svn commit: r313082 - head/audio/milkytracker

[Alexey Dokuchaev]

On Sun, Mar 17, 2013 at 09:15:23AM +0000, Chris Rees wrote:
> On 17 March 2013 01:48, Eitan Adler <eadler at freebsd.org> wrote:
> > On 28 February 2013 03:54, Emanuel Haupt <ehaupt at freebsd.org> wrote:
> >>  .if ${ARCH} == "powerpc"
> >>  BROKEN=                Does not compile on powerpc: fails in assembler
> >> +.else
> >> +BROKEN=                does not build
> >>  .endif
> >
> > Is it intended that this port eventually compile on powerpc?  If not,
> > perhaps its best to use NOT_FOR_ARCH?
> 
> NOT_FOR_ARCHS should be for ports that are designed not to work on an
> arch, eg something that deals with internals of CPU or hardware that
> could only possibly be used on certain architectures.

Correct.

> This could be useful if someone submitted a patch, or a change in base
> may cause it to work again.

The problem here that while ports are often marked as BROKEN on those tier-2
arches my portmgr@ (since they have access to the hardware), it's usually
hard to impossible for maintainers to fix them, since they don't have access
to such hardware, and AFAIK, we currently do not offer such access (not
even speaking of jails with full root access to be able to install required
dependencies).  Simple ports can be tested and fixed on pluto/flame/etc.,
but that covers only tiny fraction of broken ports.

./danfe