state of each LED as the start.
 
   // We can't use a direct lerp smoothing here though - that would require creating a copy of the original
-  // state of each LED at the start of the transition.
+  // state of each LED at the start of the transition. Instead, we "fake" the look of lerp by calculating
-  // Instead, we "fake" the look of the LERP by using an exponential average over time and using
+  // the delta between the current state and the target state, assuming that the delta represents the rest
-  // dynamically-calculated alpha values to match the look.
+  // of the transition that was to be applied as of the previous transition step, and scaling the delta for
+  // what should be left after the current transition step. In this manner, the delta decays to zero as the
+  // transition progresses.
+  //
+  // Here's an example of how the algorithm progresses in discrete steps:
+  //
+  // At time = 0.00, 0% complete, 100% remaining, 100% will remain after this step, so the scale is 100% / 100% = 100%.
+  // At time = 0.10, 0% complete, 100% remaining, 90% will remain after this step, so the scale is 90% / 100% = 90%.
+  // At time = 0.20, 10% complete, 90% remaining, 80% will remain after this step, so the scale is 80% / 90% = 88.9%.
+  // At time = 0.50, 20% complete, 80% remaining, 50% will remain after this step, so the scale is 50% / 80% = 62.5%.
+  // At time = 0.90, 50% complete, 50% remaining, 10% will remain after this step, so the scale is 10% / 50% = 20%.
+  // At time = 0.91, 90% complete, 10% remaining, 9% will remain after this step, so the scale is 9% / 10% = 90%.
+  // At time = 1.00, 91% complete, 9% remaining, 0% will remain after this step, so the scale is 0% / 9% = 0%.
+  //
+  // Because the color values are quantized to 8 bit resolution after each step, the transition may appear
+  // non-linear when applying small deltas.
 
-  float denom = (1.0f - smoothed_progress);
+  if (smoothed_progress > this->last_transition_progress_ && this->last_transition_progress_ < 1.f) {
-  float alpha = denom == 0.0f ? 1.0f : (smoothed_progress - this->last_transition_progress_) / denom;
+    int32_t scale = int32_t(256.f * std::max((1.f - smoothed_progress) / (1.f - this->last_transition_progress_), 0.f));
-
+    for (auto led : this->light_) {
-  // We need to use a low-resolution alpha here which makes the transition set in only after ~half of the length
+      led.set_rgbw(subtract_scaled_difference(this->target_color_.red, led.get_red(), scale),
-  // We solve this by accumulating the fractional part of the alpha over time.
+                   subtract_scaled_difference(this->target_color_.green, led.get_green(), scale),
-  float alpha255 = alpha * 255.0f;
+                   subtract_scaled_difference(this->target_color_.blue, led.get_blue(), scale),
-  float alpha255int = floorf(alpha255);
+                   subtract_scaled_difference(this->target_color_.white, led.get_white(), scale));
-  float alpha255remainder = alpha255 - alpha255int;
+    }
-
+    this->last_transition_progress_ = smoothed_progress;
-  this->accumulated_alpha_ += alpha255remainder;
+    this->light_.schedule_show();
-  float alpha_add = floorf(this->accumulated_alpha_);
-  this->accumulated_alpha_ -= alpha_add;
-
-  alpha255 += alpha_add;
-  alpha255 = clamp(alpha255, 0.0f, 255.0f);
-  auto alpha8 = static_cast<uint8_t>(alpha255);
-
-  if (alpha8 != 0) {
-    uint8_t inv_alpha8 = 255 - alpha8;
-    Color add = this->target_color_ * alpha8;
-
-    for (auto led : this->light_)
-      led.set(add + led.get() * inv_alpha8);
   }
 
