ESP8266 Community Forum

- Mon Apr 25, 2016 12:22 pm #46176 martinayotte
Thank you for responding and caring.
I'm a little more convinced that there is an issue with the core. It also appears the issue accumulates ( several packets) then it blows up.
Something akin to a memory leak.
I'm relatively new to the Arduino IDE for esp8266 having used for decades PIC MCU's with a realtime debugger ( step a line at a time until the crash throws you off the cliff) but Arduino is lexigraphic and seems to rely on the external file date ( no internal dating within the code itself and no change log) .
Lazarus for example has a good change log and a good error reporting log. Maybe Arduino has it too. I was unaware of the stack trace feature so it is probable Arduino has a good change and bug reporting log and I just don't know about it.

Another fact is that the error is always the same it only varies as to how many packets get displayed before the error occurs.

- Mon Apr 25, 2016 12:27 pm #46178 This bug is really strange ... The line 43 of core_esp8266_main.cpp doesn't give any clues, the stack is probably corrupted somehow ...
Maybe it is only related with UDP, which I never use, because I have no problem with the 2.2.0 TCP connections.

- Mon Apr 25, 2016 2:51 pm #46192 I agree I haven't had any issues with TCP. MQTT subscribe and publish works fine etc.
The UDP broadcast is fine I can see the packets with wireshark and they are correctly formed.
It is the UDP receive that blows up after a short while.....I have only tested with multicast packets so far.

- Mon Apr 25, 2016 3:34 pm #46195 The code works OK using IDE 1.6.8, board package 2.2.0, board type NodeMCU 1.0 and a simple netcat sender test.

Send test data using netcat from Linux. Should work on Mac as well.

Code: Select all$ nc -u 239.0.0.57 5001
;lkjasdf;lkjas;dkflja;skldj;lj
;lkajsdf;lkjasd;flkjas;dfkjasd;lfkas;dlkjfa;lksfdj;aksjdf
;alskdf;laksjdf;laksjdf;klajsfd;aksjdf;klasjdf;lkajsdf;lkajsdf;lkjasdf;lkjas;dflkja;sklfdj;alksdjf;lkasjf;lkajsdf;klajsdfa;ksjd
;kajsdfa;lskdjf


Serial monitor output.

Code: Select all821:Packet of 31 received from 192.168.1.11:33185
;lkjasdf;lkjas;dkflja;skldj;lj

824:Packet of 58 received from 192.168.1.11:33185
;lkajsdf;lkjasd;flkjas;dfkjasd;lfkas;dlkjfa;lksfdj;aksjdf

830:Packet of 7 received from 192.168.1.11:33185
a;ksjd

833:Packet of 16 received from 192.168.1.11:33185
;kajsdfa;lskdjf



Note the code skipped over the packet longer than 64 bytes. The code does not read packets longer than 64 bytes which I thought might jam up the UDP stack. I think it would be safer to read all packets even thought this simple test appears to work.