Skip to content

Commit

Permalink
Indicate that new developments have been made
Browse files Browse the repository at this point in the history
  • Loading branch information
obones authored and cpainchaud committed May 31, 2023
1 parent 1c288a1 commit ba83027
Showing 1 changed file with 5 additions and 13 deletions.
18 changes: 5 additions & 13 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -19,27 +19,19 @@ This branch allows the inclusion of [rtl_433](https://github.com/merbanan/rtl_43
* list.c
* pulse_demod.c
* r_api.c
1. Edit `src\data.c` to add the following lines right before the first `typedef void`
```cpp
#ifdef ESP32
#define _POSIX_HOST_NAME_MAX 16
#endif
```
* r_util.c
1. Apply the changes from this [PR](https://github.com/merbanan/rtl_433/pull/1728)
1. Edit `src\devices\secplus_v2.c` at the end of the file to change the value for `disabled` to `1` as this plugin uses too much stack
1. Make sure the main task is configured to use a 24kb stack
1. Open `C:\Users\YOURLOGIN\.platformio\packages\framework-arduinoespressif32\cores\esp32\main.cpp`
1. On the call to `xTaskCreateUniversal` near the end, replace `CONFIG_ARDUINO_LOOP_STACK_SIZE` by `24576`

Note that this local edit will not be required once the following [PR](https://github.com/espressif/arduino-esp32/pull/5173) has been released as it will allow us to simply change the value from the `platformio.ini` file
Note that this local edit will not be required once the following [PR](https://github.com/espressif/arduino-esp32/pull/5173) has been released as it will allow us to simply change the value from the `platformio.ini` file

The edits in `secplus_v2.c` and `main.cpp` can be avoided if you apply the changes suggested in this [issue](https://github.com/merbanan/rtl_433/issues/1726)

**What's next / Questions**

1. Create a PR at rtl_433 for the change in `src\data.c`
Is the value 16 coherent? Not that it would matter much, it is used in the `data_output_syslog_t` structure that I believe can't be used anyway on ESP32 (no syslog?)
1. Create an issue at rtl_433 for discussing the possibility of making `bitbuffer_t` an opaque pointer, thus forcing to allocate on heap which will save stack space. This structure is used a lot and is the main culprit for stack consumption.
This has to be tested on the real rtl_433 first: I have already started investigating as I have no problem building under VisualStudio
3. Modify `data_acquired_handler` inside the bridge so that it converts the `data_t` element it receives to items that are valid for the various `display_XX` methods
This would avoid the conversion to json and allow calling `RFLink::sendMsgFromBuffer()` that takes care of sending to Ser2Net, MQTT and the serial port as if it was a more regular plugin.
1. Should `processReceivedData()` return the number of decoded messages?
This could then be used to break out of `ScanEvent` early but it should return `false` to prevent the caller of `ScanEvent` from sending the same message as already sent by `data_acquired_handler`

Expand Down

0 comments on commit ba83027

Please # to comment.