ESP-32 plans for ESPEasy (?)

Message

Martinus · #21 Post by **Martinus** » 04 Mar 2018, 17:49

VoBo wrote: ↑02 Mar 2018, 18:06 Hi everybody

I´m playing around with R20100 on my ESP32 board.
Installing was easy and everything what I need, is already working
"One" of the advantages of the ESP32 ist the increased amount of GPIO Ports; but the amount of devices in R20100 ist only 12.
Is it possible to increase also the amounts of devices ? I think that 24 or 32 should be enough. As far as I know, there are only 12 device because of the limits at an 8266.
The ESP32 should have enough ressources for more devices ?
Is it possible to change this ? And, if yes, can somebody compile a version with more devices ? (for me, that is too deep in arduino)

greetings from germany
Volker

It's done. Beware that you will experience a full reset because this version has a different memory layout for the config.dat file.

Now we have a theoretical max of 32 tasks to be configured. It doesn't imply that performance will be sufficient to actually run 32 tasks as this has never been tried before. I can imagine that when you run a large number of tasks that use the 10PERSECOND polling scheme, the system could slow down to a hog...

But hey, this route is highly experimental anyway...

soif · #22 Post by **soif** » 05 Mar 2018, 09:28

Happy to see that the ESP32 dev is strongly on its way : Thanks for that

My plan was also to add support for writing all stored structs to JSON and read them again from JSON.
That will make it possible to add more of a REST interface which will make the project more scalable to use. And also upgrading settings to a newer version and easier sharing of settings

That would be really GREAT !

BTW would you please consider to split the device configuration + hardware config from the whole config file:
viewtopic.php?f=6&t=4965

Perhaps even a central 'backup' of settings and deploy of settings to multiple nodes.

My EspBuddy (see signature) already do a whole backup in one command

, but splitting files as explained in my linked post above, would pave the way for great centralized features, ie modify global settings on all nodes at once, or duplicate a node into another in a matter of seconds

Also the web interface could then be separated from the project. Using a single (e.g. REST-like) interface for interaction will also lead to fewer bugs (single path of interface) and perhaps a smaller codebase.
And generating the web page in C++ like how it is done now, uses a lot of heap resources and is very hard to maintain.

+1000, we might then integrate bootstrap, to get beautiful, mobile friendly web UI

VoBo · #23 Post by **VoBo** » 05 Mar 2018, 23:00

Hi,

Martinus wrote: ↑04 Mar 2018, 17:49 It's done. Beware that you will experience a full reset because this version has a different memory layout for the config.dat file.

Now we have a theoretical max of 32 tasks to be configured. It doesn't imply that performance will be sufficient to actually run 32 tasks as this has never been tried before. I can imagine that when you run a large number of tasks that use the 10PERSECOND polling scheme, the system could slow down to a hog...

But hey, this route is highly experimental anyway...

thank you very much

But the 32 Tasks seem to be a too much

I´ve connected 8 Input switches and 8 Relais, which are switched by rules (like "sonoff´s")
If I press the first switch, the relay switches immediately (because it´s the first one in the "rules")
if I press switch Nr.8, which is the last one in the rules, it takes nearly 2 seconds until the relay ist activated

... also the "wroom" get´s quiet hot

greetings,
Volker

Martinus · #24 Post by **Martinus** » 12 Mar 2018, 20:21

VoBo wrote: ↑05 Mar 2018, 23:00 Hi,

Martinus wrote: ↑04 Mar 2018, 17:49 It's done. Beware that you will experience a full reset because this version has a different memory layout for the config.dat file.

Now we have a theoretical max of 32 tasks to be configured. It doesn't imply that performance will be sufficient to actually run 32 tasks as this has never been tried before. I can imagine that when you run a large number of tasks that use the 10PERSECOND polling scheme, the system could slow down to a hog...

