quackmore wrote:hey, this is a cool thread!
I have a question
didn't read the doc yet
pardon me if it will sound goofy ...
is your library intended just for testing or also for diagnostic ?
I mean trying to register what goes wrong and it's so hard to reproduce later ...
Thanks for your reply; I'm glad there is further interest.
To your question: I think it is well suited to al aspects really. From initial problem determination (tracking code execution, logging data), through to final optimisation (profiling execution and memory/resource usage) and then into operation (logging of certain events, data etc). Obviously tracing too much in operation might impact it's normal use so that has to be used in the relevant way according to your project.
Right now, it is possible to produce text reports, either over serial/uart, to spiffs or over udp/tcp. I also have an option to stream raw trace data (far smaller and compressed than text), over the same routes (serial, spiffs and tcp/udp) which is then analysed off-line on your PC (the same analysis software runs on the pc as does the esp8266 so it is guaranteed to be the same results). I have designed all the software to be modular and platform agnostic so all of the tools run on Linux as well as esp8266 - I actually develop a lot of it on PC - I can also use the trace/profiling in the same way on the PC as I do internal on the esp8266.
The PC TRAP tool has a Linux/posix 'wrapper' around the core and uses libConfig; so it will not run on Windows. Perhaps someone would help to port onto windows once it is ready for use in a couple of weeks.
Hope that helps.
Carry on the questions as you have them - I enjoy the fedback.