core: assert correct msq queue size on creation

[OlegHahm]

The return value was never checked. Hence, this runtime check was rather pointless. Better assert the correct size during development.

[kaspar030]

ACK.

[kaspar030]

wow, someone actually did check the return value:

zep:native:
Building application "zep" for "native" with MCU "native".

/data/riotbuild/sys/net/gnrc/application_layer/zep/gnrc_zep.c: In function '_event_loop':
/data/riotbuild/sys/net/gnrc/application_layer/zep/gnrc_zep.c:564:9: error: void value not ignored as it ought to be
     if (msg_init_queue(msg_q, GNRC_ZEP_MSG_QUEUE_SIZE)) {
         ^

[OlegHahm]

Should we grant an award for this? Like "World-1st-msg_init_queue-Checker"?

Fixed.

[cgundogan]

why? cib_init is already asserting that. And the only reason for asserting that here is, because the msg_queue uses cib?

[OlegHahm]

Ah, didn't check this. Well, then I think we can remove the check completely. 👍

[OlegHahm]

addressed

[kaspar030]

pls squash!

[cgundogan]

regarding the assert: the cib only checks for power of 2 values, but also allows 0 in the assert. The former assert in msg prevented 0. Maybe use an assert(num) instead in msg_init_queue?

[OlegHahm]

regarding the assert: the cib only checks for power of 2 values, but also allows 0 in the assert. The former assert in msg prevented 0. Maybe use an assert(num) instead in msg_init_queue?

I noticed that, but I came to the conclusion that calling msg_init_queue() with size zero sounds really esoteric.

[miri64]

ACK by @kaspar030 seconded.

[OlegHahm]

squashed