But hey, this route is highly experimental anyway...
thank you very much
But the 32 Tasks seem to be a too much
I´ve connected 8 Input switches and 8 Relais, which are switched by rules (like "sonoff´s")
If I press the first switch, the relay switches immediately (because it´s the first one in the "rules")
if I press switch Nr.8, which is the last one in the rules, it takes nearly 2 seconds until the relay ist activated
... also the "wroom" get´s quiet hot

greetings,
Volker

Could you check the processing time on the rules? To see if this matched the 2 seconds delay.

Martinus · #25 Post by **Martinus** » 12 Mar 2018, 20:23

ESP32 binary has been updated to the latest github dev version (2018-03-12). Using the latest libraries.
https://www.letscontrolit.com/wiki/index.php/ESPEasy32
C010 is fixed.
core_version.h no longer needed for ESP32.

VoBo · #26 Post by **VoBo** » 17 Mar 2018, 21:22

Martinus wrote: ↑12 Mar 2018, 20:21 Could you check the processing time on the rules? To see if this matched the 2 seconds delay.

how can I do that ?

Martinus · #27 Post by **Martinus** » 30 Mar 2018, 16:38

ESP32 binary has been updated to the latest github dev version (2018-03-30)
https://www.letscontrolit.com/wiki/index.php/ESPEasy32

budman1758 · #28 Post by **budman1758** » 30 Mar 2018, 20:33

Martinus wrote: ↑30 Mar 2018, 16:38 ESP32 binary has been updated to the latest github dev version (2018-03-30)
https://www.letscontrolit.com/wiki/index.php/ESPEasy32

Thanks for keeping this updated. I really hope to see this continue to be developed. I think ESPEasy is a fantastic firmware for the ESP32.
I am working on a fairly complex project for it and so far the only real bug is the OLED issue reported here.
https://github.com/letscontrolit/ESPEasy/issues/1029

TD-er · #29 Post by **TD-er** » 31 Mar 2018, 01:53

It started with some very strange build error after a very normal and sound commit of Martinus.
It took me almost all evening and now it is fixed... and there is a build entry for ESP32 firmware in the PlatformIO.ini file.

Now there are 2 kinds of developers; those who start building and compiling and those who read the commit first

Martinus · #30 Post by **Martinus** » 31 Mar 2018, 14:49

TD-er wrote: ↑31 Mar 2018, 01:53 It started with some very strange build error after a very normal and sound commit of Martinus.
It took me almost all evening and now it is fixed... and there is a build entry for ESP32 firmware in the PlatformIO.ini file.

Now there are 2 kinds of developers; those who start building and compiling and those who read the commit first

There are also "developers" with poor skills and that only use Arduino IDE

Maybe i should really consider to retire from the project entirely.

TD-er · #31 Post by **TD-er** » 31 Mar 2018, 15:30

Martinus wrote: ↑31 Mar 2018, 14:49
TD-er wrote: ↑31 Mar 2018, 01:53 It started with some very strange build error after a very normal and sound commit of Martinus.
It took me almost all evening and now it is fixed... and there is a build entry for ESP32 firmware in the PlatformIO.ini file.

Now there are 2 kinds of developers; those who start building and compiling and those who read the commit first
There are also "developers" with poor skills and that only use Arduino IDE
Maybe i should really consider to retire from the project entirely.

Please don't stop working on the project.
My remark was just a wink at the comment I placed near the PlatformIO entry.

Code: Select all

### ESP32 pre-alpha test build ###########################################################
# Very limited build for ESP32, to start testing regular building for ESP32.             #
# Will probably not work, not tested and guaranteed to take a few hours time of some     #
# still trying to build the version without reading this warning.                        #
##########################################################################################
[env:esp32dev]

And it is just to make sure I also see the build errors I may cause when changing stuff in ESP8266-land.
It is just that we have had a few 'build' issues last few weeks where I was not able to see what's going on and was thus unable to help fixing stuff in ESP32 versions.
Let's hope I did not change stuff to cause build errors again in Arduino IDE using ESP32, because I did not test that yet.
At the moment we have about 4 build environments:
- PlatformIO with ESP8266 and ESP32
- Arduino IDE with ESP8266 and ESP32.

