These are chat archives for esp8266/Arduino

Jul 2015
Jul 19 2015 09:00
add Serial.setDebugOutput(true); after Serial.begin, if you get an exception message do a onjdump of the elf file and look at the address of the error
xtensa-lx106-elf-objdump -S your_elf_file.elf > dum.dobj
Jul 19 2015 09:22
so i get:

Fatal exception (28):
epc1=0x402029af, epc2=0x00000000, epc3=0x00000000, excvaddr=0x00000000, depc=0x00000000

Exception (28):
epc1=0x402029af epc2=0x00000000 epc3=0x00000000 excvaddr=0x00000000 depc=0x00000000

ctx: sys
sp: 3ffeb5e0 end: 3fffffb0 offset: 01a0

3ffeb780: 00000000 00000001 3ffeae76 402040d4

is 3ffeb780 what i am looking for
or 0x402029af?
Jul 19 2015 09:53
0x402029af is the address of the error
you try to access invalid memory may a ptr goes wrong in some way
Jul 19 2015 10:07
argh, "No heap available, failed to malloc"
Jul 19 2015 10:20
ok then its a null ptr problem
Jul 19 2015 10:25
Thanks @Links2004. I amtrying to convert some C++ code that was not really designed for the esp
Jul 19 2015 12:57

Hi Guys. Loving the work.

I could be wrong but the portOutputRegister and portModeRegister macros in Arduino.h don't seem to work like they do on AVR.

Shouldn't it be

define portOutputRegister(port) ((volatile uint32_t*) &GPO)

instead of

define portOutputRegister(port) ((volatile uint32_t*) GPO)

either that or

define ESP8266_REG(addr) ((volatile uint32_t )(0x60000000+(addr)))

needs to be

define ESP8266_REG(addr) ((volatile uint32_t *)(0x60000000+(addr)))

in esp8266_peri.h

I'm using the Adafruit_SSD1306 library and all works well with very minor mods except for this change to the core.


Jul 19 2015 18:38
question, is the OTA code supposed to stay like it is now or will there be a solution that allows the sketch to stay untouched but still OTA-enabled?
Me No Dev
Jul 19 2015 21:35
it will be an included service
chad cormier roussel
Jul 19 2015 22:51
i got that error :
it show me that : 40201c01: 42d8 l32i.n a13, a2, 16
// amount of stack taken by interrupt or exception handler
// and everything up to __wrap_system_restart_local
// (determined empirically, might break)