Welp that doesn’t look good. According to the reverse-engineered memory map, there’s nothing there. Internal mask-ROM code is at 40000000h of length 10000h, instruction RAM at 40100000h, then SPI flash is at 40200000h, variable length. This is the only places where it should ever execute code. 0x00010100 seems to to me that either the current instruction pointer hasn’t been read out correctly or some other weird fault is occuring.
You mean ~/.platformio/packages/toolchain-xtensa/bin/xtensa-lx106-elf-gdb not gdb right?
Exactly, I got the same reference table from another post in the forum. It was exactly my thought, something weird is going on; but what ?
I was using the ~/.arduino15/packages/esp8266/tools/xtensa-lx106-elf-gcc/2.5.0-4-b40a506/bin/xtensa-lx106-elf-gdb;
but with ~/.platformio/packages/toolchain-xtensa/bin/xtensa-lx106-elf-gdb the results is the same - just checked.
What is the most minimal sketch with which it doesn’t anymore? The gdbstub also accesses Serial so one has to be careful. I find it very interesting that the sketch can do a Serial.printf() at all, with other gdbstubs like avr-stub that’s not possible. At the very least, no other library should read from the serial, I guess.
i already had started to minimize my sketch to a point where i found the issue - created an issue on ArduinoJson that as it seems was my fault
regarding the Serial out, indeed i was thinking the same, but the original sketch contained it, so i gave a try - interestingly it works for now for simple outputs