-  this->last_transition_progress_ = smoothed_progress;
-  this->light_.schedule_show();
-
   return {};
 }
 

esphome/components/light/addressable_light.h

@@ -113,7 +113,6 @@ class AddressableLightTransformer : public LightTransformer {
  protected:
   AddressableLight &light_;
   float last_transition_progress_{0.0f};
-  float accumulated_alpha_{0.0f};
   Color target_color_{};
 };
 

esphome/components/light/color_mode.h

@@ -1,7 +1,7 @@
 #pragma once
 
 #include <cstdint>
-#include "esphome/core/enum_bitmask.h"
+#include "esphome/core/finite_set_mask.h"
 
 namespace esphome {
 namespace light {
@@ -105,7 +105,7 @@ constexpr ColorModeHelper operator|(ColorModeHelper lhs, ColorMode rhs) {
   return static_cast<ColorMode>(static_cast<uint8_t>(lhs) | static_cast<uint8_t>(rhs));
 }
 
-// Type alias for raw color mode bitmask values (retained for compatibility)
+// Type alias for raw color mode bitmask values
 using color_mode_bitmask_t = uint16_t;
 
 // Number of ColorMode enum values
@@ -127,8 +127,8 @@ constexpr ColorMode COLOR_MODE_LOOKUP[COLOR_MODE_BITMASK_SIZE] = {
     ColorMode::RGB_COLD_WARM_WHITE,    // bit 9
 };
 
-// Type alias for ColorMode bitmask using generic EnumBitmask template
+// Type alias for ColorMode bitmask using generic FiniteSetMask template
-using ColorModeMask = EnumBitmask<ColorMode, COLOR_MODE_BITMASK_SIZE>;
+using ColorModeMask = FiniteSetMask<ColorMode, COLOR_MODE_BITMASK_SIZE>;
 
 // Number of ColorCapability enum values
 constexpr int COLOR_CAPABILITY_COUNT = 6;
@@ -159,16 +159,21 @@ constexpr uint16_t CAPABILITY_BITMASKS[] = {
     compute_capability_bitmask(ColorCapability::RGB),                // 1 << 5
 };
 
-/// Check if any mode in the bitmask has a specific capability
+/**
-/// Used for checking if a light supports a capability (e.g., BRIGHTNESS, RGB)
+ * @brief Helper function to convert a power-of-2 ColorCapability value to an array index for CAPABILITY_BITMASKS
-inline bool has_capability(const ColorModeMask &mask, ColorCapability capability) {
+ * lookup.
-  // Lookup the pre-computed bitmask for this capability and check intersection with our mask
+ *
-  // ColorCapability values: 1, 2, 4, 8, 16, 32 -> array indices: 0, 1, 2, 3, 4, 5
+ * This function maps ColorCapability values (1, 2, 4, 8, 16, 32) to array indices (0, 1, 2, 3, 4, 5).
-  // We need to convert the power-of-2 value to an index
+ * Used to index into the CAPABILITY_BITMASKS lookup table.
+ *
+ * @param capability A ColorCapability enum value (must be a power of 2).
+ * @return The corresponding array index (0-based).
+ */
+inline int capability_to_index(ColorCapability capability) {
   uint8_t cap_val = static_cast<uint8_t>(capability);
 #if defined(__GNUC__) || defined(__clang__)
   // Use compiler intrinsic for efficient bit position lookup (O(1) vs O(log n))
-  int index = __builtin_ctz(cap_val);
+  return __builtin_ctz(cap_val);
 #else
   // Fallback for compilers without __builtin_ctz
   int index = 0;
@@ -176,8 +181,15 @@ inline bool has_capability(const ColorModeMask &mask, ColorCapability capability
     cap_val >>= 1;
     ++index;
   }
+  return index;
 #endif
-  return (mask.get_mask() & CAPABILITY_BITMASKS[index]) != 0;
+}
+
+/// Check if any mode in the bitmask has a specific capability
+/// Used for checking if a light supports a capability (e.g., BRIGHTNESS, RGB)
+inline bool has_capability(const ColorModeMask &mask, ColorCapability capability) {
+  // Lookup the pre-computed bitmask for this capability and check intersection with our mask
+  return (mask.get_mask() & CAPABILITY_BITMASKS[capability_to_index(capability)]) != 0;
 }
 