And as long as we don't test all of them regularly, there is a guarantee there will be new broken builds on one of them.

So my attempt to incorporate the ESP32 build on PlatformIO is just to make sure less build errors will occur.

I was hoping you could help with the SPIFF settings needed for the ESP32.

Martinus · #32 Post by **Martinus** » 31 Mar 2018, 15:46

TD-er wrote: ↑31 Mar 2018, 15:30 I was hoping you could help with the SPIFF settings needed for the ESP32.

We don't select this anymore within the Arduino IDE. The ESP32 seems to use some sort of partition table that seems to default to:

Code: Select all

# Name,   Type, SubType, Offset,  Size, Flags
nvs,      data, nvs,     0x9000,  0x5000,
otadata,  data, ota,     0xe000,  0x2000,
app0,     app,  ota_0,   0x10000, 0x140000,
app1,     app,  ota_1,   0x150000,0x140000,
eeprom,   data, 0x99,    0x290000,0x1000,
spiffs,   data, spiffs,  0x291000,0x16F000,

file: \hardware\espressif\esp32\tools\partitions\defaults.csv

Martinus · #33 Post by **Martinus** » 31 Mar 2018, 15:57

budman1758 wrote: ↑30 Mar 2018, 20:33 Thanks for keeping this updated. I really hope to see this continue to be developed. I think ESPEasy is a fantastic firmware for the ESP32.
I am working on a fairly complex project for it and so far the only real bug is the OLED issue reported here.
https://github.com/letscontrolit/ESPEasy/issues/1029

I think we need another can of devs to get the ESP32 development on the road, because there's a lot of work to be done here.
I merely tried to get the code through the compiler and even that can be a struggle sometimes. (the Arduino core developer teams on ESP8266 and ESP32 are not really in 'sync' at the moment)

I can't spend as much time as back in the 'old days' and my testlab has also been reduced:

: Testlabs.png (295.07 KiB) Viewed 37178 times

Just enough to verify some core features of ESPEasy32.

LisaM · #34 Post by **LisaM** » 31 Mar 2018, 17:33

TD-er wrote: ↑31 Mar 2018, 15:30 At the moment we have about 4 build environments:
- PlatformIO with ESP8266 and ESP32
- Arduino IDE with ESP8266 and ESP32.

Four? Pfff, i only have one...

Even the one is to much, i hope to make it zero. Using standard micropython firmware on a ESP32-WROVER with 4MB PsRam, i can just ftp my source code over and run that. With 3MB of flash drive there's plenty of room left, so there's no need for optimization anymore. To me it would be ideal to be just an app on micropython and let the other micropython groups deal with building the firmware, that would also expand the number of people working on uPyEasy significantly (by outsourcing a lot of firmware build&burn issues).

TD-er · #35 Post by **TD-er** » 31 Mar 2018, 19:26

LisaM wrote: ↑31 Mar 2018, 17:33
TD-er wrote: ↑31 Mar 2018, 15:30 At the moment we have about 4 build environments:
- PlatformIO with ESP8266 and ESP32
- Arduino IDE with ESP8266 and ESP32.
Four? Pfff, i only have one...
Even the one is to much, i hope to make it zero. Using standard micropython firmware on a ESP32-WROVER with 4MB PsRam, i can just ftp my source code over and run that. With 3MB of flash drive there's plenty of room left, so there's no need for optimization anymore. To me it would be ideal to be just an app on micropython and let the other micropython groups deal with building the firmware, that would also expand the number of people working on uPyEasy significantly (by outsourcing a lot of firmware build&burn issues).

I agree the Python way of doing things is the way to the future.
But until then, it is nice to see what can be done with really limited resources.
And I don't know how well you can assign tasks in Python to the 2nd core?
I think that's a great new feature, to allow to run some tasks really realtime and all other tasks on the other core.
Quite often you see feature requests here that need some realtime aspect, which the current ESPeasy can't do.

