Lex every build flag entry, warn on custom MMU sizes without the knob, note the scanf omission

esphome/build_gen/arduino8266.py

@@ -112,6 +112,8 @@ _CCFLAGS = [
     "-free",
     "-fipa-pta",
 ]
+# Upstream's -u _scanf_float is deliberately absent: it is re-added from
+# KEY_SCANF_FLOAT at emission (the remove_float_scanf extra script's job).
 _LINKFLAGS = [
     "-Os",
     "-nostdlib",
@@ -166,14 +168,10 @@ def _flag_defines() -> dict[str, str]:
     """Map define name -> full ``NAME[=VALUE]`` for every -D build flag."""
     defines: dict[str, str] = {}
     for flag in CORE.build_flags:
-        # Shell-lex multi-token entries the way PlatformIO does, so a knob
-        # in "-DKNOB -DOTHER" or a spaced "-D KNOB" is still detected;
-        # single tokens pass verbatim to keep quoting in their bodies intact.
-        tokens = (
-            join_flag_args(split_flag_entry(flag, "esphome"), "esphome")
-            if " " in flag
-            else (flag,)
-        )
+        # Shell-lex every entry the way PlatformIO's ParseFlags does, so a
+        # knob in "-DKNOB -DOTHER", a spaced "-D KNOB", and quoted bodies all
+        # read identically to _project_flags (and the compile line).
+        tokens = join_flag_args(split_flag_entry(flag, "esphome"), "esphome")
         for tok in tokens:
             if tok.startswith("-D") and len(tok) > 2:
                 body = tok[2:]
@@ -225,6 +223,15 @@ def _resolve_build_config(defines: dict[str, str]) -> _BuildConfig:
                 body for name, body in defines.items() if name.startswith("MMU_")
             )
         else:
+            if "MMU_IRAM_SIZE" in defines or "MMU_ICACHE_SIZE" in defines:
+                # Same diagnostic the PlatformIO builder prints: without the
+                # knob the linker script keeps the default layout while the
+                # compile line carries the custom sizes
+                _LOGGER.warning(
+                    "Detected custom MMU flags; use "
+                    "-DPIO_FRAMEWORK_ARDUINO_MMU_CUSTOM to disable the "
+                    "default configuration"
+                )
             mmu = list(_MMU_DEFAULT)
 
     return _BuildConfig(

tests/unit_tests/build_gen/test_arduino8266.py

@@ -422,3 +422,20 @@ def test_flag_defines_joins_spaced_define() -> None:
     defines = _flag_defines()
     assert "PIO_FRAMEWORK_ARDUINO_LWIP2_HIGHER_BANDWIDTH_LOW_FLASH" in defines
     assert "" not in defines
+
+
+def test_build_config_custom_mmu_without_knob_warns(
+    caplog: pytest.LogCaptureFixture,
+) -> None:
+    """Custom MMU sizes without the CUSTOM knob keep the default layout and
+    warn, as the PlatformIO builder does."""
+    _set_flags("-DMMU_IRAM_SIZE=0xC000")
+    config = _resolve_build_config(_flag_defines())
+    assert config.mmu_defines == ["MMU_IRAM_SIZE=0x8000", "MMU_ICACHE_SIZE=0x8000"]
+    assert "Detected custom MMU flags" in caplog.text
+
+
+def test_flag_defines_lexes_quoted_single_tokens() -> None:
+    """A quoted single-token define reads the same as on the compile line."""
+    _set_flags('-DMMU_SEC_HEAP="0x40108000"')
+    assert _flag_defines()["MMU_SEC_HEAP"] == "MMU_SEC_HEAP=0x40108000"