 }  // namespace light
@@ -186,7 +198,7 @@ inline bool has_capability(const ColorModeMask &mask, ColorCapability capability
 // Template specializations for ColorMode must be in global namespace
 //
 // C++ requires template specializations to be declared in the same namespace as the
-// original template. Since EnumBitmask is in the esphome namespace (not esphome::light),
+// original template. Since FiniteSetMask is in the esphome namespace (not esphome::light),
 // we must provide these specializations at global scope with fully-qualified names.
 //
 // These specializations define how ColorMode enum values map to/from bit positions.
@@ -194,7 +206,7 @@ inline bool has_capability(const ColorModeMask &mask, ColorCapability capability
 /// Map ColorMode enum values to bit positions (0-9)
 /// Bit positions follow the enum declaration order
 template<>
-constexpr int esphome::EnumBitmask<esphome::light::ColorMode, esphome::light::COLOR_MODE_BITMASK_SIZE>::enum_to_bit(
+constexpr int esphome::FiniteSetMask<esphome::light::ColorMode, esphome::light::COLOR_MODE_BITMASK_SIZE>::value_to_bit(
     esphome::light::ColorMode mode) {
   // Linear search through COLOR_MODE_LOOKUP array
   // Compiler optimizes this to efficient code since array is constexpr
@@ -208,8 +220,8 @@ constexpr int esphome::EnumBitmask<esphome::light::ColorMode, esphome::light::CO
 /// Map bit positions (0-9) to ColorMode enum values
 /// Bit positions follow the enum declaration order
 template<>
-inline esphome::light::ColorMode esphome::EnumBitmask<esphome::light::ColorMode,
+inline esphome::light::ColorMode esphome::FiniteSetMask<
-                                                      esphome::light::COLOR_MODE_BITMASK_SIZE>::bit_to_enum(int bit) {
+    esphome::light::ColorMode, esphome::light::COLOR_MODE_BITMASK_SIZE>::bit_to_value(int bit) {
   return (bit >= 0 && bit < esphome::light::COLOR_MODE_BITMASK_SIZE) ? esphome::light::COLOR_MODE_LOOKUP[bit]
                                                                      : esphome::light::ColorMode::UNKNOWN;
 }