LisaM · #36 Post by **LisaM** » 31 Mar 2018, 21:10

TD-er wrote: ↑31 Mar 2018, 19:26
LisaM wrote: ↑31 Mar 2018, 17:33
TD-er wrote: ↑31 Mar 2018, 15:30 At the moment we have about 4 build environments:
- PlatformIO with ESP8266 and ESP32
- Arduino IDE with ESP8266 and ESP32.
Four? Pfff, i only have one...
Even the one is to much, i hope to make it zero. Using standard micropython firmware on a ESP32-WROVER with 4MB PsRam, i can just ftp my source code over and run that. With 3MB of flash drive there's plenty of room left, so there's no need for optimization anymore. To me it would be ideal to be just an app on micropython and let the other micropython groups deal with building the firmware, that would also expand the number of people working on uPyEasy significantly (by outsourcing a lot of firmware build&burn issues).
I agree the Python way of doing things is the way to the future.
But until then, it is nice to see what can be done with really limited resources.
And I don't know how well you can assign tasks in Python to the 2nd core?
I think that's a great new feature, to allow to run some tasks really realtime and all other tasks on the other core.
Quite often you see feature requests here that need some realtime aspect, which the current ESPeasy can't do.

MicroPython task is running on the ESP32 App core, everything else on the first core. The current uPyEasy for ESP32 firmware is however using a single core approach only (so far).
Realtime requires good hardware interrupts and that's problem with the ESP32 (and ESP8266). The ESP's have software interrupts, while the STM32's have hardware interrupts. The result is that the GPIO latency with STM32's is only 1/10 of the ESP's. The ESP's will never be a good realtime platform, for that you'll need to go the STM32's (like anybody else using realtime).

TD-er · #37 Post by **TD-er** » 01 Apr 2018, 00:44

OK, that's good to know about the hardware ints, or actually the lack of.

About the ESPeasy + ESP32 combo.
I just tested an ESP32 module with the PlatformIO build I made yesterday and it just works

So expect some more debugging on that platform too. (Already found the first bug...)

Martinus · #38 Post by **Martinus** » 01 Apr 2018, 13:40

TD-er wrote: ↑31 Mar 2018, 19:26 I agree the Python way of doing things is the way to the future.
But until then, it is nice to see what can be done with really limited resources.
And I don't know how well you can assign tasks in Python to the 2nd core?
I think that's a great new feature, to allow to run some tasks really realtime and all other tasks on the other core.
Quite often you see feature requests here that need some realtime aspect, which the current ESPeasy can't do.

We actually have two different discussions here:

1) Latency on ISR handling
==========================
AFAIK, the ESP8266 and ESP32 do have hardware interrupts, but for proper handling and low latency, it is required that the response handling fits entirely in iRAM. I guess that this requirement prohibits higher level languages like microPython from handling this properly.
As explained at the end of this topic
https://forum.micropython.org/viewtopic ... 4&start=30
But using lean and mean C-language ISR handlers within ESPEasy could still do some nice jobs.

2) Realtime operating system
=========================
I think this is really the issue that Gijs is talking about. ESPEasy could really benefit from having an RTOS that will give the impression of a real multitasking system. As the SDK behind the ESP32 Arduino core is based on freeRTOS, i think that it would be possible to schedule more RTOS tasks within ESPEasy running on the ESP32.

Right now, the main.cpp only schedules a single task to run the entire user code like this:

Code: Select all

xTaskCreatePinnedToCore(loopTask, "loopTask", 8192, NULL, 1, NULL, ARDUINO_RUNNING_CORE);

Maybe it's possible to schedule more tasks for specific tasks/plugins?

Martinus · #39 Post by **Martinus** » 01 Apr 2018, 13:43

TD-er wrote: ↑01 Apr 2018, 00:44 OK, that's good to know about the hardware ints, or actually the lack of.

