Small nitpick: The title and commit message should be cpu/native instead of cpu: native:.

Otherwise I just wonder how we could ensure that this would caught by CI in the future (but that's not really part of this PR).

Murdock results

✔️ PASSED

af6da34 cpu/native: use correct isr_stack for valgrind

Artifacts

• Documentation preview

Totally fine with changing this but do we have well-defined rules for that? But indeed there seems to be a preference toward the proposed format:

> git log --oneline | grep -c "cpu: native:"
19
> git log --oneline | grep -c "cpu/native:" 
145

But some people seem to prefer native: only:

> git log --oneline | grep -cE "[[:xdigit:]]{10} native:"
139

Totally fine with changing this but do we have well-defined rules for that?

Indeed, there is the commit convention: https://github.com/RIOT-OS/RIOT/blob/master/CONTRIBUTING.md#commit-conventions

It seems like we can blame Kaspar mostly for the exceptions 🤣

cbuec@W11nMate:~/RIOTstuff/riot-gettingstarted/RIOT$ git log --pretty=format:"%an, %s%n" | grep "cpu: native:"
Oleg Hahm, cpu: native: use correct isr_stack for valgrind
Kaspar Schleiser, cpu: native: uart: adapt to changed DEBUG
Kaspar Schleiser, cpu: native: always select periph_uart
Kaspar Schleiser, cpu: native: reorganize Makefile.features
Martine Lenders, cpu: native: make syscalls vfs ready (introduce real_fcntl)
Kaspar Schleiser, cpu: native: add vfs wrappers
Kaspar Schleiser, cpu: native: fix thread_yield_higher inisr() case
Kaspar Schleiser, cpu: native: use gnu99 C dialect
Iván Briano, cpu: native: Add [v]fprintf to syscalls
Kaspar Schleiser, cpu: native: make netdev2_tap internal functions static
Kaspar Schleiser, cpu: native: netdev2_tap: don't pass isr_arg on rx complete event
Kaspar Schleiser, cpu: native: adapt xtimer backoff values to periph/timer change
Kaspar Schleiser, cpu: native: netdev2_tap: make use of netdev2_eth module
Kaspar Schleiser, cpu: native: remove hwtimer traces
Kaspar Schleiser, cpu: native: minor timer-related doxygen updates
Kaspar Schleiser, cpu: native: remove hwtimer_compat dependency
Benoît Canet, cpu: native: switch to hwtimer_compat
Kaspar Schleiser, cpu: native: work around shared errno in _native_lpm_sleep
Kaspar Schleiser, cpu: native: add tap implementation of dev_eth (ng_nativenet)

To be honest this looks to me rather like an example than a rule. So, if we want to streamline this, we should be probably more specific in that document. Question here is: do you see an advantage of being more pedantic here?

To be honest this looks to me rather like an example than a rule. So, if we want to streamline this, we should be probably more specific in that document. Question here is: do you see an advantage of being more pedantic here?

I do believe that it makes sense to keep the commit messages uniform. Of course there are exceptions, but sometimes something slips through of course and there might be some stuff from "back in the day" as well.

Maybe that discussion should be outsourced to an Issue or the Matrix chat though.