esphome/core/enum_bitmask.h

@@ -1,159 +0,0 @@
-#pragma once
-
-#include <cstddef>
-#include <cstdint>
-#include <initializer_list>
-#include <iterator>
-#include <type_traits>
-
-namespace esphome {
-
-/// Generic bitmask for storing a set of enum values efficiently.
-/// Replaces std::set<EnumType> to eliminate red-black tree overhead (~586 bytes per instantiation).
-///
-/// Template parameters:
-///   EnumType: The enum type to store (must be uint8_t-based)
-///   MaxBits: Maximum number of bits needed (auto-selects uint8_t/uint16_t/uint32_t)
-///
-/// Requirements:
-///   - EnumType must be an enum with sequential values starting from 0
-///   - Specialization must provide enum_to_bit() and bit_to_enum() static methods
-///   - MaxBits must be sufficient to hold all enum values
-///
-/// Example usage:
-///   using ClimateModeMask = EnumBitmask<ClimateMode, 8>;
-///   ClimateModeMask modes({CLIMATE_MODE_HEAT, CLIMATE_MODE_COOL});
-///   if (modes.count(CLIMATE_MODE_HEAT)) { ... }
-///   for (auto mode : modes) { ... }  // Iterate over set bits
-///
-/// For complete usage examples with template specializations, see:
-///   - esphome/components/light/color_mode.h (ColorMode example)
-///
-/// Design notes:
-///   - Uses compile-time type selection for optimal size (uint8_t/uint16_t/uint32_t)
-///   - Iterator converts bit positions to actual enum values during traversal
-///   - All operations are constexpr-compatible for compile-time initialization
-///   - Drop-in replacement for std::set<EnumType> with simpler API
-///
-template<typename EnumType, int MaxBits = 16> class EnumBitmask {
- public:
-  // Automatic bitmask type selection based on MaxBits
-  // ≤8 bits: uint8_t, ≤16 bits: uint16_t, otherwise: uint32_t
-  using bitmask_t =
-      typename std::conditional<(MaxBits <= 8), uint8_t,
-                                typename std::conditional<(MaxBits <= 16), uint16_t, uint32_t>::type>::type;
-
-  constexpr EnumBitmask() = default;
-
-  /// Construct from initializer list: {VALUE1, VALUE2, ...}
-  constexpr EnumBitmask(std::initializer_list<EnumType> values) {
-    for (auto value : values) {
-      this->insert(value);
-    }
-  }
-
-  /// Add a single enum value to the set (std::set compatibility)
-  constexpr void insert(EnumType value) { this->mask_ |= (static_cast<bitmask_t>(1) << enum_to_bit(value)); }
-
-  /// Add multiple enum values from initializer list
-  constexpr void insert(std::initializer_list<EnumType> values) {
-    for (auto value : values) {
-      this->insert(value);
-    }
-  }
-
-  /// Remove an enum value from the set (std::set compatibility)
-  constexpr void erase(EnumType value) { this->mask_ &= ~(static_cast<bitmask_t>(1) << enum_to_bit(value)); }
-
-  /// Clear all values from the set
-  constexpr void clear() { this->mask_ = 0; }
-
-  /// Check if the set contains a specific enum value (std::set compatibility)
-  /// Returns 1 if present, 0 if not (same as std::set for unique elements)
-  constexpr size_t count(EnumType value) const {
-    return (this->mask_ & (static_cast<bitmask_t>(1) << enum_to_bit(value))) != 0 ? 1 : 0;
-  }
-
-  /// Count the number of enum values in the set
-  constexpr size_t size() const {
-    // Brian Kernighan's algorithm - efficient for sparse bitmasks
-    // Typical case: 2-4 modes out of 10 possible
-    bitmask_t n = this->mask_;
-    size_t count = 0;
-    while (n) {
-      n &= n - 1;  // Clear the least significant set bit
-      count++;
-    }
-    return count;
-  }
-
-  /// Check if the set is empty
-  constexpr bool empty() const { return this->mask_ == 0; }
-
-  /// Iterator support for range-based for loops and API encoding
-  /// Iterates over set bits and converts bit positions to enum values
-  class Iterator {
-   public:
-    using iterator_category = std::forward_iterator_tag;
-    using value_type = EnumType;
-    using difference_type = std::ptrdiff_t;
-    using pointer = const EnumType *;
-    using reference = EnumType;
-
-    constexpr Iterator(bitmask_t mask, int bit) : mask_(mask), bit_(bit) { advance_to_next_set_bit_(); }
-
-    constexpr EnumType operator*() const { return bit_to_enum(bit_); }
-
-    constexpr Iterator &operator++() {
-      ++bit_;
-      advance_to_next_set_bit_();
-      return *this;
-    }
-
-    constexpr bool operator==(const Iterator &other) const { return bit_ == other.bit_; }
-
-    constexpr bool operator!=(const Iterator &other) const { return !(*this == other); }
-
-   private:
-    constexpr void advance_to_next_set_bit_() { bit_ = find_next_set_bit(mask_, bit_); }
-
-    bitmask_t mask_;
-    int bit_;
-  };
-
-  constexpr Iterator begin() const { return Iterator(mask_, 0); }
-  constexpr Iterator end() const { return Iterator(mask_, MaxBits); }
-
-  /// Get the raw bitmask value for optimized operations
-  constexpr bitmask_t get_mask() const { return this->mask_; }
-
-  /// Check if a specific enum value is present in a raw bitmask
-  /// Useful for checking intersection results without creating temporary objects
-  static constexpr bool mask_contains(bitmask_t mask, EnumType value) {
-    return (mask & (static_cast<bitmask_t>(1) << enum_to_bit(value))) != 0;
-  }
-
-  /// Get the first enum value from a raw bitmask
-  /// Used for optimizing intersection logic (e.g., "pick first suitable mode")
-  static constexpr EnumType first_value_from_mask(bitmask_t mask) { return bit_to_enum(find_next_set_bit(mask, 0)); }
-
-  /// Find the next set bit in a bitmask starting from a given position
-  /// Returns the bit position, or MaxBits if no more bits are set
-  static constexpr int find_next_set_bit(bitmask_t mask, int start_bit) {
-    int bit = start_bit;
-    while (bit < MaxBits && !(mask & (static_cast<bitmask_t>(1) << bit))) {
-      ++bit;
-    }
-    return bit;
-  }
-
- protected:
-  // Must be provided by template specialization
-  // These convert between enum values and bit positions (0, 1, 2, ...)
-  static constexpr int enum_to_bit(EnumType value);
-  static EnumType bit_to_enum(int bit);  // Not constexpr due to static array limitation in C++20
-
-  bitmask_t mask_{0};
-};
-
-}  // namespace esphome