About the ESPeasy + ESP32 combo.
I just tested an ESP32 module with the PlatformIO build I made yesterday and it just works
So expect some more debugging on that platform too. (Already found the first bug...)

Good news! Does that mean that we will have nightly builds also for ESP32?

chrille · #40 Post by **chrille** » 01 Apr 2018, 20:11

TD-er wrote: ↑01 Apr 2018, 00:44 About the ESPeasy + ESP32 combo.
I just tested an ESP32 module with the PlatformIO build I made yesterday and it just works
So expect some more debugging on that platform too. (Already found the first bug...)

Did anyone succeed running on other ESP32 boards than the LoLin32 lite? I have two different ESP32 boards (the basic Lolin ESP32 and the Goouuu ESP32). I have no success with any of the two boards. I tried both the version from https://www.letscontrolit.com/wiki/index.php/ESPEasy32 and building the latest version from github with platform-io

Code: Select all

INIT : Booting version: (custom)
INIT : Cold Boot
FS   : Mounting...
CRC  : No program memory checksum found. Check output of crc2.py
CRC  : Settings CRC           ...OK
CRC  : SecuritySettings CRC   ...OK
INIT : Free RAM:168836
INIT : I2C
INIT : SPI not enabled
ets Jun  8 2016 00:22:57

rst:0xc (SW_CPU_RESET),boot:0x13 (SPI_FAST_FLASH_BOOT)
configsip: 0, SPIWP:0xee
clk_drv:0x00,q_drv:0x00,d_drv:0x00,cs0_drv:0x00,hd_drv:0x00,wp_drv:0x00
mode:DIO, clock div:1
load:0x3fff0010,len:4
load:0x3fff0014,len:812
load:0x40078000,len:0
load:0x40078000,len:10164
entry 0x400789f8
U

After the U it just boots again

It seems that the first boot where the initial config is written is OK, but I haven't managed to start a terminal program fast enough after esptool finishes, to see the initial boot message

- Jan

chrille · #41 Post by **chrille** » 01 Apr 2018, 22:55

chrille wrote: ↑01 Apr 2018, 20:11 Did anyone succeed running on other ESP32 boards than the LoLin32 lite? I have two different ESP32 boards (the basic Lolin ESP32 and the Goouuu ESP32). I have no success with any of the two boards. I tried both the version from https://www.letscontrolit.com/wiki/index.php/ESPEasy32 and building the latest version from github with platform-io

Me bad! It turned out to be a bad USB cable. As soon as the wifi stack was initialized the module rebooted. I observed the exact same behavior with micropython on the same module, which got me on track
After replacing the USB cable it just works!

- Jan

TD-er · #42 Post by **TD-er** » 01 Apr 2018, 23:03

Martinus wrote: ↑01 Apr 2018, 13:43
TD-er wrote: ↑01 Apr 2018, 00:44 OK, that's good to know about the hardware ints, or actually the lack of.

About the ESPeasy + ESP32 combo.
I just tested an ESP32 module with the PlatformIO build I made yesterday and it just works
So expect some more debugging on that platform too. (Already found the first bug...)
Good news! Does that mean that we will have nightly builds also for ESP32?

For the nightly builds, we probably need a change made by Edwin.
But is really easy now to build an ESP32 version.
There are some strange things yet with the WiFi and the email notification.
I've been testing a bit this morning, but I've not had much time today.
I did find this discussion which may or may not be related.
Not sure yet how to upgrade the ESP32 core library.

LisaM · #43 Post by **LisaM** » 01 Apr 2018, 23:27

TD-er wrote: ↑01 Apr 2018, 00:44 OK, that's good to know about the hardware ints, or actually the lack of.

About the ESPeasy + ESP32 combo.
I just tested an ESP32 module with the PlatformIO build I made yesterday and it just works
So expect some more debugging on that platform too. (Already found the first bug...)

