re: Automated report: NetBSD-current/i386 build failure

[matthew green <mrg%eterna.com.au@localhost>]

Andreas Gustafsson writes:
> The build is still failing as of source date 2018.06.03.05.55.08,
> now with the following error messages:
> 
> --- dependall-kdump ---
> In file included from /tmp/bracket/build/2018.06.03.05.55.08-i386/src/sys/net/if.h:99:0,
>                  from kdump-ioctl.c:23:
> /tmp/bracket/build/2018.06.03.05.55.08-i386/src/sys/altq/if_altq.h:48:2: error: unknown type name 'kmutex_t'
>   kmutex_t *ifq_lock;
>   ^~~~~~~~
> In file included from kdump-ioctl.c:23:0:
> /tmp/bracket/build/2018.06.03.05.55.08-i386/src/sys/net/if.h:211:2: error: unknown type name 'kmutex_t'
>   kmutex_t *ifq_lock;
>   ^~~~~~~~
> 
> Looks like this latest error was introduced by the following commit:
> 
>   2018.06.02.20.07.15 mrg src/usr.bin/kdump/mkioctls 1.51
> 
> The build has now been broken in four different ways within 24 hours,
> without a single successful build inbetween.
> 
> If would be helpful if developers could (a) compile test their changes
> with a full release build before committing, (b) fix or revert their
> commits quickly if they nonetheless turn out to break the build, and
> (c) wait until the tree is building successfully before committing
> unrelated changes, so that those changes can benefit from automated
> testing.

FWIW, i tested my change on about 5 platforms, but all
netbsd host.  i can't reproduce the above, unfortunately.


.mrg.