[PATCH] Planner notes - revise logic for gasmix output

[Dirk Hohndel]

On Tue, Jun 23, 2015 at 09:15:25AM +0300, Benjamin wrote:
> Can the tests for NULL hurt? If not, it's always a good thing to have them.

They can make code harder to read and they do take a tiny amount of time,
but when in doubt... checking for NULL is almost always a good idea.
Coverity sometimes goes crazy and complains about code that is clearly
correct. But if it shows a potential NULL dereference, I try to add a
test.

> After all, tasks dying due to random NULL pointers are so fun to see in
> production systems. Not that I'm crying about code quality in various
> things at $WORK at all...

Hehe...

/D