> Thus, I would rather report this upstream (possibly fixing the dnsperf
> copy in contrib/) than add a configure option to BIND as a workaround.
> Am I missing something?

I think originally dnsperf was building fine, then we turned on
ISC_BUFFER_USEINLINE in buffer.h and it started emitting a warning,
then we turned it into a boolean value and it broke completely.

Being performance-sensitive, it was reasonable for dnsperf to use
this option, and it remains reasonable even if libisc isn't built
with it.

I'm fine with reporting it upstream and fixing it in contrib but
I also think this is a good change to make.