rework sensor instantiation to saves memory by removing the static allocation

[Links2004]

rework I2C sensor init.

The goal is to only instantiate sensors that are pressend to save memory.

Side effacts:

• easier sensor integration (less C&P code, sensor only needs to be "registerd" at one place)
• nodeTelemetrySensorsMap can be removed when all devices are migrated to save more RAM

🤝 Attestations

• I have tested that my proposed changes behave as described.
• I have tested that my proposed changes do not cause any obvious regressions on the following devices:
  • Heltec (Lora32) V2.1
  • Heltec (Lora32) V3
  • LilyGo T-Deck
  • LilyGo T-Beam
  • RAK WisBlock 4631
  • Seeed Studio T-1000E tracker card
  • Other (please specify below)

[Copilot]

Pull Request Overview

This pull request replaces static allocation of the bsecState buffer with dynamic allocation using malloc() to save 238 bytes of static memory. The change moves the buffer allocation from a class member to local variables within the methods that use it.

Key Changes

• Removed static bsecState array from BME680Sensor class member variables
• Added dynamic allocation with malloc() in loadState() and updateState() methods
• Added proper error handling and memory cleanup with free()

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
BME680Sensor.h	Removes static bsecState array declaration from class members
BME680Sensor.cpp	Implements dynamic allocation with malloc/free in loadState() and updateState() methods

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[mverch67]

A much larger portion of static allocation could be saved if all the sensor instances are declared as pointer and allocated in heap space during init. Then no changes in this BSEC sensor class are required, e.g. see here:

firmware/src/modules/Telemetry/EnvironmentTelemetry.cpp

Line 72 in b49496d

BME680Sensor bme680Sensor;

[Links2004]

true, had a look at EnvironmentTelemetry.cpp and it looks like there is a lot more possibility for optimizion.
In my opinion there is a loot to gain there, the current desigen looks very inefficient over all.

looked a bit deeper and I think the I2C device handling overall needs some work.

[Links2004]

have working code that elimenates the need for nodeTelemetrySensorsMap and only creates the sensor instances when the I2C scan found a device.

moved over 4 sensors to show case the idea, but want some feedback before doing all of them.
moving this 4 sensors over saved already 1152 bytes of static allocations.

I dont see that we get memory fragmentation problem with this code, since the only cases where we delete / free something is when the sensor failed it initialisation.

tested this with a BMP-280 connected to a Heltec V2.1

please review and give feedback if this code is something that can be merged (will migrate the other sensors then).

[Links2004]

did a second batch of sensors

RAM -816
Flash -916

[thebentern]

Tagging @oscgonfer for visibility

[Links2004]

The PR now covers all sensors that where allocated staticly in EnvironmentTelemetry.

Total saved:

RAM -2160
Flash -4332

we will save CPU cycles too since we do a lot less hasSensor calls.

some sensors like the INA ones are part of PowerTelemetry and EnvironmentTelemetry not sure why but I did not touch them.

Ready for review.