From the ESP32 board:

I have done a measurement and delay from external trigger to application-provided ISR handler is around 2us (at 240MHz clock), which is around 500 cycles.

https://www.esp32.com/viewtopic.php?t=422
So it takes the ESP32 2us just to start the interrupt handler...
A technical explanation what interrupt latency is: https://community.arm.com/processors/b/ ... processors
And here is somebody complaining it takes a very long time for a STM32F4 ISR to react to an external interrupt: 200ns

Complaining because it should be possible to have a STM32F4 react within 20ns...

I've heard that the ESP32 is so slow because the ESP32 has a hidden software interrupt handler inside the ESP32, micro code, which explains why it takes so long to even start the ISR...

Conclusion: the ESP32 is a nice SOC, but not for realtime purposes... (don't use it in your nuclear power plant!

)

LisaM · #44 Post by **LisaM** » 02 Apr 2018, 00:43

Martinus wrote: ↑01 Apr 2018, 13:40 I guess that this requirement prohibits higher level languages like microPython from handling this properly.
As explained at the end of this topic
https://forum.micropython.org/viewtopic ... 4&start=30

The reported maximum latency is 380us, that leaves 19.620us for the micropython ISR to complete the ISR (considering the max of 50 measurements per second).
Considering that only 5 plugins are using the max of PLUGIN_FIFTY_PER_SECOND, i'm not to worried.

LisaM · #45 Post by **LisaM** » 03 Apr 2018, 18:52

If your still worried that micropython is to slow, here's a discussion about the speed: https://forum.micropython.org/viewtopic.php?f=2&t=1349
To be short, the speed at toggling pin's:
Viper: 18,618,506 toggles/s, 9.31 MHz, 18 Clock Cycles/iter
ASM: 27,929,060 toggles/s, 14 MHz, 12 Clock Cycles/iter
Where Viper is micropython in Viper mode (very limited set of commands) and ASM is micropython embedded assembly. There's also Native mode which is faster, but more limited, then normal micropython mode. These numbers are all for an STM32F405. So far i've spent practically zero time on optimizing the code.
In case you really need speed (which i think is rare), they are possibilities.

waspie · #46 Post by **waspie** » 05 Apr 2018, 13:54

do i see in github that the esp32 build is now running both cores?

i don't know how to read this stuff that well but i got the impression from some commits

happytm · #47 Post by **happytm** » 13 Apr 2018, 21:54

Dear Martinus,

Does it compile with following plugins ? :

_P059_Encoder
_P062_MPR121_KeyPad
_P017_PN532

I know you mentioned a while ago IRTX and IRRX still do not compile but have you made any progress after that?

I tried and failed to compile so asking for your help.

Thanks

hamster · #48 Post by **hamster** » 14 Apr 2018, 19:12

Latest esp32 version http://www.letscontrolit.com/downloads/ ... R20100.zipseems very good and quite stable.
I was wondering how to compile a version myself as I want to add a sensor [Nextion display] from the playground
I tried and managed to create one in PlatformIO but the result only had a few devices available and not the nextion.
Any assistance much appreciated.

TD-er · #49 Post by **TD-er** » 14 Apr 2018, 20:34

hamster wrote: ↑14 Apr 2018, 19:12 Latest esp32 version http://www.letscontrolit.com/downloads/ ... R20100.zipseems very good and quite stable.
I was wondering how to compile a version myself as I want to add a sensor [Nextion display] from the playground
I tried and managed to create one in PlatformIO but the result only had a few devices available and not the nextion.
Any assistance much appreciated.

The Nextion plugin is on its way, so have a little patience.

grovkillen · #50 Post by **grovkillen** » 15 Apr 2018, 00:10

TD-er wrote: ↑14 Apr 2018, 20:34 The Nextion plugin is on its way, so have a little patience.

Tommmii · #51 Post by **Tommmii** » 15 Apr 2018, 10:44

what's the best place to watch, to know if there's a fresh build ?