esphome/core/finite_set_mask.h

@@ -0,0 +1,165 @@
+#pragma once
+
+#include <cstddef>
+#include <cstdint>
+#include <initializer_list>
+#include <iterator>
+#include <type_traits>
+
+namespace esphome {
+
+/// Generic bitmask for storing a finite set of discrete values efficiently.
+/// Replaces std::set<ValueType> to eliminate red-black tree overhead (~586 bytes per instantiation).
+///
+/// Template parameters:
+///   ValueType: The type to store (typically enum, but can be any discrete bounded type)
+///   MaxBits: Maximum number of bits needed (auto-selects uint8_t/uint16_t/uint32_t)
+///
+/// Requirements:
+///   - ValueType must have a bounded discrete range that maps to bit positions
+///   - For 1:1 mappings (contiguous enums starting at 0), no specialization needed
+///   - For custom mappings (like ColorMode), specialize value_to_bit() and/or bit_to_value()
+///   - MaxBits must be sufficient to hold all possible values
+///
+/// Example usage (1:1 mapping - climate enums):
+///   // For enums with contiguous values starting at 0, no specialization needed!
+///   using ClimateModeMask = FiniteSetMask<ClimateMode, CLIMATE_MODE_AUTO + 1>;
+///   ClimateModeMask modes({CLIMATE_MODE_HEAT, CLIMATE_MODE_COOL});
+///   if (modes.count(CLIMATE_MODE_HEAT)) { ... }
+///   for (auto mode : modes) { ... }  // Iterate over set bits
+///
+/// Example usage (custom mapping - ColorMode):
+///   // For non-contiguous enums or custom mappings, specialize value_to_bit() and/or bit_to_value()
+///   // See esphome/components/light/color_mode.h for complete example
+///
+/// Design notes:
+///   - Uses compile-time type selection for optimal size (uint8_t/uint16_t/uint32_t)
+///   - Iterator converts bit positions to actual values during traversal
+///   - All operations are constexpr-compatible for compile-time initialization
+///   - Drop-in replacement for std::set<ValueType> with simpler API
+///   - Despite the name, works with any discrete bounded type, not just enums
+///
+template<typename ValueType, int MaxBits = 16> class FiniteSetMask {
+ public:
+  // Automatic bitmask type selection based on MaxBits
+  // ≤8 bits: uint8_t, ≤16 bits: uint16_t, otherwise: uint32_t
+  using bitmask_t =
+      typename std::conditional<(MaxBits <= 8), uint8_t,
+                                typename std::conditional<(MaxBits <= 16), uint16_t, uint32_t>::type>::type;
+
+  constexpr FiniteSetMask() = default;
+
+  /// Construct from initializer list: {VALUE1, VALUE2, ...}
+  constexpr FiniteSetMask(std::initializer_list<ValueType> values) {
+    for (auto value : values) {
+      this->insert(value);
+    }
+  }
+
+  /// Add a single value to the set (std::set compatibility)
+  constexpr void insert(ValueType value) { this->mask_ |= (static_cast<bitmask_t>(1) << value_to_bit(value)); }
+
+  /// Add multiple values from initializer list
+  constexpr void insert(std::initializer_list<ValueType> values) {
+    for (auto value : values) {
+      this->insert(value);
+    }
+  }
+
+  /// Remove a value from the set (std::set compatibility)
+  constexpr void erase(ValueType value) { this->mask_ &= ~(static_cast<bitmask_t>(1) << value_to_bit(value)); }
+
+  /// Clear all values from the set
+  constexpr void clear() { this->mask_ = 0; }
+
+  /// Check if the set contains a specific value (std::set compatibility)
