Kaspar Schleiser
Kaspar Schleiser
This is running `tests/bench_xtimer` with `USEMODULE=ztimer_xtimer_compat` on nrf52840dk, first on master, then this PR (rebased on master): ``` main(): This is RIOT! (Version: buildtest) xtimer benchmark application. set() one 6501...
I like this, too!
> We still could consider testing on FreeBSD, as `native` was originally developed with support for it. And let's not forget armhf. There are plenty of RasPi's that could test...
> @kaspar030 Does the pifleet also test `native` (that is, on Linux/ARM)? Nope, not at the moment. They're also not really equipped to compile much themselves.
There's also #4294. It can use drift between two crystals when there are no "natural" random sources.
... and they are supposed to be used in compile-time situations, I guess?
> How about asking the caller of thread_create() to allocate a thread_t instead? IMO that would just require two allocations instead of one: thread_t and stack itself. I know there's...
> I strongly agree. Also @benpicco was in favor, but @kaspar030 shot that idea down. I changed my mind on that one :/
> I don't see the problem with That would change to sth like ``` struct thread_info { thread_t thread; char stack[MY_MODULE_STACK_SIZE]; } int mymodule_init() { for (i=0; i