Right now i'm watching https://www.letscontrolit.com/wiki/index.php/ESPEasy32

TD-er · #52 Post by **TD-er** » 15 Apr 2018, 12:30

There are plans to include it in the nightly build. Not sure about the status.

hamster · #53 Post by **hamster** » 16 Apr 2018, 19:25

TD-er wrote: ↑14 Apr 2018, 20:34
hamster wrote: ↑14 Apr 2018, 19:12 Latest esp32 version http://www.letscontrolit.com/downloads/ ... R20100.zipseems very good and quite stable.
I was wondering how to compile a version myself as I want to add a sensor [Nextion display] from the playground
I tried and managed to create one in PlatformIO but the result only had a few devices available and not the nextion.
Any assistance much appreciated.
The Nextion plugin is on its way, so have a little patience.

looking forward to it

hamster · #54 Post by **hamster** » 29 Apr 2018, 00:35

any pointers how to compile a image for esp32 ?

fazambuja · #55 Post by **fazambuja** » 29 Apr 2018, 23:56

You can do it with Platformio IDE. It will show up on the list of what can you build. The number of plugins included is really minimal, you need to include the ones you want and test. But first, try to build and upload the firmware

https://www.letscontrolit.com/wiki/inde ... Platformio

https://www.letscontrolit.com/wiki/inde ... platformio

Tommmii · #56 Post by **Tommmii** » 30 Apr 2018, 10:54

@hamster : if you do have a successful build, would you mind sharing the files to flash ?

hamster · #57 Post by **hamster** » 02 May 2018, 23:58

within Platformio I can build esp32dev

However i end up with an image with only a few devices

: esp32.png (14.14 KiB) Viewed 12606 times

So being a novice and not knowing how to select the devices for the build I edited the Framed OLED 36 device added the Nextion content, renaming all references to 75 to 36

but this then fails when compiling with the following errors? Any ideas?

Linking .pioenvs\esp32dev\firmware.elf
.pioenvs\esp32dev\src\ESPEasy.ino.cpp.o:(.literal._Z10Plugin_036hP11EventStructR6String+0x34): undefined reference to `ESPeasySoftwareSerial::ESPeasySoftwareSerial(unsigned char, unsigned char, bool, unsigned short)'
.pioenvs\esp32dev\src\ESPEasy.ino.cpp.o:(.literal._Z10Plugin_036hP11EventStructR6String+0x38): undefined reference to `ESPeasySoftwareSerial::begin(long)'
.pioenvs\esp32dev\src\ESPEasy.ino.cpp.o: In function `StringSumHelper::StringSumHelper(String const&)':
C:\Users\noobed\.platformio\packages\framework-arduinoespressif32\cores\esp32/WString.h:47: undefined reference to `ESPeasySoftwareSerial::ESPeasySoftwareSerial(unsigned char, unsigned char, bool, unsigned short)'
.pioenvs\esp32dev\src\ESPEasy.ino.cpp.o: In function `Plugin_036(unsigned char, EventStruct*, String&)':

C:\Users\noobed\.platformio\packages\framework-arduinoespressif32\cores\esp32/WString.h:47: undefined reference to `ESPeasySoftwareSerial::begin(long)'
collect2.exe: error: ld returned 1 exit status
*** [.pioenvs\esp32dev\firmware.elf] Error 1

TD-er · #58 Post by **TD-er** » 05 May 2018, 00:04

The list of devices included in the ESP32 build is very limited, because the rest does not yet work very well (or not at all) on ESP32.
I just added a few that work and others will follow, but lately there have so many issues with the 82xx builds, that took all my time to fix.

whatsupskip · #59 Post by **whatsupskip** » 13 Jul 2018, 00:35

I couldn't flash my ESP-32, but given the development is still in its early days I won't worry about it.

I am very grateful for all the work by the developers on the ESP8266 version of the ESPEasy.

Who is online