+  /// Returns 1 if present, 0 if not (same as std::set for unique elements)
+  constexpr size_t count(ValueType value) const {
+    return (this->mask_ & (static_cast<bitmask_t>(1) << value_to_bit(value))) != 0 ? 1 : 0;
+  }
+
+  /// Count the number of values in the set
+  constexpr size_t size() const {
+    // Brian Kernighan's algorithm - efficient for sparse bitmasks
+    // Typical case: 2-4 modes out of 10 possible
+    bitmask_t n = this->mask_;
+    size_t count = 0;
+    while (n) {
+      n &= n - 1;  // Clear the least significant set bit
+      count++;
+    }
+    return count;
+  }
+
+  /// Check if the set is empty
+  constexpr bool empty() const { return this->mask_ == 0; }
+
+  /// Iterator support for range-based for loops and API encoding
+  /// Iterates over set bits and converts bit positions to values
+  /// Optimization: removes bits from mask as we iterate
+  class Iterator {
+   public:
+    using iterator_category = std::forward_iterator_tag;
+    using value_type = ValueType;
+    using difference_type = std::ptrdiff_t;
+    using pointer = const ValueType *;
+    using reference = ValueType;
+
+    constexpr explicit Iterator(bitmask_t mask) : mask_(mask) {}
+
+    constexpr ValueType operator*() const {
+      // Return value for the first set bit
+      return bit_to_value(find_next_set_bit(mask_, 0));
+    }
+
+    constexpr Iterator &operator++() {
+      // Clear the lowest set bit (Brian Kernighan's algorithm)
+      mask_ &= mask_ - 1;
+      return *this;
+    }
+
+    constexpr bool operator==(const Iterator &other) const { return mask_ == other.mask_; }
+
+    constexpr bool operator!=(const Iterator &other) const { return !(*this == other); }
+
+   private:
+    bitmask_t mask_;
+  };
+
+  constexpr Iterator begin() const { return Iterator(mask_); }
+  constexpr Iterator end() const { return Iterator(0); }
+
+  /// Get the raw bitmask value for optimized operations
+  constexpr bitmask_t get_mask() const { return this->mask_; }
+
+  /// Check if a specific value is present in a raw bitmask
+  /// Useful for checking intersection results without creating temporary objects
+  static constexpr bool mask_contains(bitmask_t mask, ValueType value) {
+    return (mask & (static_cast<bitmask_t>(1) << value_to_bit(value))) != 0;
+  }
+
+  /// Get the first value from a raw bitmask
+  /// Used for optimizing intersection logic (e.g., "pick first suitable mode")
+  static constexpr ValueType first_value_from_mask(bitmask_t mask) { return bit_to_value(find_next_set_bit(mask, 0)); }
+
+  /// Find the next set bit in a bitmask starting from a given position
+  /// Returns the bit position, or MaxBits if no more bits are set
+  static constexpr int find_next_set_bit(bitmask_t mask, int start_bit) {
+    int bit = start_bit;
+    while (bit < MaxBits && !(mask & (static_cast<bitmask_t>(1) << bit))) {
+      ++bit;
+    }
+    return bit;
+  }
+
+ protected:
+  // Default implementations for 1:1 mapping (enum value = bit position)
+  // For enums with contiguous values starting at 0, these defaults work as-is.
+  // If you need custom mapping (like ColorMode), provide specializations.
+  static constexpr int value_to_bit(ValueType value) { return static_cast<int>(value); }
+  static constexpr ValueType bit_to_value(int bit) { return static_cast<ValueType>(bit); }
+
+  bitmask_t mask_{0};
+};
+
+}  // namespace esphome

