Commit Graph
15 Commits
Author SHA1 Message Date
J. Nick Koston c4e1360cdf [api] Check the encoded end against the reserved size under ESPHOME_DEBUG_API
The fixed32 store helper moves to a private section since it neither bounds checks nor
advances the cursor, its comment describes the path each target takes, the generated
file scan flags any ProtoEncode call that does not assign the cursor, and StateWaiter
timeouts can carry a label so gathered waits are told apart.
2026-09-07 15:57:10 +02:00
J. Nick Koston 709a1e1eb6 [api] Mark the raw encode helpers nodiscard too and drop a duplicate cursor test
The generated file scan already covers every emitted call, so the parametrized copy of
the same assertion goes.
2026-09-07 13:48:53 +02:00
J. Nick Koston 822b701792 [api] Mark the cursor returning encode helpers nodiscard
A call that drops the returned cursor would silently truncate the message, so the
compiler now warns on it and a unit test scans the generated file for the same mistake.
Also corrects the outlining comment for ESP8266, where the inline write is a few byte
stores rather than one, and the RAW_ENCODE_MAP annotation.
2026-09-07 12:37:48 +02:00
J. Nick Koston adbbda4072 [api] Emit every encode call through one generator helper
_encode_call() owns the cursor assignment and the _force suffix, so
the convention lives in one place instead of at every emission site;
the fixed32 fast path is an arm of the generic encode_content keyed by
a per type value template. write_fixed32_le uses convert_little_endian
instead of its own byte order switch. The integration test shares a
StateWaiter from state_utils and leaves the disconnect to the fixture.
2026-09-07 11:09:07 +02:00
J. Nick Koston 490aca17e6 [api] Share the fixed32 emission between float and fixed32 fields
One helper next to the other precomputed tag paths decides how a
single byte tag fixed32 field is written; the float and fixed32 types
only differ in the value expression. Drop the non forced std::string
encode_string overload, which the generator never emits, and build the
generator tests from one block of field type constants.
2026-09-07 10:33:21 +02:00
J. Nick Koston 3b14f4dfc8 [api] Pass the encode cursor by value through the protobuf helpers
The ProtoEncode helpers took the write cursor by reference and a
bool force flag. At -Os the compiler outlines most of them, so every
call site had to keep pos in a stack slot and pass its address, plus
a constant for the flag. The helpers now take the cursor by value and
return the advanced cursor, so consecutive calls chain through the
return register; forced fields call a _force overload instead of
passing a flag.

The fixed32 writers use __builtin_memcpy, which stays a builtin under
ESP-IDF's -fno-builtin-memcpy, and are outlined on embedded targets so
each fixed32 or float field is a short call instead of an inline
memcpy call. Non-forced float and fixed32 fields with a single-byte
tag share the same writer behind a zero check.

Generated encode bodies shrink by 18 percent on an ESP32 IDF proxy
build (2360 to 1932 bytes for 27 messages); entity messages gain the
most, for example ListEntitiesSensorResponse::encode 190 to 134 bytes
and SensorStateResponse::encode 78 to 49 bytes.
2026-09-07 09:21:54 +02:00
J. Nick Koston 145373ed13 [api] Widen message type storage to uint16_t (#18526) 2026-08-25 18:04:41 +00:00
Keith BurzinskiandJ. Nick Koston b97c877992 [api] Acknowledge proxy subscribe and configuration requests (#18312)
Co-authored-by: J. Nick Koston <nick@koston.org>
2026-08-24 20:27:24 +00:00
J. Nick Koston 37eae9b466 [api] Gate the bluetooth connection messages on their own define (#18281) 2026-08-11 21:31:13 -05:00
Jesse Hills f697cf2011 [api] Add DeviceCapabilities message for optional-feature flags (#17984) 2026-08-11 14:52:19 +12:00
J. Nick Koston 4424af8c5b [api] Move runtime log client out of the component package (#18043) 2026-08-03 19:57:52 -05:00
J. Nick Koston 2a59c0ad9e [logs] Cap the logs reconnect backoff for deep-sleep devices (#17656) 2026-07-17 17:16:14 -10:00
J. Nick Koston 5b4ae22f58 [api] Don't spam tracebacks when decoding a crash without a local build (#17597) 2026-07-16 08:08:28 -04:00
Diorcet Yann 690a197346 [main] Move stacktrace handling out of platformio_api and FlashImage into platform components/util (#16186) 2026-05-04 15:07:31 -04:00
J. Nick Koston 9ddb828da3 [api] Don't tear down log connection on stack-trace decode failure (#16196) 2026-05-03 20:05:13 -05:00