Closed ntavish closed 5 years ago
CC @rlubos
The various prj*.conf files look quite out-dated and probably need to be cleaned up a bit.
The new standard is to design a nice generic prj.conf file which gets used by default and then overlay-*.conf for various specialized cases.
@ntavish: Thanks for submitting this ticket, was looking to create one myself ;-).
So, @rlubos, the problem here is that this sample should not have stuff like #if defined(CONFIG_NET_L2_BT)
. We have net_config subsys to handle networking setup.
Also from @mike-scott on IRC:
The multiple prj_*.conf files threw me off with all of their net_buf/net_pkt settings
TBH, most of these settings could be cleaned up
The fact is BLE stuff in the sample might be outdated, I did not touch it at all during the MQTT update. Will look into this.
The fact is BLE stuff in the sample might be outdated,
It is outdated, as mentioned, stuff like that was replaced long ago by net_app_config which was replaced by net_config. It seems to have some display hacks, like rotating progress indicator waiting for BLE conn, so will be hard to remove ;-).
Take a look at this #11765, it doesn't handle connecting properly is BT is not connected, but it does connect if you are fast enough in connecting after boot.
Closing as this got fixed by #11765.
Describe the bug MQTT publisher sample is stuck trying to connect over TCP.
Following is after making a change to fix a bug(?) which enables bt twice and quits after
EALREADY
is returned:To Reproduce Steps to reproduce the behavior:
Output:
(before earlier fix, it would simply exit
network_setup()
and not attempt to connect)Expected behavior
Application should attempt to connect to MQTT server. I do not see any connection attempts. Through the shell, I can actually connect to mosquitto:
Impact Unable to test a very common use case for BLE/6lowpan, and also things to test after this like TLS support for MQTT.
Environment (please complete the following information):