script/determine-jobs.py

@@ -336,7 +336,7 @@ def _component_has_tests(component: str) -> bool:
     Returns:
         True if the component has test YAML files
     """
-    return bool(get_component_test_files(component))
+    return bool(get_component_test_files(component, all_variants=True))
 
 
 def _select_platform_by_preference(
@@ -496,7 +496,7 @@ def detect_memory_impact_config(
 
     for component in sorted(changed_component_set):
         # Look for test files on preferred platforms
-        test_files = get_component_test_files(component)
+        test_files = get_component_test_files(component, all_variants=True)
         if not test_files:
             continue
 

script/split_components_for_ci.py

@@ -49,9 +49,9 @@ def has_test_files(component_name: str, tests_dir: Path) -> bool:
         tests_dir: Path to tests/components directory (unused, kept for compatibility)
 
     Returns:
-        True if the component has test.*.yaml files
+        True if the component has test.*.yaml or test-*.yaml files
     """
-    return bool(get_component_test_files(component_name))
+    return bool(get_component_test_files(component_name, all_variants=True))
 
 
 def create_intelligent_batches(

tests/script/test_determine_jobs.py

@@ -574,6 +574,105 @@ def test_main_filters_components_without_tests(
     assert output["memory_impact"]["should_run"] == "false"
 
 
+def test_main_detects_components_with_variant_tests(
+    mock_should_run_integration_tests: Mock,
+    mock_should_run_clang_tidy: Mock,
+    mock_should_run_clang_format: Mock,
+    mock_should_run_python_linters: Mock,
+    mock_changed_files: Mock,
+    capsys: pytest.CaptureFixture[str],
+    tmp_path: Path,
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """Test that components with only variant test files (test-*.yaml) are detected.
+
+    This test verifies the fix for components like improv_serial, ethernet, mdns,
+    improv_base, and safe_mode which only have variant test files (test-*.yaml)
+    instead of base test files (test.*.yaml).
+    """
+    # Ensure we're not in GITHUB_ACTIONS mode for this test
+    monkeypatch.delenv("GITHUB_ACTIONS", raising=False)
+
+    mock_should_run_integration_tests.return_value = False
+    mock_should_run_clang_tidy.return_value = False
+    mock_should_run_clang_format.return_value = False
+    mock_should_run_python_linters.return_value = False
+
+    # Mock changed_files to return component files
+    mock_changed_files.return_value = [
+        "esphome/components/improv_serial/improv_serial.cpp",
+        "esphome/components/ethernet/ethernet.cpp",
+        "esphome/components/no_tests/component.cpp",
+    ]
+
+    # Create test directory structure
+    tests_dir = tmp_path / "tests" / "components"
+
+    # improv_serial has only variant tests (like the real component)
+    improv_serial_dir = tests_dir / "improv_serial"
+    improv_serial_dir.mkdir(parents=True)
+    (improv_serial_dir / "test-uart0.esp32-idf.yaml").write_text("test: config")
+    (improv_serial_dir / "test-uart0.esp8266-ard.yaml").write_text("test: config")
+    (improv_serial_dir / "test-usb_cdc.esp32-s2-idf.yaml").write_text("test: config")
+
+    # ethernet also has only variant tests
+    ethernet_dir = tests_dir / "ethernet"
+    ethernet_dir.mkdir(parents=True)
+    (ethernet_dir / "test-manual_ip.esp32-idf.yaml").write_text("test: config")
+    (ethernet_dir / "test-dhcp.esp32-idf.yaml").write_text("test: config")
+
+    # no_tests component has no test files at all
+    no_tests_dir = tests_dir / "no_tests"
+    no_tests_dir.mkdir(parents=True)
+
+    # Mock root_path to use tmp_path (need to patch both determine_jobs and helpers)
+    with (
+        patch.object(determine_jobs, "root_path", str(tmp_path)),
+        patch.object(helpers, "root_path", str(tmp_path)),
+        patch("sys.argv", ["determine-jobs.py"]),
+        patch.object(
+            determine_jobs,
+            "get_changed_components",
+            return_value=["improv_serial", "ethernet", "no_tests"],
+        ),
+        patch.object(
+            determine_jobs,
+            "filter_component_and_test_files",
+            side_effect=lambda f: f.startswith("esphome/components/"),
+        ),
+        patch.object(
+            determine_jobs,
+            "get_components_with_dependencies",
+            side_effect=lambda files, deps: (
+                ["improv_serial", "ethernet"]
+                if not deps
+                else ["improv_serial", "ethernet", "no_tests"]
+            ),
+        ),
+        patch.object(determine_jobs, "changed_files", return_value=[]),
+    ):
+        # Clear the cache since we're mocking root_path
+        determine_jobs._component_has_tests.cache_clear()
+        determine_jobs.main()
+
+    # Check output
+    captured = capsys.readouterr()
+    output = json.loads(captured.out)
+
+    # changed_components should have all components
+    assert set(output["changed_components"]) == {
+        "improv_serial",
+        "ethernet",
+        "no_tests",
+    }
+    # changed_components_with_tests should include components with variant tests
+    assert set(output["changed_components_with_tests"]) == {"improv_serial", "ethernet"}
+    # component_test_count should be 2 (improv_serial and ethernet)
+    assert output["component_test_count"] == 2
+    # no_tests should be excluded since it has no test files
+    assert "no_tests" not in output["changed_components_with_tests"]
+
+
 # Tests for detect_memory_impact_config function
 
