[riot-notifications] [RIOT-OS/RIOT] board: add hamilton board support (#9013)
notifications at github.com
Mon Apr 1 17:46:34 CEST 2019
Here's my assessment of the issues and a proposed course of action.
Stdio_rtt has a dependency on xtimer being initialised. It's meant to act the same way as stdio_uart - it shares the API functions `stdio_read`, `stdio_write` and `stdio_init`. However, one difference is that stdio_rtt is dependent on xtimer, whereas stdio_uart is not. And xtimer is initialised centrally, by auto_init, so it would break current paradigms to initialise it in stdio_rtt. But we've currently got this "all-or-nothing" approach to initializing modules in auto_init, which effectively means that to make RTT work in current paradigms we have to initialise everything. Ok usually, but not for tests.
Personally I think if we want to make stdio_rtt a transparent swap-in for stdio_uart, we should configure an RTT timer in the board config code and use this instead of xtimer. Then we wouldn't need auto_init.
Checking the Ruuvitag and Thingy52, it seems they have the same treatment - i.e. stdio is provided over RTT by default. So, no `make test` available.
*Proposed way forward*
For `make test`, I suggest we stick with current paradigms and merge as-is, document the flashing method (obviously), and raise an issue describing the situation. For the xtimer dependency issue, currently we don't have an RTT stdio which imitates the uart stdio transparently, and so it shouldn't share a test. I would propose for now a shell_rtt test with xtimer_init() in the main() function, and raise an issue to merge this into the other shell test once the dependency on xtimer is removed.
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the notifications