 
@@ -785,6 +884,51 @@ def test_detect_memory_impact_config_skips_base_bus_components(tmp_path: Path) -
     assert "i2c" not in result["components"]
 
 
+def test_detect_memory_impact_config_with_variant_tests(tmp_path: Path) -> None:
+    """Test memory impact detection for components with only variant test files.
+
+    This verifies that memory impact analysis works correctly for components like
+    improv_serial, ethernet, mdns, etc. which only have variant test files
+    (test-*.yaml) instead of base test files (test.*.yaml).
+    """
+    # Create test directory structure
+    tests_dir = tmp_path / "tests" / "components"
+
+    # improv_serial with only variant tests
+    improv_serial_dir = tests_dir / "improv_serial"
+    improv_serial_dir.mkdir(parents=True)
+    (improv_serial_dir / "test-uart0.esp32-idf.yaml").write_text("test: improv")
+    (improv_serial_dir / "test-uart0.esp8266-ard.yaml").write_text("test: improv")
+    (improv_serial_dir / "test-usb_cdc.esp32-s2-idf.yaml").write_text("test: improv")
+
+    # ethernet with only variant tests
+    ethernet_dir = tests_dir / "ethernet"
+    ethernet_dir.mkdir(parents=True)
+    (ethernet_dir / "test-manual_ip.esp32-idf.yaml").write_text("test: ethernet")
+    (ethernet_dir / "test-dhcp.esp32-c3-idf.yaml").write_text("test: ethernet")
+
+    # Mock changed_files to return both components
+    with (
+        patch.object(determine_jobs, "root_path", str(tmp_path)),
+        patch.object(helpers, "root_path", str(tmp_path)),
+        patch.object(determine_jobs, "changed_files") as mock_changed_files,
+    ):
+        mock_changed_files.return_value = [
+            "esphome/components/improv_serial/improv_serial.cpp",
+            "esphome/components/ethernet/ethernet.cpp",
+        ]
+        determine_jobs._component_has_tests.cache_clear()
+
+        result = determine_jobs.detect_memory_impact_config()
+
+    # Should detect both components even though they only have variant tests
+    assert result["should_run"] == "true"
+    assert set(result["components"]) == {"improv_serial", "ethernet"}
+    # Both components support esp32-idf
+    assert result["platform"] == "esp32-idf"
+    assert result["use_merged_config"] == "true"
+
+
 # Tests for clang-tidy split mode logic